~~~~g~

POLICE USE OF

DEADLY FORCE

Edited by
James J. Fyfe

PoliceFoundotion
The Police Foundation is a privately funded, independent, non-profit organiza­
tion founded by the Ford Foundation in 1970 and dedicated to supporting
innovation and improvement in policing. This book of readings is published by
the Foundation as a service. Opinions and recommendations are those of the
authors and not necessarily those of the Police Foundation.

Copyright c 1982, Police Foundation

Library of Congress Catalog Number 81-86057

 FOREWORD

The distinguishing characteristic of policing is the authority

to use force. With this authority, of course, comes the responsi­
bility never to misuse force . This responsibility translates into
an imperative on the part of police management to control po­
lice discretion so that officers employ only that degree of force
necessary to do their job fairly and humanely.
The use of force at its most extreme is the use of deadly
force which, with rare exception, can be described as a decision
of a police officer to point a service revolver at another human
being and fire it. This is the most momentous decision a human
being can make-to take another life. Limiting such decisions
to those instances when the use of force is absolutely necessary
is one of the most important goals for the police agency. This is
so, not only to reduce death and injury, but also to diminish the
often woeful impact that police woundings and killings have on
citizens' perception of the fairness and decency of police agen­
cies.
In strictly supervising the authority to use deadly force,
police chiefs and their fellow administrators need all the reli­
able information they can obtain so that their decisions are
based on a solid foundation. To help police administrators with
the task of formulating and enforcing deadly force policies, the
Police Foundation has published this anthology of what we be­
lieve to be the best available -current research on the subject.
Although this volume will be of use to researchers, legisla­
tors, lawyers, and judges who must deal with deadly force, it is
meant primarily for police chiefs and other local municipal offi­
cials charged with controlling crime and maintaining order. Our
goal is for this volume to be a principal reference tool for police

lll
chiefs, mayors, managers, and council members as they go
about the job of preparing and implementing police department
rules and regulations on how and when deadly force is to be
used.

Patrick V. Murphy
President
Police Foundation

iv
 ACKNOWLEDGMENTS
Each of the readings in this book originally appeared in
another publication . Thus, acknowledgement is owed both the
authors of the readings included in this volume, and the publica­
tions which consented to their consolidation in one volume.
I wish also to express appreciation to Police Foundation
President Patrick V. Murphy and to Police Foundation Re­
search Director Lawrence W. Sherman, both of whom strongly
encouraged publication of this book. Finally, I also owe thanks
to Thomas V. Brady, Mara Adams, and Sheila Bodner of the
Police Foundation for their valuable assistance.

James J. Fyfe

v
 CONTENTS

FOREWORD Ill

ACKNOWLEDGMENTS v
I. INTRODUCTION 3

II. FREQUENCY OF POLICE USE OF DEADLY FORCE

Measuring Homicide by Police Officers 12
Lawrence W. Sherman and
Robert H. Langworthy
Analysis ofShooting Incidents 42
Catherine Milton, Jeanne Wahl Halleck,
James Lardner, and Gary L. Abrecht
III. HISTORY AND LAW
Use ofDeadly Force to Arrest a Fleeing Felon-
A Constitutional Challenge, Parts I and III 65
J. Paul Boutwell
Execution Without Trial: Police Homicide
and the Constitution 88
Lawrence W. Sherman
IV. RACE AND POLICE USE OF DEADLY FORCE
Minorities as Victims of Police Shootings:
Interpretations of Racial Disproportionality
and Police Use of Deadly Force 128
JohnS. Goldkamp
Police Shootings at Minorities:
The Case ofLos Angeles 152
Marshall W. Meyer
Race and Extreme Police-Citizen Violence 173
James J. Fyfe
A Garrison State in a "Democratic" Society 195
Paul Takagi
v. DEADLY FORCE POLICIES AND PRACTICES
Use ofDeadly Force to Arrest a Fleeing Felon-
A Constitutional Challenge, Part II 214
J. Paul Boutwell
Police Policy on the Use ofFirearms 224
Samuel G. Chapman
Administrative Interventions on Police Shooting
Discretion: An Empirical Examination 258
James J. Fyfe
Always Prepared: Police Off-Duty Guns 282
James J. Fyfe
Observations on Police Deadly Force 297
James J. Fyfe

VI. ABOUT THE AUTHORS 315

 INTRODUCTION

In 1972, I was a New York City police sergeant on leave to

pursue graduate studies in criminal justice. Anxious to com­
plete my degree requirements with a dissertation that would be
useful to police administrators, I wrote a proposal for an empir­
ical study of the relationship of police corruption to the charac­
teristics of police organizations and individual officers. Since
the data required for such a study consisted largely of records
possessed by my employer, I forwarded the proposal to a de­
partmental "research liaison committee," with which I subse­
quently met to discuss the research.
The committee's reaction to the proposed study was luke­
warm at best. After several hours of wrangling, the committee
granted me access to the necessary data, letting me know in no
uncertain terms that I had its permission, but not its blessing,
to conduct the proposed corruption study. Corruption, commit­
tee members argued, was a highly politicized phenomenon
which could not be studied in a meaningful way by academics
whose tools were limited to the computer and the statistical
techniques of social science. I disagreed with their assessment
of the limits of social science research, and with their assess­
ment of me as an academic: I was a police sergeant who had
gone to graduate school right out of eight years in a patrol car.
I was also a realist, and I got their message. The message
was that corruption was a controversial topic that had just been

3
 given a highly visible public airing in the Knapp Commission
hearings and in the revelations of Frank Serpico and Robert
Leuci. The department would allow me to conduct my research,
but it really wished that attention would be focused on other
less volatile subjects while they attempted to address the cor­
ruption problems that were still in the process of being identi­
fied.
I put my corruption study on hold for a while, and returned
to patrol duty in June of 1973. Shortly thereafter, I attended a
firearms training program, and learned from instructors that
they had collected a data base on shootings by the department's
officers. I looked at the data, found that they included all re­
ports of shootings by police and serious assaults upon officers,
and felt that I had found my dissertation topic. Deadly force
was important, I was not aware of any significant literature on
it, and, best of all, it was neither controversial nor politicized. It
promised that I could do my research in an academically appro­
priate, tension-free environment.
I knew that some individual police shootings had raised
questions, and I knew that some of the riots of the 1960s-New
York's, for example-had followed police shootings, but I also
knew that, with very rare exceptions, officers used their guns
only reluctantly and with the greatest respect for human life. I
had never shot anybody, but I knew officers who had, and I
knew how greatly they were affected by having had to kill or
wound someone.
I began a review of the literature on deadly force, and
found that very little had been published on this topic. There
existed some law review articles and comments, some police
training material, a few small-scale empirical studies, and some
policy recommendations that had been written on the heels of
the urban disorders of the previous decade. Even though news­
papers regularly reported accounts of police shootings, the
scholarly literature included no large-scale empirical studies of
police use of deadly force.
Over the next five years, I designed the study, obtained
permission to do it from both the State University at Albany

4
and the New York City Police Department, and did it. I got
only cooperation from the police department and was able to do
my research with no restrictions whatever. Any tensions cre­
ated by controversial shootings which occurred during the
course of my research never affected my access to data, and
never resulted in limitations of any kind. When my work was
done, I felt quite good about my findings. The data I analyzed
included some shootings of questionable necessity but, in the
main, I found that my colleagues exercised their life or death
powers in a controlled and humane way.
In the course of my research, I became aware that other
people were studying police deadly force. I also became aware,
however, that researchers who had attempted to study deadly
force in other departments had encountered far greater resist­
ance than had my corruption study proposal; studying this sub­
ject was not tension-free everywhere, and much work was
aborted or delayed by denial of access to police data. Similarly,
after completion of my work, I found that many people chal­
lenged my remarks about the prudence exercised by police in
their use of firearms. As I reviewed data and research from
other jurisdictions, I realized that I had made the mistake of
assuming that my single-city findings were generally true of
police agencies across the country. New York has a reputation
as a crime ridden city, and I had found the frequency of police
shootings there to be closely associated with levels of crime and
violence within the city's different communities. I had assumed,
therefore, that New York City police officers were as likely to
kill as were police officers in any American jurisdiction. As I
have learned since, that is not so.
The small, but growing, body of literature shows that police
in some cities are far more likely to kill than are police in other
cities. In some cases, relatively high probability that an officer
will kill a citizen is associated with high levels of violence and
crime. But in other cases, police in comparatively crime-free
jurisdictions use their guns with great frequency . In some juris­
dictions, almost all those shot by police are armed and threaten­
ing, and leave officers no real choice but to shoot or be shot or

5
 stabbed themselves. In other cities, most of those shot by police
are unarmed and shot in the back as they flee from crimes
against property. In some jurisdictions, police are commended
for shooting people in circumstances that would result in dis­
missal had they involved officers in other police departments.
There are about 17,000 police departments in the United
States, and their operating policies and practices vary widely.
Some police departments enforce traffic laws vigorously, while
others often overlook traffic violations. In some police depart­
ments, officers regularly provide merchants with services that
would be regarded as compromising and unethical in other jur­
isdictions. Naively, however, I had assumed that police policies
and practices where questions of life or death are concerned
would be relatively constant across jurisdictions. I learned that
I was wrong, and that police deadly force policies and practices
vary as much as policies and practices concerning parking en­
forcement .
One possible reason for this variance among police depart­
ments is that no national reporting system on police use of
deadly force exists. True, the National Health Service annually
publishes figures on persons killed by the police, but, as
Sherman and Langworthy point out in this book's first chapter,
there is reason to doubt the accuracy of those figures. Even if
accurate, however, the National Health Service's Vital Statis­
tics would be instructive only on questions concerning persons
killed by police use of deadly force; they include no data on
persons shot and wounded or shot at and missed by officers
exercising their deadly force powers. Because no agency re­
ports in a systematic manner on use of deadly force among
American jurisdictions, there is no way for police in one juris­
diction to assess their use of deadly force in relation to use of
deadly force in other similar jurisdictions. By referring to the
Uniform Crime Reports, police chiefs can obtain some measure
of the relative frequency of crime in their jurisdictions. By re­
ferring to the same reports, citizens can obtain information on
the relative effectiveness of their police in solving crime. Even
though the Uniform Crime Reports annually reports on killings

6
 of police and assaults upon officers, there exists no such report­

ing system to assess the relative frequency of deadly force

among American jurisdictions, or to compare the circumstances

in which citizens are killed by police in those jurisdictions. If

such a system did exist, it would show that police in some

. jurisdictions are as much as ten times as likely to kill citizens as

are police in other similar jurisdictions.

It is important to collect and disseminate information about
deadly force policies and practices among American jurisdic­
tions. Such information would embarrass some police agencies,
but that consideration is not an adequate reason to refrain from
providing American citizens and officials with information on
how often their police officers use deadly force. Neither is it the
point of disseminating such information to embarrass police
departments in which rates of deadly force are high. The rea­
son that it is important to disseminate such information is to
make citizens and police in jurisdictions with high shooting
rates aware that police in other jurisdictions cause far less
bloodshed while doing the same job and facing the same haz­
ards with equal effectiveness and no greater risk to themselves.
Since it is the primary obligation of the police to protect life,
such information would encourage citizens and police in juris­
dictions with high rates of deadly force to reexamine their poli­
cies and practices. The limited information now available on
police deadly force has meant that policies and practices among
this country's 17,000 police departments often have been devel­
oped in a vacuum, and often are based on untested assumptions
rather than upon the collective experience of police throughout
the country.
Ironically, it is important to provide information on police
use of deadly force because it is employed so infrequently and
is, thus, often regarded as insignificant and not problematic.
True, every violent death is a tragedy and, in comparison to
deaths by execution, deaths at the hands of the police occur
with great frequency. Since 1967, we have executed four con­
victed murderers in this country; estimates of the numbers of
people shot and killed by police in the same period vary be­
tween 4,000 and 8,000. There is, of course, no reliable way to

7
estimate the number of people wounded or missed by police
bullets during those same years. Thus, in comparison to deaths
by execution, deaths by police use of deadly force are a phenom­
enon of enormous proportions.
From another angle, however, the frequency of deadly
force looks quite small. There are 17,000 police departments in
the United States, and American police have killed somewhere
between 4,000 and 8,000 people over the last 15 years. Thus,
arithmetic indicates that most police departments have killed
nobody over the last 15 years. There are nearly a half million
sworn police personnel in this country, and they encounter
thousands of potentially violent situations every day; but even
the greatest estimates of killings by police indicate that fewer
than one in 60 officers have killed anybody over the last 15
years. Further, the limited data available suggest that most
police shootings take place in large cities, so that rates of police
killings in America's small jurisdictions are probably far lower
than this one in 60 aggregate rate.
Understandably, therefore, police use of deadly force is not
viewed as an issue of importance by citizens and police in many
of America's smaller jurisdictions. Frequently, I hear police
chiefs from smaller departments question the relevance to their
departments of both research into deadly force and recommen­
dations for unambiguous departmental deadly force policies
which further limit the broad shooting discretion found in most
state laws. "Deadly force is a big city problem," they argue,
"my department has 30 people, and the only person we've ever
shot was a rum runner in 1926. We've been doing just fine the
way things are. I know my men, and I trust their judgment.
The law in my state says that they can use their guns to defend
themselves or innocent citizens, or to apprehend fleeing felons.
Why should I make waves and get more specific than that if we
have no problem?"
One reason for such a chief to make waves is that, in addi­
tion to its consequences upon its victim, a single shooting can
have severe consequences for the community, for the depart­
ment, for the officer involved, and for the chief. How would the

8
citizens of that community react if, in a bad moment, one of the
chief's officers exercised the great discretion available under
law by shooting a local youth fleeing from the scene of such a
felony as the theft of a bicycle? Such a shooting can unravel
years of good police work. Such a shooting can expose a police
agency to great civil liability. Such a shooting can torment an
officer who acted within the vague rules defined for him or her,
and who did not realize the enormity of the act until it was too
late for any corrective action. Such a shooting can be prevented
by the simple expedient of directing officers not to use their
guns in such nonthreatening circumstances. By refraining from
"making waves" while deadly force is not a problem, such a
chief may be postponing action until too late .
This book is an attempt to provide information on deadly
force policies and practices. It draws upon the literature and
research of deadly force, most of which has appeared in the last
several years. Its origins may be traced to constant requests to
the Police Foundation for information on deadly force . These
have come from mayors, county executives, legislators, city at­
torneys, private attorneys, public interest law groups, police
chiefs, sheriffs, police organizations, citizens' groups, and the
media. Often these requests come from small and medium sized
jurisdictions in which single shootings have made deadly force a
major concern. In some cases, that concern has also expressed
itself in disorder, protests, and tensions which have led to the
downfall of city administrations and police chiefs, and in enor­
mous burdens to tax payers. A small midwestern jurisdiction,
for example, was recently directed in a jury decision to pay
$5.75 million to the survivors of a young man shot and killed by
its police department.
The authors whose works appear in this book address some
of the more frequent and volatile questions about deadly force.
The book opens with a section on the frequency of police deadly
force. In the first chapter, Lawrence W. Sherman and Robert
H. Langworthy test the adequacy of existing data on police
deadly force, and conclude that there is considerable room for
improvement. Next, in an excerpt from their seminal study of

9
deadly force, Catherine H. Milton and her colleagues analyze
police shooting incidents in several cities.
In the book's second section, separate chapters by J. Paul
Boutwell and Lawrence W. Sherman discuss one of the most
controversial deadly force related issues: the power of the po­
lice to shoot suspected fleeing felons.
Questions about the relationship of race to deadly force are
treated in four chapters in the book's next section . Regardless
of its inadequacies, the information available clearly indicates
that minorities-especially blacks-are shot by police in num­
bers that greatly exceed their representation in the general
population. John S. Goldkamp provides a thoughtful essay
about the relatively great frequency of police killings of mem­
bers of minority groups, and describes two alternative interpre­
tations of this racial disproportion. Three analyses of deadly
force by Marshall W. Meyer, this editor, and Paul Takagi are
then presented.
Issues concerning internal police departmental policies on
deadly force and firearms are discussed and analyzed in the last
section of the book. J. Paul Boutwell of the Federal Bureau of
Investigation starts this section with a discussion of the civil
liability of police evolving from use of deadly force. Next, in one
of the earliest and most influential discussions of deadly force,
Samuel G. Chapman, who studied this issue for the President's
Commission on Law Enforcement and Administration of Jus­
tice, addresses the need for internal departmental shooting pol­
icies. The book closes with three of my own pieces on deadly
force. One describes the effects of a restrictive deadly force
policy in the New York City Police Department_; a second ana­
lyzes questions and data related to use of firearms by off-duty
officers, and the third examines policy related questions, and
offers specific recommendations on effective management of
police deadly force .
My own feelings about this book are mixed. On the one
hand, I am delighted to have witnessed the development of a
body of literature on police deadly force. On the other hand, it
is troubling that this body of literature is of such recent vin­

10
tage: one would have thought, that by this stage in the develop­
ment of our democracy, questions about the authority of the
state to kill would long ago have been settled. It is troubling
also that deadly force is such a divisive topic. As I noted earlier,
in New York City in 1973, deadly force was perceived as nei­
ther controversial nor politicized, and it remains that way there
and in most American police jurisdictions. In other places, how­
ever, it is highly controversial and highly politicized. Access to
data on when and under what circumstances police kill fre­
quently is denied, and those who have merely reported upon the
findings of their research have been castigated for doing so. I
hope that this book, which includes the perspectives of govern­
ment officials, police practitioners, and academic criminologists
of both the traditional and radical persuasion contributes to
both the enhancement of knowledge about deadly force and the
reduction of tension caused by studying it and discussing it.

James J. Fyfe

11
 CHAPTER 1

MEASURING HOMICIDE BY POLICE

OFFICERS

LAWRENCE W. SHERMAN AND

ROBERT H. LANGWORTHY

Criminologists have long viewed homicide as the least diffi­

cult type of crime to measure. 1 The difficulty of disposing of
bodies, the generally high level of agreement between the Uni­
form Crime Reports and the Vital Statistics of the United
States, 2 and the monitoring function of coroners in recording
homicide events all support the view that official statistics pro­
vide a highly accurate measure of homicide. The excellence of
this official measurement, however , is confined to citizens kill­
ing other citizens. The official measurement of officials killing
citizens falls far short of excellence. The widespread American
belief that official killings do not constitute violence3 is reflected
by the complete absence of such killings from the Uniform
Crime Reports, 4 most police departments' annual reports , and
the limited summary treatment they receive in the Vital Statis­
tics, where no figures are published below the state level.
The paucity of official data on official killings has become
more noticeable in recent years as both public and scholarly
interest in police-caused homicide has intensified. Public policy
debates questioning the propriety of police use of deadly force,
SOURCE: "Measuring Homicide by Police Officers," Journal ofCriminal Law
and Criminology (December, 1979),70:4:546·560. Copyright 1!!79 by Northwest·
ern University School of Law. Reprinted by special pe rmission of the Journal of
Criminal Law and Criminology.

12
 often prompted by major protest demonstrations after specific
police-homicide incidents in minority communities, have com­
manded the attention of the United States Civil Rights Com­
mission, the Department of Justice, and even the White House. 5
Both legal scholarship6 and empirical research 7 have reflected
the growing public concern with this category of homicide, to
which the Vital Statistics attributes 1.77 percent of all homi­
cides in the United States from 1971 to 1975.8 But that figure,
like many others used in this area, has yet to be examined
critically through comparisons with other sources of data. Pol­
icy discussions and empirical research both require that the
problems of measurement be addressed before any conclusions
are drawn from the available data.
The adequacy of current methods of measuring homicide by
police officers poses three important questions. A first question
is whether the number of these killings occurring each year
throughout the country can be measured . While the quest for
an accurate count of the "absolute incidence" of any form of
conduct may be futile, 9 it is not unreasonable to expect a society
to know how many of its citizens are killed by officials acting
under what is ruled by other officials (i.e., police chiefs, prose­
cutors, grand juries, judges, or juries) after the fact to be
proper use of the authority of the state. Without some approxi­
mation of the actual number of events that fit some consistent
definition of police killings, it is difficult to address the public
policy issues raised by those events at the national level.
A second question is how well the relative incidence of po­
lice killings from one police department to the next can be
measured. Local public policy debates over the quality of police
services often focus on specific police shooting events, but they
could just as easily focus on comparisons to other cities. For
example, the fact that city X has twice the rate of police killings
as city Y, which is similar to X in other important respects,
could be most relevant to the evaluation of a police chief's per­
formance, the selection of a new firearms policy, or a decision
about what size gun the police should carry. All of these deci­
sions require accurate measurement of the relative incidence of
police homicides across specific cities.
13
 The third question, and the one most relevant to crimino­
logical theory, is whether the pattern of differences across po­
lice departments in police homicide rates can be measured to
explain that pattern with theoretical and public policy variables.
This question is related to, but distinct from , the question of
how accurately specific cities can be compared. For as it will be
shown, ... available measures contain too much error either to
estimate the national incidence of police killings or to make
reliable comparisons of specific cities, but not too much error to
compute apparently valid statistical relationships between po­
lice homicide rates and other characteristics of police depart­
ments and the communities they serve.

AVAILABLE SOURCES OF DATA

Three basic sources of data on homicides by police officers
are generally available: death certificates, police department
internal affairs records, and newspaper stories. A fourth
source, the supplemental homicide reports filed by police de­
partments with the Uniform Crime Reporting Section of the
FBI, is not generally available to researchers because of the
FBI's reservations about the quality of those data.W Each of the
available data sources has substantial limitations.

DEATH CERTIFICATES
If the American system of vital statistics actually worked in
the manner its federal overseers intend it to, then death certifi­
cates would provide a nearly perfect count of official homicides
by police officers throughout the country. Assuming that the
system works as intended, most of the empirical studies of
police homicides have made some use of the national and state
level tabulations of the death certificates reporting the cause of
death published by the National Center for Health Statistics.11
Unfortunately, at least six major flaws in the system cause it to
grossly underestimate the number of "deaths by legal
intervention-police," defined by the International Classification
of Diseases as "injuries inflicted by the police or other law­
enforcing agents, including military.
on duty,
.
in the course of

14
arresting or attempting to arrest lawbreakers, suppressing dis­
turbances, maintaining order and other legal action." 12
American vital statistics are part of a world health statis­
tics system in which causes of death are defined and agreed
upon by the periodic Geneva conventions that revise and
promulgate the International Classification of Diseases (ICD).
Membership in the system and use of the ICD at all levels is
voluntary, and within the United States it extends down the
federal ladder to each county's chief medico-legal officer (usu­
ally either an elected coroner or an appointed medical exam­
iner). The system employs a standard death certificate (or a
variant which contains the same information) which each state
must use in order to participate in the national death registra­
tion system. 13 "Natural" or usual deaths may be certified by
any licensed medical doctor. Medico-legal officers must fill out
the death certificates on violent and other unusual deaths (their
usual jurisdiction amounting to about 20 percent of all death
certifications nationally), 14 ideally supplying all the information
necessary for classification of the cause of death according to
the ICD categories. The death certificate then goes to the fu­
neral director, who in turn secures a burial permit from the
local registrar, who then records the death and forwards the
death certificate to the state registrar. The state registrar re­
cords the death and sends an official copy of the death certifi­
cate to the National Center for Health Statistics (NCHS),
where coders assign each case to one of the ICD categories and
enter them into the national mortality data published in the
annual Vital Statistics of the United States. 15
Almost every step of this system is vulnerable to serious
flaws. The first flaw is the often poor quality of the medical
diagnoses of the causes of death. Two studies conducted in the
early 1950s showed high rates of error by either attending phy­
sicians or coroners' physicians. One study found 39 percent of a
Pennsylvania sample of death certificates to be based on
"sketchy" diagnostic information, with 18 percent having an
equally likely or preferred diagnosis. 16 More relevant was an
independent study of 1,889 autopsied deaths in Albany, New

15
York, in which the medical researchers concluded from their
own evaluation of the recorded clinical information, autopsy
protocols (reports), and laboratory reports that 57 percent of
the homicide and suicide deaths in the sample could have been
misclassified as to the circumstances of death. 17
No matter how accurate the diagnosis, however, a second
flaw in the system seriously hinders accurate data collection:
the apparently widespread lack of the coroners' awareness of,
support for, and legal obligation to comply with the system's
request for the full information necessary to code the causes of
death according to ICD categories. One leading medical exam­
iner has claimed that his colleagues around the country are
generally "turned off" by the ICD categories, particularly
where any stigma to the victim or his family may result from
the use of the categories. 18 A board-certified forensic patholo­
gist (a level of technical qualification only some medico-legal
officers attain) observed that those with her qualifications may
be more likely to be aware of the ICD categories, but not neces­
sarily more likely to employ them or provide information con­
sistent with them. 19 Even the Model State Vital Statistics Act
published by the NCHS fails to make any mention of the ICD
categories, let alone require compliance with them. 20
The lack of concern for the ICD categories exacerbates a
third flaw in the system: the vagueness of the instructions for
completing the Standard Death Certificate. This vagueness fa­
cilitates the omission of the information necessary to distin­
guish a civilian-caused homicide from a death by legal interven­
tion of police. This is especially true since the critical
information is supplied in item 20d of the certificate, "How
Injury Occurred," which has a very small space with room for
only five or six words. The NCHS handbook on death registra­
tion fo r medico-legal officers paradoxically urges both "com­
plete reporting" and the use of "as few words as possible [to]
describe the injury-producing situation."2 1
The latter principle is clearly evident in one of the hand­
book's examples that might be relevant to police-caused homi­
cide. In the example, a pulmonary hemorrhage due to stab

16
wounds is described in item 20d as "stabbed by a sharp instru­
ment."22 No mention is made of who did the stabbing; it could
have been either a criminal assailant or a police officer defend­
ing himself when attacked during a family fight. Since there
are known instances of facts being omitted,Z3 it is likely that
critical information about police officers is omitted from the
responses to the vague question of "How Injury Occurred." 24
In fact, omission of the police role in a killing may often be
quite probable given a fourth flaw in the system: the close
relationship between the local police and the medico-legal of­
fice. A case study of a rural coroner's office found that
[tlhe coroner is enmeshed in the legal-political structure
of the county in which he practices. This immersion
places upon him certain informal controls which can be
exercised to insure continuing cooperation between the
Coroner, Sheriff, Prosecuting Attorney and the medical
community. These informal restrictions may be as sig­
nificant as the law in determining cause of death proce­
dures.25
This relationship may well lead medico-legal officials to
omit police involvement from the information they provide on
how the injury occurred. One forensic pathologist observed:
The ease of doin!f the job and serving the public in a
medical examiners or coroner's office largely depends
upon the cooperation of the police. So it doesn't help to
antagonize the police unnecessarily. On the other hand,
the doctors won't pull a cover-up job. When you sign
the certificate, you have to put down homicide. You just
may not put down the full background circumstances of
death. 26
The relationship between the doctors and the police may be
as much individual as it is organizational, which exposes a fifth
flaw in the system: diversity of procedures used (and complete­
ness of information supplied on the death certificate) among
different coroners, even within the same office. In the New
York City Medical Examiner's office, for example, the older
examiners rarely indicate that police effected a homicide be­
cause they feel it places an "unnecessary onus" on the police. A
recent chief medical examiner in New York City encouraged his

17
colleagues to indicate police involvement, a policy contrary to
that of his predecessors. But each examiner still makes his own
decisions about how to fill out the death certificate Y
The sixth flaw in the system is that the transmission and
coding of the data suffers both mechanical and conceptual er­
rors. On one occasion, 6,000 death certificates were lost during
transmission from Massachusetts to the NCHS. 28 More impor­
tant, however, may be the complete lack of any coding instruc­
tions, other than the ICD definition quoted above,Z9 for death by
legal intervention of the police. Thus, while NCHS is able to say
publicly how it would code borderline situations such as an off­
duty police officer killing his wife in self-defense, 30 it is not clear
that the coding would always follow the publicly provided inter­
pretations. Ambiguity of the coding rules is further suggested
by the disagreement between the tabulations of the New York
City Health Department (equivalent to a state·level death regis­
trar reporting directly to NCHS) and those of NCHS. In 1971,
the NYCHD counted thirty-three police homicides while NCHS
counted thirty-two; in 1972, the respective figures were thirty­
four and twenty-four; in 1973, thirty-seven and forty-one; in
1974, twenty-three and twenty-five; and in 1975, eighteen and
twenty.31 Since the differences vary in direction from year to
year, one may infer that the differences in coding decisions are
arbitrary rather than systematic.

POLICE INTERNAL AFFAIRS RECORDS

In large, bureaucratized police departments, specialized in­
ternal affairs units are usually responsible for all investigations
of possible serious criminal misconduct by police officers.32 This
often includes investigations of police use of deadly force, al­
though other units occasionally investigate such incidents.
Even when homicides by police are investigated by other units,
records of the investigations and the incidents may be stored at
the internal affairs unit. These records provide the basis for the
counts of homicides by police that some, but not all, police
departments supply on request to the news media and social
scientists.

18
 In the opinion of several police researchers, these records
usually provide fairly accurate counts of deaths caused by spe­
cific police departments. As the basis for national data collec­
tion on the incidence of police homicides, however, these re­
cords are limited, for they are generally not kept in smaller
police departments where police homicides also occur. 33 While
some states (California, Oregon, Minnesota, and others) now
require all police agencies to report these data as part of their
general homicide statistics to a state level crime statistics unit,
this practice is far from universal.
Although police records are not gathered for the purpose of
comparative analysis across large cities, they have been used in
that manner. 34 These data have at least four limitations as a
basis for comparative analysis. One is that many police depart­
ments refuse to make the data available to the public or to
researchers. Another limitation is that the figures that are
released sometimes are different from figures obtained from
other sources. Responding to a request from the New York
City Police Department, for example, the Dallas, Texas, police
department reported a lower count than had been reported in a
study of that department's records done by a local university. 35
A third limitation is the considerable cost involved in obtaining
data from hundreds or thousands of separate police depart­
ments. A fourth limitation arises even when figures can be
obtained, as differences in definitions may undermine the com­
parability of the data from one department to the next. Some
departments, for example, may omit accidental deaths, police
officer suicides, off-duty killings, or killings taking place outside
L1e city limits, while others may include them. In short, police
records seem to be as problematic as death certificates for both
nation-wide and cross-city measurement.

NEWSPAPER STORIES
In some cities, newspaper stories may provide the most
accurate count of police homicides. This will be true only where
a newspaper's editorial policy defines all homicides as newswor­

19
thy. An exhaustive reading of the back issues of such a newspa­
per, while highly labor-intensive, should yield a complete annual
count of such incidents. The Kansas City, Missouri, police de­
partment records, for example, show the exact count of police
homicides for the year 1974 as an exhaustive reading of the
Kansas City Star. 36 Yet editorial policies are subject to change,
and they vary from one city to the next. Many police homicides,
in the few large cities in which they are a common occurrence,
such as New York, are not reported in local newspapers. Conse­
quently, newspaper stories are of limited use for assessing the
relative incidence of police homicides across cities.
For similar reasons, news reports provide a poor basis for
measuring the absolute incidence of police homicides around
the nation. One study employed a national news-clipping serv­
ice throughout much of the 1960s, collecting over a thousand
reports of police homicides. 37 Our secondary analysis of a three­
year period of these data, however, showed that they yielded
substantially lower counts at the state and national levels than
the NCHS statistics derived from death certificates, with 53
percent fewer deaths nationally in 1966, 41 percent fewer in
1967, and 56 percent fewer in 1968. In only six states in 1968
did the newspaper count yield a higher figure than the NCHS
count. From seven to eleven states showed equal figures from
the two counts each year, but all of these had either zero or one
death reported per year. Not one state showed consistently
higher news-based counts than NCHS counts over the full
three-year period examined.
Every data source has certain problems, and what may ap­
pear on conceptual grounds to be a major flaw in the collection
of data may make little difference in practice. The flaws in
news-based counts of police homicide seem to be serious enough
to eliminate them from further consideration as a possibly use­
ful data source for most purposes, and the preceding empirical
analysis of those data supports that conclusion. The rest of the
article subjects the other two data sources to an empirical anal­
ysis designed to answer the three central questions about the
adequacy of the measurement they provide.

20
VITAL STATISTICS AS A NATIONAL MEASURE OF
POLICE HOMICIDE
The only nationwide data collection system on police homi­
cide is the vital statistics compilation of death certificate data.
Our empirical evaluation of the adequacy of vital statistics as a
national measure of police homicide consists of a comparison of
a nonrandom, convenience sample of those data to police­
generated data matched by place and time at the state level of
the jurisdictions examined and the county level for New York
City (see Table 1). The thirteen jurisdictions of the comparisons
include all those at the state and county level for which we
could obtain police-generated statistics. In nine of the thirteen
jurisdictions (not counting New York City totals) the death
counts from police-generated data for the total years available
exceed the counts of the vital statistics compiled by the Na­
tional Center for Health Statistics. In only three of the thirteen
do the NCHS figures exceed those based on police-generated
data, and in one of those jurisdictions (Nebraska) the difference
is only three deaths over three years. Moreover, in the two
jurisdictions besides Kings County (Brooklyn) in which NCHS
figures are larger, the police-generated data are derived from
the supplemental homicide reports to the FBI which the FBI
defines as unreliable. 38 The NCHS figure for Kings County is
larger than the police figure for two apparent reasons: 1) the
Brooklyn medical examiners probably provide full information
on the death certificates, as their chief indicated some of his
colleagues do, and 2) the Transit Authority Police, Housing
Authority Police, and other law enforcement agencies in New
York City also kill people, with those deaths possibly included
in the NCHS count but definitely not included in the New York
City Police Department count.
The most striking aspect of Table 1 is the more than 50
percent underreporting of the NCHS data relative to the police­
generated data, not just overall, but also within differing ele­
ments of the data: in New York City (total), in California, in the
heavily urban areas grouped together, and in the less urban
areas grouped together. According to NCHS national data, the

21
t-0 TABLE 1 VITAL STATISTICS• AND POLICE-GENERATED DATA ON POLICE HOMICIDES BY JURISDICTION AND YEAR
t-0

YEAR TOTAL YEARS AVAILABLE

FOR BOTH SOU RCES
1970 1971 1972.. 1973 1974 1975 1976
JURISOICTION VS PG % OIFF. VS PG o/o OIFF. VS PG % OIFF. VS PG % OIFF . VS PG % OIFF. VS PG % OIFF . VS PG % OIFF VS PG % OIFF .

1. Heavily Urban Areas

CALIFORNIA ' 41 46 - 11 56 93 - 40 16 76 - 79 37 64 - 42 35 84 - 58 36 87 - 59 36 94 - 36 257 544 - 53
NEW JERSEY' 10 2 +400 8 NA - 12 NA - 3 1 +200 10 4 + 150 2 NA - 4 NA - 23 7 +229
NEW YORK COUNTY'
BRONX COUNTY'
1 NA
2 NA
-
-
638
5 13
-
-
84
62
0 23
4 15
-
- 73
10 25
11 14
- 60
- 21
4 18
313
- 78
- 77
5 23
6 15
- 78
- 60
5NA
7NA
-
-
25
29
127
70
- 80
- 59
KINGS COUNTY '
QUEENS COUNTY'
1 NA
3 NA --
17 16
4 15
•
-
6
73
16 20
2 8
-
-
20
75
13 12
6 8
•- 8
25
13 9
5 6
• 44
- 17
8 8
1 4
0
- 75
4 NA
2 NA
-- 67
18
65
41
+ 3
- 56
STATEN ISLAND' 0 NA 0 1 - 2 1 + 100 1 0 - 0 1 - 0 2 - 0 NA 3 5 - 40
(NEW YORK CITY TOTAL)' 17 NA - 32 83 - 61 24 67 - 64 41 59 - 31 25 47 - 47 20 52 - 62 18 NA 142 308 - 54

SUBTOTAL 422 859 - 51

2. Less Urban and Non-Urban Areas

ALASKA'
NEBRASKA'
1 NA
2NA
-
-
1 NA
3 NA
-
-
0 NA
2 0
-
--
0 2
3 0
-
-
0 1
3 5
-
- 40
1 NA
1 NA
-- 0 NA
1 NA -
0
8
3
5 + 60
OREGON ' 2 NA 3NA - 0 NA 2 NA - 4 9 - 56 2 4 -50 2 4 - 50 8 17 - 53
SOUTH CAROLINA ' 1 NA 5 NA - 0 NA - 4 NA - 8 NA 2 NA - 2 6 - 66 2 6 - 66
VEAMONT2 1 0 - 0 0 - 0 0 0 1 - 1 1 0 ONA - 0 NA - 2 2 0
WISCONSIN' 1 NA - 1 4 - 75 0 3 3 8 - 63 3 7 - 57 4 NA - 1 NA - 7 22 - 68

SUBTOTAL 27 55 - 51

TOTAL 449 914 - 51

NA = Not Available
VS = Vital Statistics Data
PG = Police-Generated Data
1 = PG Data Reported to State Statistical Analysis Center (SAC)
2 = PG Data Reported to FBI on Supplemental Homicide Forms
3 = PG Data Compiled by New York City Police
·Years 1970-73 are taken f rom published data; 1974-76 data are taken from computer tapes
• • 1972 is a 50 percent sample
jurisdictions in Table 1 accounted for 25 percent of all deaths by
legal intervention of police for 1971-75. Yet these jurisdictions
show a combined underreporting of 51 percent during the per­
iod 1970-76 (with some years omitted in some jurisdictions). At
the very least, then, the total national incidence of police homi­
cide in that period was probably about 26 percent higher than
the NCHS data reported.
It is always dangerous to generalize from a nonrandom
sample, even when 1) the sample constitutes one-fourth of the
count obtained from the entire universe; 2) the bias in much of
the sample (California and New York) seems to be toward more
professional (and perhaps more complete) reporting by coroners
and medical examiners; and 3) the sample shows similar under­
reporting rates for both heavily urban and less urban areas. If
such a generalization were made, however, the 51 percent un­
derreporting rate applied nationwide would yield an estimate of
3,673 police homicides throughout the country during the per­
iod 1971-75. Dividing this estimate of police homicides by the
total of 101,665 homicides from all causes throughout the coun­
try during that period39 shows that the police may be responsi­
ble for 3.61 percent of all homicides-about one out of every
twenty-eight. In New York City alone, the figure was even
higher: 3.7 percent; in California, it was higher still at 4.18
percent.
Yet it must be stressed that generalizing the underreport­
ing rate is a suspect procedure. There is no way of being cer­
tain that death certificates for police homicide are reported as
incompletely in the majority of jurisdictions for which we were
unable to obtain police-generated data. Moreover, as Table 2
shows, in some big cities the NCHS figures exceed those de ­
rived from other sources. Regional variations in rates of both
homicide and police homicide further complicate the procedure,
as well as the possibility that regular homicides are also under­
reported at varying rates.
What Table 1 does suggest is that the NCHS data cannot
be used to measure the national incidence of homicide by police
officers. Since the police-generated data do not encompass the
entire nation, it is safe to say that this country simply does not
23
know how many of its own citizens it kills each year under the
authority of the state.
MEASURES OF RELATIVE INCIDENCE ACROSS CITIES
The second question facing the available data sources is
whether they can be used to measure the relative incidence of
police homicide from one police department or city to another.
The method used here to evaluate the NCHS data for this pur­
pose is to compare those data on decedent's city of residence
(not place of death)-the only form in which city level data are
available-to data obtained from a variety of alternate sources
(primarily but not only police-generated data) on the number of
people killed by police in each city (place of death) or by the
city's main police department (agency responsible for death). 40
There are four sources of error, then, built into this compari­
son: the place of residence may differ from place of death, place
of death may differ from agency responsible for death, place of
residence may differ from agency responsible for death, and
alternate data sources vary across cities and also change from
year to year within cities. In some years up to three different
figures from alternate data sources are averaged to obtain the
comparison figure reported in Table 2. Given this mixture of
definitions and types of data, the level of agreement for each
city in each year is surprisingly high.
Both NCHS and alternate data were obtained for a total of
133 city-years from thirty-six jurisdictions of over 250,000 pop­
ulation (counting New York's five boroughs separately). The
raw death counts provided by the two sources of data show a
substantial positive association (r = .64, r 2 = .41, r, = .62,
annual data not displayed). When the death counts are stand­
ardized by population, the strength of the correlations is re­
duced somewhat but the Pearson' s coefficient remains substan­
tial (r = .53, r = .28, r. = .38, annual data not displayed).
Computations omitting California cities and 1972 data (in which
year the NCHS based its statistics on only a 50 percent sample)
show insignificant differences from the computations using all
133 city-years. None of the correlations, however, account for
even half of the variance.
24
TABLE 2 MEAN ANNUAL DEATHS AND DEATH RATES FROM HOMICIDE
BY POLICE OFFICERS BASED ON VITAL STATISTICS AND
ALTERNATE DATA IN 36 JURISDICTIONS FOR VARIOUS YEARS
FROM 1966 TO 1976.
MEAN DEATHS
PER RATIO OF
NO. OF YEARS PER ANNUM 1oo.ooo pa· MEAN DEATHS
PER ANNUM PER ANNUM
CITY COMPARED vs· A"" vs A AIVS

1. ATLANTA 4 6.25 10.50 1.41 2.37 1.68

2. BALTIMORE 2 3.00 8.00 0.34 0.91 2.67
3. BIRMINGHAM 5 1.80 6.00 0.63 2.10 3.33
4. BOSTON 2 2.00 2.50 0.32 0.40 1.25
5. CHICAGO 7 9.29 33.00 0.29 1.03 3.55
6. CLEVELAND 2 12.50 10.50 1.84 1.55 0.84
7. COLUMBUS 2 2.50 2.00 0.46 0.37 0.80
8. DALLAS 2 10.50 7.50 1.29 0.92 0.71
9. DENVER 2 1.00 4.00 0.19 0 .78 4.00
10. DETROIT 3 15.67 29.67 1.13 2.14 1.89
11 . DISTRICT OF
COLUMBIA 3 4.67 10.67 0.64 1.45
12. HONOLULU 2 0.00 0.50 0.00 0.07
13. HOUSTON 2 0.50 15.00 0.04 1. 14 30.00
14. INDIANAPOLIS 3 7.00 4.00 0.96 0 .55 0.57
15. JACKSONVILLE 2 0.50 5.50 0.10 1.05 11 .00
16. KANSAS CITY, MO. 3 0.00 3.33 0.00 0.94 0.00
17. LONG BEACH 4 0.50 1.75 0.14 0.50 3.50
18. LOS ANGELES 4 7.50 21.25 0.27 0.76 2.83
19. MEMPHIS 3 0.13 5.25 0.02 0.81 40.39
20. MILWAUKEE 2 1.00 3.00 0.14 0.43 3.00
21. OAKLAND 5 1.40 2.00 0.40 0.57 1.43
22. PHILADELPHIA 11 9.45 14.18 0.50 0.74 1.50
23. PHOENIX 2 0.00 1.50 0.00 0.24
24. PORTLAND 4 0.75 1.00 0 .20 0.27 1.33
25. SAN ANTONIO 2 1.00 3.00 0 .13 0.40 3.00
26. SAN DIEGO 4 2.00 1.25 0.28 0.17 0.63
27. SAN FRANCISCO 4 2.25 3.25 0 .32 0.46 1.44
28. SAN JOSE 4 2.00 1.50 0.41 0.31 0.75
29. SEATILE 2 1.00 3.50 0.20 0.70 3.50
30. ST. LOUIS 2 4.00 6.50 0 .72 1.16 1.63
31. SACRAMENTO 4 3.00 2.00 1.1 4 0.76 0.67
32. NEW YORK COUNTY 5 5.00 24.80 0.34 1.70 4.96
33. BRONX COUNTY 5 5.80 14.00 0 .41 0.98 2.41
34. KINGS COUNTY 5 13.40 13.00 0.54 0.52 0.97
35. QUEENS COUNTY 5 3.60 8 .20 0. 18 0.42 2.28
36. STATEN ISLAND 5 0.60 1.00 0.19 0.32 1.67
(New York City Total) ' (5) (28.40) (61 .60) (0.37) (0.80) (2.1 7)

·vs = Vital Statist ics z = 3.59 r = .69 r = .56

.. A = Alternate Source of
Data s = 1.98 r2 = .48 r' = .3 1 6 = 1.66
r, = .67 r, = .44
n = 36 .50 r = .31 n = 36
·Not included is calculat ion of statistics n = 36 n = 36

25
 Given the extreme rarity of police homicide events, much of
the variation of both data sources for the 133 city-years is prob­
ably mere year-to-year statistical instability. As Table 2 shows,
none of the jurisdictions in the computations exceeded a mean
vital statistics count of sixteen deaths per year, and only four
jurisdictions exceeded that level using the alternate data
sources. Consequently, both the death counts and death rates
per 100,000 are much more stable when their mean levels for
all available years (from two to eleven years per city) are em­
ployed (Table 2). This procedure increases the correlations be­
tween both the death counts and the death rates per 100,000
population provided by the two data sets, but it still leaves over
one-half of the variance to be accounted for.
Table 2 also shows that the absolute differences between
the NCHS data and the alternate sources are much higher at
the city level, or at least in certain cities, than at the state level.
Contrary to the ratio of two-to-one found in Table 1, the mean
ratio of NCHS to alternate data for the cities in Table 2 is
almost four-to-one. This ratio, however , is heavily influenced by
two outlier cases, Houston and Memphis, and is moreover inap­
propriate to calculate since there are three values of infinity in
the data set. A least squares estimate, however, is appropriate,
and it yields a b of 1.66, which is slightly lower than the two-to­
one ratio found in Table 2. Because of the differences in place
of death versus place of residence present in the city level data,
however, it is questionable whether this ratio or the one derived
from Table 1 (which relies much less on city level data) is more
appropriate.
On the other hand, eight cities (including Brooklyn) in Table
2 show higher death counts with NCHS data than with the
alternate data sources. In three of the eight, the alternate
sources of data contain the Uniform Crime Report supplemen­
tary homicide reports, for which certain cities fail to complete
the section describing the circumstances of the homicide. What­
ever the reason, the fact that the alternate data do not produce
consistently higher death counts prevents any conclusion that
the alternate data provide a "better" measure of the relative
incidence of police homicides across cities.

26
 Rather, two conclusions about measuring relative incidence
are suggested by Table 2. One is that while the two data sets
show roughly the same patterns of relative incidence, several
cities, especially Kansas City, Houston, and Memphis, show
radical disagreement. Figure 1 illustrates both the general simi­
larity of the overall pattern and the wide discrepancies in par­
ticular cases. The second conclusion is, therefore, that although
the data are not accurate enough to be used to compare one
specific city to another, either data set might be appropriate to
use as a measure of the pattern of variation in police homicide
rates in relation to independent variables that might explain
that variation. And as Figure 1 demonstrates, there is a great
deal of variation to be explained.

MEASURES OF PATTERNED VARIATION ACROSS

CITIES
The third question facing the available data sources is
whether they provide a reliable measure of patterns of varia­
tion. This question is tentatively answered by the interpreta­
tions of Table 2 and Figure 1. Since both data sources show
roughly the same patterns, it appears more likely that each of
them is reliably measuring the true patterns. 41 The level of
agreement on the patterns between the two data sources is low
enough, however, so that further empirical comparisons are
necessary. Another method of comparison is to examine the
correlations of the rates produced by the two data sources with
the same theoretically relevant independent variables. If the
data sources are both approximating the same patterns, then
the rates they produce should yield similar correlations with the
independent variables.
Table 3 displays the correlations of the rates from the two
data sources with seventeen independent variables from three
separate theoretical domains. Many other variables and theo­
retical domains offer possible explanations for variations in po­
lice homicide rates; these variables are offered only as illustra­
tions. A substantive analysis of these correlations is beyond the

27
 FIGURE 1. VITAL STATISTICS AND ALTERNATE DATA MEAN
ANNUAL RATES OF HOMICIDE BY POLICE
OFFICERS PER 100,000 POPULATION IN 32
CITIES .
CITY 0 0.5 10 15 20 25

1 ATLANTA
2 DETROIT

3 BIRMINGHAM

4 CLEVELAND

5 DISTRICT OF COLUMBIA

6 ST. LOUIS

7 HOUSTON

8 JACKSONVILLE

9 CHICAGO
--
, _.
10 KANSAS CITY
11 DALlAS
12 BALTIMORE ~
13 MEMPHIS ,..
14 NEW YORK CITY ~
15 DENVER ~
16 SACRAMENTO
17 LOS ANGELES ~
18 PHILADELPHIA
19 SEATTLE
20 OAKLAND
.....
21 INDIANAPOLIS
22 LONG BEACH ,...
23 SAN FRANCISCO ~
24 MILWAUKEE ~
,._

­
25 BOSTON KEY:
26 SAN ANTONIO ,.. A ALTERNATE DATA
27 COLUMBUS VS VITAL STATISTICS
1---1
28 SAN JOSE
29 PORTLAND ~
~
30 PHOENIX
31 SAN DIEGO
32 HONOLULU p
0 0.5 1.0 1.5 2.0 2.5

28
scope of this article, but the theoretical rationales for and pre­
dicted directions of the relationships with the independent vari­
ables can be briefly summarized.
Among community characteristics, it can be predicted that
population density would be positively related to police homi­
cide rates, both because it is related to other kinds of violence
which might prompt police homicide and because shots fired in
denser areas are more likely to hit someone (whether or not the
person hit is the intended target). Gun density would be ex­
pected to be positively related to police homicide because
greater gun density should increase the frequency of defense of
life situations in which police homicides occur. Unemployment
and suicide, as measures of declining social cohesion, would be
expected to be positively related to police homicide, given the
theory that governmental social control increases as social co­
hension declines. It would be expected that the violent index
crime rate and the homicide rate would be positively related to
the police homicide rate because these rates may increase police
perceptions of danger in their work and make them more prone
to use violence as a possible preemption of attacks on them.
The police per 1,000 population ratio and the violent arrest rate
both should be positively related to police homicide, since both
of those variables provide an increased risk or exposure of citi­
zens to police use of deadly force.
Among police organizational structure variables, it can be
predicted that geographic decentralization (precincts per
square mile) would be negatively related with police homicide,
since it is associated with a "watchman" 42 style of low-level Jaw
enforcement. Administrative intensity (percentage of personnel
in support units), span of control (number of supervisors per
line officer), differentiation (percentage of all personnel in other
units than basic patrol), and self-regulation (percentage of per­
sonnel assigned to internal investigations) would be predicted
to be positively related to the police homicide rate, since all are
rough measures of bureaucratization and more bureaucratized
law enforcement agencies may be expected to mete out more
legal sanctions of all forms, including killings."1

29
<:>:> TABLE 3 A COMPARISON OF CORRELATIONS OF SELECTED VAR IABLES WITH NCHS AND
0
POLICE-GENERATED POLICE HOMICIDE RATES FOR SELECTED YEARS AND CITIES
Correlatfon ( r) Wtlh

Matched C ities and Years . Selected Years , 1967-76

(2)
(1 ) Annual Mean Alternate
Annual Mean NCHS Data Source Rate o l (4)
Rate of Pohce Pohce Homteides (3) Alternattve Data
lndependenl Predocled HomiCides Per Per 100.000 Annual NCHS Rate Source Annual Rate
Vanable Sogn 100.000 Population Populat•on of Pollee Hom•c•des of Pollee HomiCides
1974·1976 (N) 1974-1976 (N) Per 100.000 Pop (N) Per 100.000 Pop (N)

Commumty CharacteristJCS
1. Populat•on Oens•ty + +.08 (48) + 04 (20) + 20 (27) + ' 15 (27)
2 Gun Oens•ty + + 28" (48) + 45 ' (16) + .25 (27) + .54 . • (27)
3 Unemployment Rate + -.20 (47) - 25 ( 17) - 22 (27) - .3s · (27)
4. Su•c•de Rate + - .04 (48) - 39 (\6) - 1\ (27) - .23 (27)
5 Vtolent Index Cnme Rate + + 37"' (48) + 47" (20) + 40' (27) + ss· • • (27)
6 UCR Hom1C1de Rate + + ss· • • (48) + 72 ... (20) + 60 ... (27) + 84 .•. ( 27)
7 Pohce Per 1.000
PopulatiOn + +AS ' ' ' (44) +so ··· (18) + 3\ (24) • ss· • (24) i
8 V1otent Arrest Rate + + 31' ( 45) + 42' (\9) + 19 (26) + 36" (26 )

PoliCe Orgamzational Structure

9 GeographiC - - 02 (47) + 08 (19) +09 (27) + 12 (27)
Oecentral•zat•on + + \2 (42) - 31 (\7) 00 (22) + 08 (22)
10 Adm1nistrat1ve Intensity + - \0 (43) + 32 (18) + 02 (23) + 2S (23)
11 . Span ol Control + - .23 (42) - 39 ( 18) -.33 (23) - 48". (23)
12 c,v,hzahon + -. \3 (43) -. so· (18) + \2 (23) +OS (23)
13. 01fferenllat10n + + .3s· (41) + .08 (17) + .37 ' (23) + 36. (23)
14 Seii-Aegulalion
Police Orgamzarronal Polic;es
15 Arrests (all olfenses) per + + 29. (4S) + .40 ' (19) + 31 (26) + 43" ( 26)
100.000 populahon - + \0 (37) + \3 (lS) + 01 (19) +06 p 9)
16 DISCiplinary FormaliSm - -.12 (36) + \9 (14) + 02 (18) - 09 (18) I
17 01SC1phnary Pressure
I
Key
• S1Qn1flcant al 05 level
• S1gn,lteant at 01 level
• • • • S1gn,hcant at .001 level
 Among police organizational policies, it can be predicted
that the overall arrest rate would be positively related to the
police homicide rate because every arrest provides an opportu­
nity for the arrestee to resist arrest, a response that could lead
to a police homicide. One would expect both disciplinary formal­
ism (the percentage of all complaints of police misconduct that
are investigated) and disciplinary pressure (the percentage of
the complaints investigated that are substantiated) to be nega­
tively related to police homicide rates, since police officers may
be less likely to kill citizens where they perceive a greater risk
of punishment for misconduct in situations where the justifiabil­
ity of the homicide may be unclear or subject to conflicting
opinions.
Table 3 presents the correlations of the 1976 values44 of
most of the independent variables45 with two sets of the mea­
sures of police homicide rates. One set (columns 1-3) is selected
to match the data for the independent variables as closely in
time as possible, although it used a three-year (1974-76) annual
mean rather than a one-year (1976 only) figure to stabilize year­
to-year fluctuations. The consequences of that selection proce­
dure, however, are to reduce the number of cities for which
alternate data are available and to bias the alternate data sam­
ple heavily in favor of California. While the data for all of the
fifty-three cities over 250,000 population for which data on the
independent variables were available (up to forty-eight are used
to compute the correlations reported in column 1 of Table 3),
only data from a maximum of twenty (ten of which are in Cali­
fornia) out of the 157 cities over 100,000 population were avail­
able to compute the correlations reported in column 2. It is not
possible to compute the correlations of the independent varia­
bles with NCHS data using only the cities included in the calcu­
lation of the same correlations using alternate data sources
since up to half of the cities included in the alternate data
sources have less than 250,000 population and NCHS figures
are therefore unavailable.
The second set of the measures of police homicide rates
(columns 3 and 4) provides direct comparisons of the two when

31
correlated with independent variables matched by both year
and city for all of the nonrandom convenience sample for which
alternate data on police homicide rates were available through
the period 1967-76. The consequence of this procedure is to
move the data on police homicide rates further away in time
from the data on the independent variables, and inconsistently
so from one city to the next. Whatever problems this conse­
quence may pose for a substantive analysis, however, it should
not be a great hindrance to achieving the present objective of
determining whether the two data sources yield similar correla­
tions.
Considering the diversity of procedures employed, the
results are remarkably consistent. Overall, the two data
sources and data sets tend to yield similar results, with the
alternate data sources producing stronger correlations and with
most of the correlations in the theoretically predicted direc­
tions. When "agreement" is defined as the correlations from
both data sets showing the same sign and statistical signifi­
cance at at least the .05 level, or showing both correlations as
not significant regardless of sign, then a comparison of columns
1 and 2 shows 88 percent agreement, and a comparison of
columns 3 and 4 shows 65 percent agreement. The level of
agreement between columns 3 and 4 would have been even
higher if the significance level of the correlations of variables
twelve and fifteen with the NCHS data (column 3) had been
greater by 11100th. The generally high level of agreement sug­
gests that both data sources are indeed tapping similar patterns
of variation.
The alternate data sources produce stronger correlations
more than three times out of four. When only statistically sig­
nificant correlations are compared, the correlations using alter­
nate data sources reported in column 2 are larger than those
using NCHS data reported in column 1 for 87.5 percent of the
comparisons; when the nonsignificant correlations are included,
the results are almost identical (88.2 percent higher). For
columns 3 and 4, the correlations with alternate data sources
are greater for 89 percent of the comparisons of statistically

32
significant correlations and for 77 percent of the comparisons
of all correlations.
Contrary to the findings on the strength of the correlations,
however, the alternate data sources do not produce correlations
that are more often in the theoretically predicted direction. For
the significant correlations, 100 percent of those derived from
NCHS data are in the predicted direction, while only 86 percent
of those in column 2 and 78 percent of those in column 4 (de­
rived from alternate data sources) are as predicted. For all
correlations, however, the findings are mixed: a higher percent­
age of NCHS correlations are as predicted in the first data set
(65 percent of column 1 versus 53 percent of column 2), while a
higher percentage of the alternate data source correlations are
as predicted in the second data set (71 percent in column 4
versus 59 percent in column 3).
The modest degree of overall success of the theoretical pre­
dictions (for 89 percent of all the significant correlations and 62
percent of all the correlations), whatever it may say about the
adequacy of the theories, lends further support to the conclu­
sion that the data sources measure similar patterns and per­
haps that they both measure an actual pattern of variation.
To the extent that these procedures are able to answer the
question of whether available data sources adequately measure
patterns of variation in police homicide rates, then, the answer
seems to be affirmative. Other independent variables, of
course, might have been selected that possibly could produce
different results. On the basis of the correlations with the varia­
bles that were selected, however, the similarity of results be­
tween the two data sources suggests that either source might
be appropriate for cross-city analysis of patterns. Since the
NCHS data are consistently available (though, unfortunately,
only on tape) for all large cities since 1967, this finding is par­
ticularly important since it means that at least one complete
data set on police homicide can be matched by year to the
corresponding data on independent variables. The matching
will allow multivariate analysis and other more sophisticated
analytic approaches.

33
CONCLUSIONS
This analysis provides tentative answers to three central
questions about the adequacy of current measurement of homi­
cide by police officers in the United States. First, it suggests
that the national incidence of police homicide is substantially
underreported, possibly by around 50 percent and that the po­
lice may account for closer to 3.6 percent of all homicides
rather than to 1.8 percent, as previously had been reported.
Even these figures, however, are largely speculative. In any
case, the analysis strongly suggests that there is no adequate
basis for arriving at accurate national estimates of the number
of citizens killed by police officers each year.
Second, the analysis reveals many instances in which the
Vital Statistics data and data from other sources on the num­
ber of police homicides in specific cities are in substantial dis­
agreement. This finding suggests the conclusion that none of
the available data sources should be used to compare police
homicide rates from one particular city to another. Since in any
particular city there is a substantial likelihood that the number
of police homicides derived from any one data source is in er­
ror, comparisons of specific cities are likely to be dangerously
misleading.
The most encouraging finding of this analysis is its answer
to the third question. Judging from the similarity of the correla­
tions of police homicide rates derived from the Vital Statistics
and from alternate data sources with theoretically relevant in­
dependent variables, both of these data sources seem to be
producing the same total patterns of variation across cities.
Either data source may therefore be appropriate to use for
correlational analysis of the factors associated with those pat­
terns. While the alternate data sources tend to produce
stronger correlations, they are only available on a haphazard
basis. The Vital Statistics data may therefore be preferable for
those analytic purposes for which a more complete data set is
required.
These three conclusions place previous research on police
homicides in a new light. The conclusions of those studies that

34
 have employed NCHS data on the national incidence and trends
of police homicide46 should be reevaluated now and treated with
great caution. Similarly, those studies of police homicide that
have made specific comparisons from one city to another or
among a small group of cities also should be used with great
caution .47 Those studies that have focused solely on correla­
tional analysis, however,~ 8 now can be viewed with greater con­
fidence from the standpoint of measurement, whatever the the­
oretical quality of the analysis.
The overwhelming implication of this analysis is that our
present procedures for measuring homicide by police officers
should be improved. Since both the Uniform Crime Report sys­
tem and the Vital Statistics system are voluntary, there may be
little that can be done with them to improve our national meas­
urement of the absolute incidence of these events. Reporting
systems at the state level, required by state law/9 however,
show a great deal of promise, and would probably be the best
long-term way to improve the measurement of both the abso­
lute and relative incidence of public homicide across states and
cities.
Finally, it is worth noting the irony in this analysis: while
the police may have the most to gain by undercounting the
number of citizens they kill and while it is true that many police
departments fail to undertake any count at all, it is the police
that have provided the largest figures on the numbers of citi­
zens killed. For whatever reason s, the source of the under­
counting of police homicides is not the police, but rather the
local medico-legal officers and the national system of vital sta­
tistics. If any general fault or blame is to be assessed on any
group for the demonstrably shoddy state of the official meas­
urement of police homicides, the medico-legal officers may be a
more appropriate target than the police.
Rather than assessing blame, however, a more useful re­
sponse would be for all institutions concerned to improve the
quality of their data. If the National Center for Health Statis­
tics, the United States Public Health Service, and the American
Association for Vital Records and Public Health Statistics re­
vised the Standard Death Certificate to include a check box for
35
police homicide; if the National Center for Health Statistics
compiled mortality data by city of occurrence rather than by
decedent's city of residence; if police departments published in
their annual reports the number of citizens they killed each
year; if the Uniform Crime Statistics published the numbers of
citizens killed as reported in the supplemental homicide reports
supplied by local police departments; and if all state legislatures
required local police departments to file a report with a state
agency whenever a citizen is killed, it would be much easier to
monitor trends and differences in the use and possible abuse of
police power. Since some democracies require a written report
to the national government every time a police officer draws a
weapon, 50 there improvements in the American system for re­
porting the taking of life would appear feasible.

This research was supported by Grant No .

IR01MH31335-01CD awarded to the Criminal Jus­
tice Research Center, Inc., Albany, N.Y., by the Cen­
ter for Studies in Crime and Delinquency, National
Institute of Mental Health. We wish to thank Paul
Zolbe of the FBI for supplying us with arrest data;
David Christianson and Mark Blumberg for assisting
in assemblin_g some of the data sources; and Dr.
Richard Staufenberger of the Police Foundation, Dr.
James Fyfe of the New York City Police Depart­
ment, and Dr. Arthur Kobler for making available
for secondary analysis some of the data reported.
We also thank the National Center for Health Statis­
tics for providing the city-level data reported here .
James Nelson, Michael Gottfredson, Herman Gold­
stein, Albert J. Reiss, Jr., Marshall W. Meyer, Mi­
chael J. Buckman, and Kenneth Adams provided
helpful comments on an earlier draft.

Notes
Q. Rev.
1. See Sellin , The Significance of Records of Crime, 67 Law
489,494 (1951); Wolfgang, A Sociological Analysis ofCriminalHomi­
cide, in Crime in America 53 (B. Cohen ed. 1970) (both are cited in

36
S.F. Messner, Income Inequality and Murder Rates; Some Cross·
National Findings (1978) (paper presented to the 73d Annual Meeting
of the American Sociological Association)).
2. See Hindelang, The Uniform Crime Reports Revisited, 2 J. Crim.
Just. 1 (1974). But see Cantor and Cohen, Comparative Measures of
Homicide Trends: Methodological and Substantive Differences in The
Vital Statistics and Uniform Crime Report Time Series (1933-75).
Working paper 7821, Program in Applied Social Statistics, Depart·
ment of Sociology, University of Illinois at Urbana-Champaign.
3. In a 1969 survey, for example, 57 percent of a national sample said
that "police shooting looters" was not an act of violence. M. Blumen·
thai, L. Chadiha, G. Cole & T. Jayartne, Justifying Violence 73 (1972),
cited in Archer and Gartner, Legal Homicide and Its Consequences, in
Violence: Perspectives on Murder and Aggression 221 (Kutash ed.
1978). Archer and Gartner also cite Professor Short's account of how
the research staff of the National Commission on the Causes and Pre­
vention of Violence, which had originally defined the scope of study
neutrally to include all uses of force , including police killings, was
influenced to narrow the scope of study to all "illegal violence"­
thereby excluding most governmental use of force. Archer and
Gartner, supra at 222-23. See also Short, The National Commission on
the Causes and Prevention of Violence: Reflections on the Contributions
ofSociology and Sociologists, in Sociology and Public Policy: The Case
of Presidential Commissions (Komarovsky ed. 1975).
4. See Takagi, A Garrison State in "Democratic" Society, in Police
Community-Relations 358-71 (Cohn and Viano eds. 1976).
5. See Sherman, Restricting the License to Kill: Recent Developments
In Police Use of Deadly Force, 14 Crim. L. Bull. 577 (1978); U.S.
Commission on Civil Rights, Police Practices and the Preservation of
Civil Rights: A Consultation. (Dec. 12-13, 1978, Washington, D.C.).
See also Gilman, In Washington, A New Zeal For Prosecuting Police,
Police Magazine, November 1978, at 18.
6. See Day, Shooting the Fleeing Felon: State of the Law, 14 Crim. L.
Bull. 285 (1978); De Roma, Justifiable Use of Deadly Force by the
Police: A Statutory Survey, 12 Wm. & Mary L. Rev. 67 (1970); Finch,
Deadly Force To Arrest: Triggering Constitutional Review, 11 Harv.
C.R.·C.L.. Rev. 361 (1976); Mayhall, Use ofDeadly Force in the Arrest
Process, 31 La. L. Rev. 131 (1970); Zittler, Policeman's Use of Deadly
Force in Illinois, 48 Chi.-Kent L. Rev. 252 [1971).
7. See Harding & Fahey, Killings By Chicago Police, 1969·70; An Em­
pirical Study, 46 S. Cal. L. Rev. 284 (1973); Jacobs and Britt, Inequal­
ity And Police Use of Deadly Force: An Empirical Assessment Of A
Conflict Hypothesis, 25 Soc. Prob. 403 (1979); Kania and Mackey, Po­

37
lice Violence as a Function of Community Characteristics, 15 Crimi­
nology 27 (1977); Kobler, Police Homicide In A Democracy, 31 J. Soc.
Issues 163 (1975); Takagi, note 4 supra; Robin, Justifiable Homicide
by Police Officers, 54 J. Crim. L. C. & P. S. 225 (1963); Delman,
Varieties of Police Policy; A Study of Police Policy Regarding the Use
of Deadly Force in Los Angeles County, 6 Loy. L.A.L. Rev. 1 (1973);
Fyfe, Shots Fired: A Typological Examination of New York City Po­
lice Firearms Discharges (1978). Unpublished Ph.D. dissertation,
State University of New York at Albany; Milton, Police Use of Deadly
Force (1977). Washington, D.C.: Police Foundation.
8. During this period, there were 1,800 deaths attributed to Jaw en­
forcement officers included in the 101,665 homicides from all causes.
Vital Statistics of the United States (1965-1974).
9. See Biderman and Reiss, On Exploring The "Dark Figure" of
Crime, in 374 The Annals of the American Academy of Political and
Social Science 1 (1967).
10. Many police agencies fail to provide some or all of the descriptive
information on those forms that is necessary to discriminate justifiable
homicides by police from other forms of homicide. Interview with Paul
M. Zolbe, Chief, Uniform Crime Reporting Section, FBI (July 5, 1978).
11. See, e.g., Goldkamp, Minorities as Victims of Police Shootings:
Interpretation of Racial Disproportionality and Police Use of Deadly
Force, 2 Just. Sys. J. 169 (1976); Jacobs and Britt, note 7 supra; Kania
and Mackey, note 7 supra; Kobler, note 7 supra; Takagi, note 4 supra;
Milton, note 7 supra.
12. National Center for Health Statistics, International Classification
of Diseases, Adapted for Use in the United States 501 (8th rev. 1967).
13. 2 Vital Statistics of the United States (Part A) 6-9 (1973).
14. See I. Wayne, Suicide Statistics in the United States: An Explora­
tion of Some Factors Affecting the Quality of Data (1969). Terminal
Progress Report MH-15104, cited in Bradshaw, The Social Construc­
tion of Suicide Rates 52 (1973), unpublished Ph.D. Dissertation, De­
partment of Sociology, Syracuse University.
15. For a description of the system, see National Center for Health
Statistics, Medical Examiners' and Coroners' Handbook on Death and
Fetal Death Registration (1971).
16. See Moriyama, Baum, Haenszel and Mattison, Inquiry Into Diag­
nostic Evidence Supporting Medical Certifications of Death, 48 Am. J.
Pub. Health (1376-87 (1958).
17. See James, Patton and Heslin, Accuracy of Cause of Death State­
ments on Death Certificates, 70 Pub. Health Rep. 39-51 (1955).

38
18. Telephone Interview with Michael Baden, M.D., then chief medical
examiner of New York City (July 17, 1978).
19. Interview with Sydney Katz, M.D. (December 19, 1978).
20. See National Center for Health Statistics, Model State Vital Sta­
tistics Act and Model State Vital Statistics Regulations, 1977, at 78­
115 (1978).
21. !d. at 8.
22. !d. at 18.
23. For example, almost 17 percent of the 1973 death certificates
reporting that an autopsy had been performed failed to complete a
simple yes-no question about the autopsy. 2 Vital Statistics of the
United States (Part A) 6-18 (1973).
24. Even when the police officer's role is described, there may be
insufficient information to discriminate between legal and illegal
actions of the police. While the ICD definition of this cause of death
implies that the death certificate is filled out after t he proper officials
have determined whether or not a police homicide was justified, in
practice that is probably not the case. The necessary review proce­
dures can go on for months after a killing, but the death certificate
typically must be completed before a burial is possible. Since burials
usually occur within a week after a death, it seems virtually impossible
for a death certificate to be based on a final ruling on the justifiability
of the death. If a police officer is convicted of murder for an on-duty
homicide a year after the fact, there seems to be no provision in the
vital statistics system for changing the cause of death from legal inter­
vention to homicide. Since officers are convicted so rarely for on-duty
murder, however, this issue may have little impact on the system's
data. Kobler found that only 3 of 1,500 officers in his sample of police
killings were convicted on criminal charges related to the killing.
Kobler, supra note 7, at 164. The first conviction of an officer for on­
duty criminal homicide charges in the history of the New York City
Police (since 1844) did not occur until the mid-1970s. See Hoffman, The
Man Who Defends Killer Cops, 10 N.Y. Magazine 76 (1977).
25. Bradshaw, supra note 14, at 53. What is true for rural coroners
may also be true for big city medical examiners as well. One line of
speculation over the reason for the dismissal of New York City Medi­
cal Examiner Michael Baden, for example, was that he had failed to be
sufficiently responsive to the wishes of the New York County prosecu­
tor. See Baden Planning to Sue the City Over His Ouster, N.Y. Times,
Aug. 8, 1979, at B4.
26. Katz interview, note 19 supra.
27. Baden interview, note 18 supra.

39
28. 2 Vital Statistics of the United States (part A) 6·19 (1973).
29. See text accompanying note 12 supra.
30. Such a situation would not be coded as death by legal intervention.
Letter from Harry Rosenberg, Chief, Mortality Branch, National Cen·
ter for Health Statistics (July 27, 1978).
31. NCHS Micro-data Detail Tape (Mortality): File figures, New York
City Department of Health.
32. L. Sherman, Scandal and Reform: Controlling Police Corruption
146·49 (1978).
33. In Fort Lupton, Colorado, for example, a 10-officer police depart­
ment shot and killed three citizens in one year. See generally Greeley
Tribune, 1977-78.
34. See, e.g., Milton, note 7 supra.
35. Fyfe, supra note 7, at 516 n.5.
36. Even the same exact count from both data sources, however, pro­
vides no assurance that all police homicides have been counted. The
following table shows how a total count of 25 deaths in one year found
in both the police records and the newspapers could be found when the
actual number of deaths was 50. A procedure that recorded the names
of the victims could capture the deaths in table cells b and c, thereby
raising the total number of deaths counted to 35. But the cases in cell
d would go unnoticed, by definition, using these two data sources, as
they would in the comparisons of two data sources made in Tables 1
and 2.
Reported in Newspapers
Reported In
Police Files
Yes No Total
Yes a) 15 b) 10 25
No ~ 10 d) 15 25
Total 25 25 50
37. See Kobler, note 7 supra.
38. See note 10 supra.
39. Vital Statistics of the United States, Annual1971-75.
40. A list of the alternate data sources used in each city for each year
is available from the authors at One Alton Road, Albany, N.Y. 12203.
41. See E. Webb, D. Campbell, R. Schwartz and L. Sechrest, Unobtru­
sive Measures (1966).

40
42. See J. Wilson, Varieties of Police Behavior (1968).
43. See D. Black, The Behavior of Law (1976).
44. The table does not include gun density and population density,
which use a 1974-76 average and 1975 data, respectively. The only
year for which data on most of the independent variables were availa­
ble was 1976.
45. Variables 1, 7, 9-14, 16, and 17 were computed from Police Foun­
dation, Police Practices: The General Administrative Survey (1978)
and Police Executive Research Forum, Survey of Police Operational
and Administrative Practices, 1977 (1978); variable 2 was computed
from data tapes supplied by the National Center for Health Statistics
using the procedure suggested by Cook, The Effect ofGun Availability
on Robbery and Robbery Murder: A Cross-Section Study of 50 Cities, 3
Pol'y Rev. Ann. 743 (1979) and is composed of the average of the
proportion of suicides committed with a gun and the proportion of
homicides committed with a gun; variable 3 was computed by averag­
ing the percent of the workforce unemployed reported in the U.S.
Bureau of Labor Statistics area trends in January, April, June, Sep­
tember and December 1976; variable 4 was computed from data tapes
supplied by the National Center for Health Statistics; variables 5 and 6
were computed from FBI Crime in the United States 1976 (1977); and
variables 8 and 15 were computed from data supplied on tape by the
Federal Bureau of Investigation. All population figures used to com­
pute 1976 rates were derived from Police Foundation, supra, and
Police Executive Research Forum, supra. A complete list of the cities
and years included for each correlation is available from the authors.
46. See, e.g., Goldkamp, note 11 supra; Takagi, note 4 supra.
47. See, e.g., Harding and Fahey, note 7 supra; Milton, note 7 supra.
48. See, e.g., Jacobs and Britt, note 7 supra; Kania and Mackey, note 7
supra; Delman, note 7 supra.
49. See, e.g., Minn. Stat. §626.533(2) (1976), which requires that a
report of all firearms discharges by police officers in the line of duty be
filed with the Minnesota Department of Public Safety .
50. See Baun, The Danish Police System, 1 Police Stud. 53 (1978).

41
 CHAPTER 2
ANALYSIS OF SHOOTING INCIDENTS
CATHERINE MILTON, JEANNE WAHL HALLECK,
JAMES LARDNER, AND GARY L. ABRECHT

This chapter contains observations about 320 shooting inci­

dents from 1973 and 1974, gathered from seven police depart­
ments. These data and selected characteristics of the cities and
police departments, together with some very preliminary con­
clusions, are presented to serve as a guide to police administra­
tors and to suggest areas for more extensive analysis in future
research.
While the small sample size taken from a two-year period
does not allow statistically significant conclusions to be drawn
at levels usually acceptable to behavioral scientists, some find­
ings are meaningful for policy-making officials. However, read­
ers should take care in making comparisons between depart­
ments and should recognize that the data merely characterize
the present level of shooting incidents in individual cities; the
data do not explain why or how the rates got there. Further­
more, it would be inappropriate to be critical of police adminis­
trators for substantial percentage increases in shooting rates
from one year to the next. Such changes many actually repre­
sent a very small number of incidents-such as five shootings
instead of three-or may be attributable to factors entirely be­
yond a chief's control.

SOURCE: Police Use of Deadly Force, Washington, D.C.: Police Foundation,

1977, pp. 13-37.

42
DATA COLLECTION
The seven cities ranged in size from Birmingham, Alabama,
with a population of slightly less than 300,000, to Detroit, Mich­
igan, with a population of almost 1.4 million persons. The field
researchers 1 visited only the city police department and did not
collect information from other law enforcement agencies opera­
ting in a jurisdiction, such as transit or housing authority po­
lice. The researchers reviewed shooting incident reports, de­
partment regulations, and descriptions of procedures for the
use of firearms. In addition, they interviewed administrative
personnel and spoke informally, while riding on patrol or in
other settings, with other members of the department.
Because information was available from all cities except
Detroit2 for calendar years 1973 and 1974, the staff reviewed a
total of 320 incidents3 involving the use of firearms by members
of the seven police departments over a two-year period. Only
incidents that involved shootings by police were included;
deaths or injuries of civilians by other means attributable to
police action were eliminated from consideration. Shootings by
both on- and off-duty personnel were tabulated, including per­
sonal disputes involving off-duty officers. Five incidents (all
nonfatal) involving shootings of police by their fellow officers
also have been included. Discharges that did not hit anyone,
shootings of animals, suicides, or instances in which a police
weapon was used by someone other than an officer were ex­
cluded from the study.

OBSERVATIONS ABOUT THE DATA

Observations about the data4 are organized into the follow­
ing three categories:
1. A description of the circumstances of the shooting inci­
dents and the characteristics of the citizens and police involved.
These observations include ratio of fatal to nonfatal incidents;
sex, age, and race of shooting victim; possession of a weapon by
shooting victim; type of incident in which the shooting oc­
curred; status and assignment of the officer involved; and adju­
dication of the incident by the department.

43
 2. An examination of the relationship between the shoot­
ing rates of individual cities and population size, police depart­
ment size, the Index crime rate, and the violent crime rate.
3. An examination of the relationship between the use of
fatal force by citizens against the police and by police against
citizens. For this purpose, nationwide figures have been used.
Although it is reasonable to suppose that relationships of
some kind do exist between various factors and shooting rates,
a serious problem exists in attempting to isolate each variable
to determine the nature and extent of its influence. Neither the
data collected for this study nor existing knowledge about these
matters is capable of providing that information. For example,
a change in administration or in the written policy of a depart­
ment might be followed by a reduction in shootings, but it is
extremely difficult to tie the two together. The change in shoot­
ing rate might have come about because of an unrelated revi­
sion of the department's training program or because of a sig­
nificant population shift in the community. Even to attempt
such an analysis, researchers would have to collect data far
beyond the scope of this pilot study. We hope that the prelimi­
nary findings presented here will encourage others to do just
that.

GENERAL CHARACTERISTICS OF SHOOTING

INCIDENTS, OFFICERS, AND CITIZENS

Ratio ofFatal to Nonfatal Shootings

Of the 320 shooting incidents analyzed in this chapter, close
to one-third (96) were fatal shootings. Of the total number of
shootings in all seven cities in both calendar years (378), 29
percent were fatal and 71 percent nonfatal, as Figure 1 shows.
The ratio differs from city to city in the study, ranging from
one-quarter to a little more than one-third of all incidents, as
shown in Table 1; however, the number of fatal shootings is
generally too small in individual jurisdictions and the study per­
iod too short to support a conclusion that major differences in
this ratio exist among cities. Furthermore, comparative data

44
from studies in other cities . . . show a similar ratio of fatal to
nonfatal shootings.
As noted in the introduction to this report, the decision to
include nonfatal shootings in this study was made primarily
because the number of fatal shootings was quite low in some
cities, and also because it was assumed that the distinction
between the two was frequently a matter of chance; however ,
FIGURE 1
RATIO OF FATAL TO NONFATAL SHOOTINGS: SEVEN CITIES
Number of

Shooting

Incidents

200
190
180
170
160
150
140
130
120
110
100
90
80
70 72% 69%
60
50
40
30
20
10

1973 1974
Key:
Fatal Shootings ~
Nonfatal Shootings L.,__ _ _...J
NOTE: The figure is based on the total number of shooting inci­
dents (378) occurring in the seven cities during the two­
year period; it includes a// incidents occurring during the
full year 1973 in Detroit.

45
subsequent analysis of all shooting incidents suggests that fatal
shootings are more likely to occur when subjects are armed (see
Figure 5). It should be noted, however, that the small sample
size and the uncertain validity of data concerning the presence
of a weapon make this only a tentative conclusion.
In 1973, 376 civilians were killed in the United States by
law enforcement officers, according to statistics gathered by
the National Center for Health Statistics, U.S. Public Health
Service. The total number of fatal shootings that year in the
seven sample cities, 51,> represents almost 14 percent of the
national total, even though the population of the seven cities is
only 1.9 percent of the entire population, based on 1973 Bureau
of the Census population estimates. 6
Sex (N = 314)
Nearly all of the subjects shot by the police were known to
be male (308 out of 320), and six were identified as female. In
the remaining six instances either the police report was incom­
plete, or officers reported shooting a suspect who escaped, and
was not subsequently located . In those cases, age, sex, and race
are unknown.
Age (N = 290)
The reported ages of the shooting victims ranged from 14
to 73. More than one-third (35 percent) were between the ages
of 19 and 24. By way of comparison, 11 percent of the popula­
tion and 26 percent of all persons arrested for Index crimes in
the seven cities were in that age category. Almost three­
quarters (73 percent) of all shooting victims whose ages were
known were under 30, and 50 percent were 24 years old and
under. The data presented in Figures 2 and 3 suggest that
existing department sanctions against the shooting of juveniles
(see Table 11) are being observed. Although young persons be­
tween the ages of 13 and 18 represented 39 percent of all
persons arrested for Index crimes in the seven sample cities in
1973, only 12 percent of the shooting victims were in that age
group-a figure in direct proportion to their representation in
the general population of those cities.
46
 TABLE 1
FATAL AND NONFATAL SHOOTING INCIDENTS, 1973 AND 1974

Percentage of Shootings Total

City and Number of
Population• Fatal Nonfatal Shootings

Birmingham 27 73 41
295,686
Oakland 24 76 17
345,880
Portland 33 67 9
378,134
Kansas City 23 77 26
487,779
Indianapolis 36 64 36
509,000
Washington, D.C. 31 69 70
733,801
Detroit 29 71 179
1,386,817

TOTAL 29 71 378b

• Cities are ranked by population size, according to Bureau of the

Census 1973 population estimates; the Indianapolis population figure
refers to police district and is based on 1970 census.
b This figure represents all incidents occurring in both calendar years

in all seven cities.

Comparative data from studies of police homicides by Robin

(1963) and Kobler (1975)1 show similar age breakdowns .
Kobler's study of 911 civilian victims over a five-year period
(1965-69) reports an age range of 12 to 81. The largest group of
victims were between 17 and 19; approximately 50 percent
were between the ages of 17 and 27 . In Robin's study of 32
homicides by police officers in Philadelphia (1950-60), exactly
half of the group was under the age of 24. Robin collected
similar data from nine other cities which showed the largest
percentage of victims to be in the 20-to-25-year-old range (32
percent); overall, 50 percent were under age 28.

47
 Race (N = 309)
Of the number of nonfatally shot civilians whose race was
known (169), almost 80 percent were black, as were 78 percent
of those killed by police use of firearms. Overall, 79 percent of
the shooting victims were black. The percentage of black shoot­
ing victims is disproportionately high in comparison with the
percentage of blacks in the total population; however, the fig­
ure corresponds quite closely to black arrest rates for Index
crimes (see Figure 4). 8
In Robin's study of Philadelphia police homicides, 87.5 per­
cent of shooting victims were black, in contrast to blacks' 22
percent representation in the city's population and 30.6 percent
representation in the arrest population during those same years
(1950-60). In his expanded study of nine additional cities, al­
most 62 percent of shooting victims were black. In contract,
Kobler's study reported a substantially lower percentage of
black victims: 42 percent. However, that study included both
rural areas and, as Kobler points out, a disproportionate num­
ber of cases from the western states. There the population
make-up is likely to differ from that of the large urban centers,
which have been the source of most of the data on this subject
gathered up to now.

Victim Armed (N = 315)

According to police reports, 57 percent of the 315 civilians
shot were armed; 45 percent (143), with guns; and 12 percent
(36), with other weapons, primarily knives. Fifty-four percent
(52) of the subjects killed were armed with guns, and 15 percent
(14) were armed with other weapons. Forty-two percent (91) of
the persons nonfatally shot were armed with guns; in addition,
10 percent (22) were found to have had other weapons. Al­
though many persons shot were unarmed (43 percent), the
data, as shown in Figure 5, suggest that those who were armed
were more likely to be fatally shot. (Suspects were considered
to be armed only if a weapon was reported to have been recov­
ered or, in a few instances, if the records contained additional
evidence that supported the officer's report.)
48
 Kobler's findings concerning the presence or absence of a
weapon in police homicides closely parallel these: In 25 percent
of the cases he studied, no weapon was recovered; in this study,
31 percent of all those persons fatally shot were unarmed. Fifty
percent of all victims had firearms, an additional 15 percent
were armed with knives, and 10 percent apparently had other
weapons.
FIGURE 2
AGE OF SHOOTING VICTIMS
Percentage Number
Ages of Cases of Cases

ALL 100 290•

14-16 6.2 18
14.5
17-18 8.3 24

19-21 16.6 48
34.9•
22-24 18.3 53

25-29 23.1 67

30-34 7.2 21

35-39 5.5 16

40-44 4.5 13

45-49 4.8 14

50-55 2.4 7

Over 55 3.1 9
•In 24 cases, ages were unknown; 5 additional cases were not in­
cluded in the table because the shooting victims were police officers.
Of the total not tabulated (30), 9 were fatal and 21 were nonfatal
shooting incidents.
• The percentage of victims under 25 is 49.4-almost half.

49
c.n Percentage FIGURE 3

0
SHOOTING VICTIMS, URB.41\J POPULATION, AND ARREST POPULATION,

70
,...- BY SELECTED AGE GROUPS
60
r-

50
r-
(39)
40
r- (35)
Seven
30
r- Sample Shooting
(26)
Cities' Victims (23)
Pop. (1973) Seven
20
r- Shooting
(12) (12) Arrested Sample
(11) Victims (11)
10
r- for Cities' (8)
Shooting Seven Seven Pop. (1973) Seven
Victims Sample Index Seven
Sample Arrested Sample
(1973)• Cities' Crimes' Sample Cities'
Cities' for
Pop.b · Cities' Pop.
Pop. Index Pop. Arrested
Crimes
for
Index
Crimes
13-18 years of age 19-24 years of age (25-29 years of age)
-- - · - - - ·- - - - ----- ·- -- --- · ---~

• The percentage of shooting victims v~ries only slightly-and in only one category-when 1974 incidents
are added to those occurring in 1973: 13-18 years of age, 14 percent; 19-24 years of age, 35 percent; 25­
29 years of age, 23 percent.
b Seven sample cities population figures are based on 1970 census data.
' Figures are derived from 1973 UCR report.
FIGURE 4
SHOOTING VICTIMS, CITY POPULATION, AND ARREST
POPULATION , BY RACE
Percentage

100
r-

90
-

80 - (76)
Shooting (73)
70
- Incidents
Involving
Black
Index
60
- Black Crime
Subjects, Arrests,
50 - Seven
Sample
Seven
(39) Sample
40
f.- Cities, Cities
1973• Black 1973•
30
f.- Population ,
Seven
20 ~ Sample
Cities•
10 ~

• Shooting incidents involving black subjects in 1973 and 1974

combined (from seven sample cities) equal 79 percent.
• City population and black population figures from 1970 Bureau of
Census data.
• Figures are derived from 1973 UCR Report.

T ype ofIncident (N = 3 20)

On examination of t he nat ure of the circumstances sur­
rounding each shooting incident, some incidents were r elatively
easy to categorize. For example, the police dispatcher received
a call requesting offi cers to respond to the sce ne of a burglary;
when officer s ar rived , subjects were fou nd on t he premises,
shot s were exchanged, and a suspect was shot. However , in
many incidents a series of event s occurred which meant t hat
officers responding to one presumed set of circumstances found

51
themselves confronted with another. To the extent possible,
incidents were classified according to the primary activity re­
flected in official reports.
According to department records, almost all 320 persons
shot were seemingly involved in criminal incidents-either di­
rectly engaged in illegal activity or acting in a suspicious man­
ner. Of the rest, four were bystanders, one was a citizen at­
tempting suicide, and five were police officers.~ Almost one­
third (32 percent) of the incidents to which the police responded
involved disturbance calls: family quarrels, fights, assaults, dis­
turbed persons, or reports of man or woman with a gun.
Twenty-one percent of the incidents involved reports of a rob­
bery in progress or pursuit of robbery suspects; a nearly equiva­
lent amount (20 percent) involved a burglary in progress, lar­
ceny, tampering with an auto, or pursuit of subjects after an
incident of this nature.
In 8 percent of the cases, persons shot were originally
stopped for a traffic offense or stolen vehicle check; in a num­
ber of these instances, the shooting occurred in the course of
pursuit. Only 4 percent of the shootings were either personal
disputes involving the officer, accidental firings at friends or
coworkers, or the result of horseplay. Another 4 percent of the
incidents involved stakeouts or decoy operations. Miscellaneous
situations, including escapes, investigation of accidents, serving
of warrants, and other circumstances, constituted 11 percent of
the total number of shootings (see Table 2).
Although some studies of shooting incidents have charac­
terized subjects as "confronting" and sometimes "resisting" or
"fleeing" at the time of shooting, our review of the data sug­
gests that it is often extremely difficult to categorize incidents
in this fashion . These postures are not always mutually exclu­
sive, and both police and victim reports of shooting incidents
are to some degree self-serving and not always easily verifiable.
A complicating factor is that the presence of other witnesses
may, depending on the circumstances, merely add to the num­
ber of conflicting accounts. Similar difficulties occur in trying
to identify shootings which are "accidental" or "by mistake"­
a problem researchers appear to share with grand juries.

52
 FIGURE 5
PERCENTAGE OF VICTIMS REPORTED ARMED AND TYPE OF WEAPON
Percentage Fatal Incidents Nonfatal Incidents Total
100

90

80

70

II

60
(15)
I

50

40
(43)

30

20

10

Armed:
Armed
Gun
I

Unarmed Armed
Unarmed: 11111111111

Unarmed Armed Unarmed

~ I I Other weapon

 Officer Status and Assignment (N = 320)

Of the 320 shooting incidents, 17 percent (55) involved off­
duty officers, and 21 percent of the shootings occurring on duty
involved plainclothes officers. 10 Several incidents involving
plainclothes officers on their way to or from work were re­
corded as off-duty incidents. According to police reports, the
majority of off-duty incidents came about because the officer
happened upon the scene of criminal activity or was in a public
place, such as a bar, when a disturbance occurred. Relatively
few off-duty incidents appear to have been initiated by officers.
As Table 4 shows, the ratio of nonfatal to fatal shootings by off­
duty and plainclothes officers is very nearly the same as that of
all incidents (2:1).
TABLE 2
TYPES OF INCIDENTS TO WHICH POLICE RESPONDED
Percentage Number of
Incident• of Total Incidents

Disturbance calls: 32 102

Family Quarrels
Disturbed Persons
Fights
Assaults
"Man With a Gun"
Robbery in Progress, Pursuit of 21 66
Robbery Suspects
Burglary in Progress , Larceny, 20 63
Tampering with Auto, or
Pursuit of Subjects
Traffic Offenses: 8 25
Pursuits
Vehicle Stops
Officer Involved in Personal Dis­ 4 14
pute, Horseplay, or Accident
Stakeout or Decoy 4 13
Other 11 37

TOTAL 100 320

• These incidents were classified from police records for this research.
The categories do not represent formal charges or final dispositions.

54
TABLE 3
STATUS AND ASSIGNMENT OF OFFICERS INVOLVED IN
SHOOTING INCIDENTS

Total
Percentage (Number)

Incidents Involving Off-Duty 17 320

Officers
Incidents Involving Plainclothes 18 320
Officers (as Percentage of On­ (21) (265)
Duty Incidents)

Information about officer status and assignment represents

the only officer-related information consistently available in all
cities. The major difficulty in collecting personal information
about officers, similar to that gathered about persons shot, is
that such information is not likely to be included in the shooting
incident report and must be retrieved from other sources. This
procedure is time consuming and, in some instances, not feasi­
ble because of concerns about the confidentiality of personnel
files. Needless to say, future research efforts looking more in­
tensively into the question of police use of firearms should in­
clude more information about officers in the scope of inquiry.

TABLE 4
RATIO OF FATAL TO NONFATAL SHOOTINGS BY OFF-DUTY AND
PLAINCLOTHES OFFICERS

Fatal Nonfatal Total

(percent) (percent) (percent)

Incidents Involving Off- 29 71 100

Duty Officers
Incidents Involving Plain- 27 73 100
clothes Officers

55
 Disposition (N = 199)
According to police department records, almost 92 percent
of shooting episodes in all cities except Detroit were found to
be justified or not to have resulted in any formal punitive
action, such as a reprimand or suspension by the department.
Detroit was not included in this tabulation because information
about adjudication of police offenses was not available from
incident reports.
Only two departments found less than 90 percent of shoot­
ing incidents in any single year to be justified. One of these
departments was the second largest city, Washington, D.C. (84
percent), which has a well established Weapons Review Board
that scrutinizes all firearms discharges. 11 This finding is consist­
ent with those of other researchers; for example, Delman found
that, on an annual basis, 88 percent of the shooting cases he
reviewed in Los Angeles County were disposed of as justified. 12
Note that review procedures, possible sanctions, and termi­
nology vary among departments, resulting in some difficulty in
interpreting the outcome of administrative review. For the inci­
dents not considered to be justified, department action gener­
ally consisted of a reprimand rather than suspension or termi­
nation.

RELATIONSHIP BETWEEN SHOOTING RATES AND

CITY CHARACTERISTICS
In addition to looking at information about shooting vic­
tims, the researchers also tried to determine whether there is a
correlation between shooting rates in individual cities and city
characteristics, such as population size and the level of criminal
activity reflected in crime rates.

Population Size
In examining the relationship between population size and
shooting rates, it appears that the data support generally as­
sumed trends. On the whole, larger cities have more shootings

56
than smaller cities, and the influence of urbanization is re­
flected in the increased rate of shootings in larger cities. Fac­
tors other than population size, however, affect the shooting
rate. Table 5 shows that more variation in rates occurs within a
group of cities of similar size than between cities of dissimilar
size. The most noteworthy example of this variation is Birming­
ham, which had a higher shooting rate than other cities of
similar size.

TABLE 5
COMPARISON OF POPULATION SIZE AND SHOOTING RATES,
1974

Rate of
Shootings
per 100,000
Population Number of People
Category City Population Shootings (1974)

295,000 Birmingham 295,686 25 8.5

to Oakland 345,880 10 2.9
475,000 Portland 378,134 6 1.6

475,000 Kansas City 487,799 10 2.1

to Indianapolis 509,000" 28 5.5
750,000 Washington, D.C. 733,801 40 5.5

More than
1,000 ,000 Detroit 1,386,817 77 5.6

• The figure refers to police district population.

Department Size
Another factor thought to influence the number of shoot­
ings is police department size. The idea that t he mor e personnel
on the street, the greater the opportunity for interaction be­
tween police and citizens is not substantiated by the data pre­
sented. Table 6 shows the varied experience of the seven cities
in this regard-variations that are particularly striking when
one compares the rates of shootings per 1,000 officers between
cities with similar ratios of officers to population (e.g., Detroit

57
and Washington, or Indianapolis and Birmingham). Any further
exploration of this subject, however, should take into account
t he proportion of t he force assigned to street work, in addi tion
to the total number of personnel in the department.

TABLE 6
RATES OF POLICE SHOOTINGS OF CIVILIANS PER 1,000
OFFICERS•

Rate of Number of
Number Shootings Officers
of per 1,000 per 1,000
City Officers Officers Population

Portland 714 4.2 1.8

Washington, D.C. 4,937 6.0 4.1
Indianapolis 1' 110 7.2 2.1
Oakland 722 9.6 2.0
Kansas City 1,310 12.2 2.6
Detroit 5,575 21.8 4.0
Birmingham 637 25.0 2.1

• Figures are derived from 1973 UCR report and 1973 police data from
the seven sample cities.

Index Crime and Violent Crime Rates

An examination of Index crime'3 rates and shooting rates in
the seven cities over the two-year period shows no consistent
relationship between changes in the number and rate of shoot­
ings and changes in Index crime rates. In some instances, In­
dex crime rates increased while the rate of shootings de­
creased; in other cases, the rever se was true. Kansas City
experienced the greatest increase in Index crime rates of the
seven cities. At the same t ime, the shooting rate decreased 38
percent in that city. A comparison of the violent crime rates
reported by the FBI (homicide, rape , robbery, and aggravated
assault) 14 with shooting rates in the seven cities over the same
two-year period produced a somewhat similar pattern. 15 This is

58
not entirely surprising, given the fact t hat a sizable number of
shooting incidents occurred in conjunction with less se rious of­
fense s which are not reflected in Index or violent crime rates
(see Table 2).
TABLE 7 RATE INCREASES AND DECREASES IN SHOOTING
INCIDENTS, INDEX CRIMES, AND VIOLENT CRIMES,
1973 AND 1974

Percentage Percentage Percentage

Change in Change in Change in
Shooting Index Crime Violent Crime
City• Rates Rates Rates

1. Indianapolis +250 +15 +40

2. Portland + 100 +15 +35
3. Birmingham + 56 +17 +20
4. Oakland + 43 - 3 + 6
5. Washington, D.C . + 33 + 7 - 0 .6
6. Detroit - 25 +18 +18
7. Kansas City - 38 +26 +27

• Ranked from highest increase in shootings (1} to greatest decrease in

shootings (7).

Although a two-year period is insufficient to document es­

tablished trends, the data from Detroit and Kansas City are
nonetheless worth noting. In both cities, shooting rates de­
creased in 1974 even though both Index and violent crime rates
increased. In early March 1974, the Detroit Police Department
abolished STRESS (Stop the Robb eries, Enjoy Safe Streets), a
controversial plainclothes unit whose members had been in­
volved in 17 fatal shootings over a three-year period. In Kan sas
City, a new police administrator took office in November 1973
and soon therafter issued a more restrictive firearms policy in
response to a particularly controversial shooting incident. In
neither instance can it be said, on the basis of available data,
that the reduction in shootings was or was not directly attribut­
able to administrative actions. The impact of those actions can
be determined only by comprehensive, long-range studies .

59
THE USE OF FATAL FORCE BY CITIZENS AND POLICE

Police administrators contemplating changes in their fire­

arms policies will quite likely want to know the answers to the
following:
1. What is the relationship between the use of fatal force
by citizens against police and by police against citizens?
2. Does a reduction in the number of shooting incidents by
police result in increased risk to the police, as measured
by serious injuries or deaths?
As noted in the introduction, we did not collect detailed
information about serious injuries or shootings of police offi­
cers. However, some observations can be made from a compari­
son of police and civilian deaths, using national figures over a
15-year period and figures from the 1973-74 study period in the
seven cities. In Table 8, the number of civilians killed by police
was obtained from the National Center for Health Statistics; as
noted earlier, at least 90 percent of the deaths are presumed to
be the result of firearms use. The data concerning deaths of law
enforcement agents killed by civilians were obtained from the
FBI; the number of deaths resulting from firearms from 1969
through 1975 is indicated in the table.
An examination of national figures fails either to support or
to refute with any certainty the proposition that a reduction in
the number of civilian deaths, possibly reflecting increased re­
straint on the part of the police officers, results in increased
risk to officers' Jives. However, a decrease is apparent in the 1­
to-5 ratio of police killed to police killings, noted by Kobler for
the ten-year period 1960-69. That ratio is now 1 to 4 for the
period 1960 through 1974, and to 1 to 3 for the period 1970
through 1974.
These figures and similar data from the seven cities shown
in Table 9 are simply presented for the reader's information
and as a suggested starting point for further research. Any
attempt to draw conclusions about such a relationship from
these data would be, at best, premature.

60
 Table 10 presents a comparison between the circumstances
of the shootings (fatal and nonfatal) of civilians in the seven
study cities and the circumstances under which police officers
were killed during the same two-year period . The results indi­
cate, as might be expected, that robbery is a high-risk venture
for all concerned; disturbance calls appear to present an even
greater risk to police officers and civilians. This latter finding
suggests that departments that place an emphasis on "danger­
ous felons" in written firearms policies and in related training
curricula may be overlooking a substantial problem area.

TABLE 8 POLICE AND CIVILIAN DEATHS : 1960-1975

Number of Law Enforce­

ment Agents Killed as a Number of Civilians
Year Result of Criminal Action• Killed by Police

1960 28 245
1961 37 237
1962 48 187
1963 55 246
1964 57 278
1965 53 271
1966 57 298
1967 76 387
1968 64 350
1969 86 (83)" 354
1970 100 (93) 333
1971 129 (124) 412
1972 116 (111) 300
1973 134 (127) 376
1974 132 (128) 375
1975 129 (127) not available

• From 1972 on, total includes federal law enforcement agents.

• Numbers in parentheses indicate those killed by firearms.

61
TABLE 9 POLICE AND CIVILIAN DEATHS, 1973 AND 1974: SEVEN
SAMPLE CITIES
1973 1974

Police Civilian Police Civilian

City Deaths• Deaths Deaths• Deaths

Birmingham 1 5 0 6
Detroit 3 28 5 24
Indianapolis 0 2 2 11
Kansas City 0 5 0 1
Oakland 3 1 2 3
Portland 0 0 1 3
Washington, D.C. 1 10 1 12
• As a result of criminal action.

TABLE 10 CIRCUMSTANCES SURROUNDING SHOOTING OF

CIVILIANS AND DEATHS OF POLICE OFFICERS,
1973-1974
Percentage Percentage of
of Civilians Fatal and
Killed by Nonfatal
Police Shootings of
Percentage of Officers, Civilians by
Police Officers Seven Police Officers,
Killed Sample Seven Sample
Type of Incident Nationwide Cities Cities

Robbery,
Pursuit of Robbery
Suspect 20 21 21
Burglary,
Pursuit of Burglary
Suspect 6 13 20
Disturbance Calls 24 36 32
Traffic Stops 14 13 8
All Other" 36 17 19
TOTAL 100 100 100

• Because information on the circumstances surrounding shootings of

civilians and police is maintained differently by different agencies,
many categories proved to be noncomparable and had to be merged
in the category " All other."

62
 Notes
1. Individuals with police experience were hired as field researchers to
visit selected cities to gather data. One is currently a lieutenant in the
District of Columbia Metropolitan Police Department; one is a writer
and former D.C. police officer; and one is a lieutenant in the Birming­
ham, Alabama, police department.
2. As noted in the introduction, the staff collected information about
shooting incidents from the Detroit police department for the entire
year 1974, but only the last six months of 1973, because of the size of
the department and the large number of cases to be reviewed. How­
ever, because the total number of shootings (both fatal and nonfatal) in
that city in 1973 is known, that figure is used in several of the tables in
this chapter and is identified as such. Similarly, there are occasional
references to the total number of shootings-378-in all seven cities
over the entire two-year period.
3. The term "incident" refers to the shooting of an individual subject
even though several individuals may have been shot in one episode.
4. For the most part, data were collected by Police Foundation field
researchers from department reports of shooting incidents. Relevant
items of information were obtained by reading through a number of
reports in each individual folder; in most instances, personnel informa­
tion was maintained in a separate location and could not within the
time available be correlated with data from shooting incident reports.
5. This figure includes all fatal shootings of civilians by police officers
in Detroit in 1973.
6. It should be kept in mind that a small number of civilian deaths
reflected in Public Health Service statistics are the result of means
other than firearms.
7. Gerald D. Robin, "Justifiable Homicides by Police," Journal of
Criminal Law, Criminology, and Police Science 54 (1963): 224; Arthur
L. Kobler, "Figures (and Perhaps Some Facts) on Police Killing of
Civilians in the United States, 1965-1969," Journal ofSocial Issues 31,
1 {1975): 185-91.
8. Age may be a very significant factor in the disproportionate num­
ber of black victims in comparison to their representation in the popu­
lation. In some jurisdictions-New York City, for example-the me­
dian age of black males is 23.1 as compared to 33.3 for white males
(Bureau of Census, 1970 data). The arrest rate for Index crimes of
persons in the age groups 13-18 and 19-24, the vast majority of whom
are male, is considerably higher than the arrest rate of persons 26
years of age and above.

63
9. Characteristics of police officers shot are not included in tabula­
tions of age, race, or presence of weapon.
10. An incident was considered to involve plainclothes officers only if
the officers were on duty and assigned as such.
11. The Washington, D.C. , department reports that in the year pre­
ceding the publication of this report, less than 60 percent of the cases
reviewed by the board were found to be justified. This figure, how­
ever, includes discharges which did not take effect. Such incidents
were excluded from analysis in this report.
12. Gerald F. Uelman, "Varieties of Police Policy: A Study of Police
Policy Regarding Use of Deadly Force in Los Angeles County," 6
Loyola L. Rev. 39 (1973).
13. Index crime offenses, as reported by the FBI, are murder, forcible
rape, robbery, aggravated assault, burglary, larceny-theft, and auto
theft.
14. Victimization rates for personal crimes of violence, as recorded by
LEAA surveys, are undoubtedly higher; however, data are not availa­
ble on a comparable basis for all seven cities.
15. Although both Portland and Indianapolis reflected the highest in­
creases in shooting and violent crime rates, it should be noted that the
actual number of shootings in Portland increased only from three to
six.

64
 CHAPTER 3

USE OF DEADLY FORCE TO ARREST A

FLEEING FELON-A CONSTITUTIONAL

CHALLENGE, PARTS I AND III

J. PAUL BOUTWELL

INTRODUCTION
From the 15th century to the present day, a law enforce­
ment officer's use of deadly force 1 to apprehend one fleeing
from a crime has been largely governed by the felony-mis­
demeanor classification of crimes. 2 An officer may use deadly
force to prevent the escape of a fleeing felon, but he may not
use such force to apprehend a fleeing misdemeanant. The ra­
tionale for permitting deadly force to be used against a felon, at
least at early common law and in 18th century America, was
that all felonies-murder, rape, manslaughter, robbery, sod­
omy, mayhem, burglary, arson, prison break, and larceny­
were punished by death.3 The use of deadly force was seen as
merely speeding up the process. "It made little difference if the
suspected felon was killed in the process of capture since, in the
eyes of the law, he had already forfeited his life by committing
the felony." 4

SOURCE: "Use of Deadly Force to Arrest a Fleeing Felon-A Constitutional

Challenge, Part I," FBI Law Enforcement Bulletin (September 1977) 46:9:27­
31;

"Use of Deadly Force to Arrest a Fleeing Felon-A Constitutional Challenge,

Part III," FBI Law Enforcement Bulletin (November 1977) 46:11:9-14.

65
 On the other hand, deadly force could not be used against a
fleeing misdemeanant under any circumstances. "[T]o permit
the life of one charged with a mere misdemeanor to be taken
when fleeing from the officer would , aside from its inhumanity,
be productive of more abuse than good... .The security of per­
son and property is not endangered by a petty offender being at
large...." 5
Through the years, the line between felonies and misde­
meanors has become less distinct. The number of crimes classi­
fied as felonies has increased significantly. Our concept of pun­
ishment has undergone substantial changes since the early days
of common law. Yet, there has not been a significant change in
the rule permitting the use of deadly force to arrest any fleeing
felon. This has meant, therefore, that deadly force is authorized
in many more situations today than existed in earlier days.
While there is general agreement that deadly force is .justi­
fied against a fleeing felon when the felony committed is a
dangerous or violent one, there is considerable controversy
over the use of such force when the felony is a minor, nonvio­
lent one. The argument is that many of today's minor felonies
are simply not analogous to the felony classification at common
law when the fleeing felon rule was formulated.
Efforts to reform the common-law fleeing felon rule have
been directed primarily toward limiting the use of deadly force
to dangerous felons. While there has been some movement
away from the "any felony" rule during this century, it has
remained essentially intact. Those who have sought to restrict
the use of deadly force in arrest situations have done so on four
fronts; namely, (1) legislative reform, (2) state civil court action,
(3) departmental policy restrictions, and (4) challenge to the
rule's constitutionality.
One state's codification of the common law fleeing felon
rule has been declared unconstitutional by a federal Court of
Appeals.6 While the U.S. Supreme Court vacated judgment in
the case, it did so on a procedural deficiency and not on t he
merits of the court's holding.7 Therefore, the opinion of the
appeals court continues to represent, at least on the merits, a

66
conflict with other federal circuits as to whether' the use of
deadly force to apprehend a nondangerous fleeing felon is a
constitutional violation.
This article, while discussing efforts made toward legisla·
tive reform and departmental policy restrictions, emphasizes
the challenge that has been made in federal court to the consti·
tutionality of the rule . This type of litigation will be distin·
guished from a state court civil suit.

THE COMMON LAW FLEEING FELON RULE

The common law rule permitting the use of deadly force to
effect the arrest of any fleeing felon has been both severely
criticized and staunchly defended . A summary of the commonly
expressed arguments for and against the rule is presented to
bring the different views into sharp focus. Even though some of
the points are more moral and sociological than legal, they
should add to our understanding of the rule and illustrate why
controversy seems to develop when the rule is discussed.

Argument for the Rule

Society requires protection against criminals. Criminal laws
are enacted to give legal content and efficiency to such protec·
tion. Enforcement of these laws requires prosecution of those
who violate them. Arrest of the violator is a condition prece·
dent to the entire enforcement procedure. The whole criminal
justice system breaks down unless society can require peace­
able surrender to the exertion of law enforcement authority.
Therefore, society benefits from that which facilitates arrest.
Obviously, the right to use deadly force facilitates arrest. Its
lawful use notifies the criminal that flight is not an option open
for his consideration. If he flees from the commission of a fel·
ony, against the order of an officer of the law, he should realize
that he invites the risk of injury or death. This does not mean
that the officer will always exercise the right to shoot, but it
should not mean that the advantage should belong to the crimi·
nal.

67
 If effective law enforcement is to be maintained, certainly
an arrest should not be made to turn on who can run the fast­
est. There is no constitutional right to commit a felony and then
escape the consequences by fleeing. There is no constitutional
right to flee from an officer lawfully exercising his authority. It
has been said that if a fleeing felon is injured or killed, he must
be regarded as the author of his own misfortune. 8
A law enforcement officer is called upon to make a difficult,
on-the-spot legal judgment. His facts are often vague and am­
biguous. Yet, his decision must be swift. If he uses force, it
must not be unreasonable. This standard presupposes that a
law enforcement officer is endowed with foresight. Of those
who would change the rule, some would require the officer,
before using deadly force, to believe the felon will use force
against others if not immediately apprehended. How can a po­
lice officer ever know, reasonably or otherwise, whether the
felon will use force against others?9
Given the long history and current status of justification,
the ready availability of handguns to the populace at large (in­
cluding nonviolent felons), and the needs of law enforcement in
a society where violence is widespread, the justifiable homicide
statutes which permit deadly force against any fleeing felon are
not unreasonable. 10
Surely a police officer should not be imprisoned if he mis­
takes a nondangerous felon for a dangerous one or a nonforci­
ble felon from a forcible one. A police officer faced with an
emergency situation makes a mistake and uses deadly force
against a nondangerous felon. He and his employing agency
may or may not be civilly liable, he may or may not be disci­
plined for not following a departmental policy, but it should not
be said, out of awareness of his difficult job in emergency situa­
tions, that he assumes the risk of going to jail for his mistake.

Argument Against the Rule

The common law distinction between felony and misde­
meanor crimes for the purpose of determining the scope of the
privilege to use deadly force is grossly inadequate for modern­
day law enforcement. A felony usually is based merely on the
68
length of sentence involved, and some misdemeanors embrace
conduct more dangerous than many felonies.
In 15th century England, as well as 18th century America,
the rule reflected the social and legal context of felonies at that
time. They were punishable by death. It made little difference if
the suspected felon were killed in the process of capture, since
in the eyes of the law he had forfeited his life by committing the
felony. It was assumed that a suspected felon facing death upon
capture was more desperate than a misdemeanant, and greater
force was required for his apprehension. Only a few crimes
were felonies. In most American jurisdictions, the social and
legal context of felonies today bears little resemblance to that
of the early common law. For example, some modern code revi­
sions classify felonies according to five different categories
ranging from Class A felonies, the most serious, down to Class
E felonies, the least serious. 11
Felonies today include numerous crimes not involving force
or violence, such as property-based crimes and compliance with
complex government regulations (e.g., income tax fraud) . Since
the felony-misdemeanor distinction is usually based merely on
the length of sentence involved, and since some misdemeanors
involve conduct more dangerous than some felonies, a deadly
force justification, which makes no distinction between felonies
or does not address the gravity and need of such force, bears
elements of irrationality. 12
Deadly force should not be permitted when the felony com­
mitted is a minor, nondangerous one. Felonies against prop­
erty, which as larceny, forgery, and counterfeiting, are re­
garded as being nondangerous. Deadly force should be
permitted only to apprehend or prevent escape of a dangerous
felon. The crime for which the arrest or recapture is sought
should involve conduct including the use or threatened use of
deadly force.
Speaking against the common law rule, Professor Michael
Mikell stated:
It has been said, 'Why should not this man be shot
down, the man who is running away with an automo- ·
bile? Why not kill him if you ca~not arrest him?' We

69
 answer: because, assuming that the man is making no
resistance to the officer, he does not deserve death ....
May I ask what we are killing him for when he steals
an automobile and runs off with it? Are we killing him
for stealing the automobile? If we catch him and try
him, we throw every protection around him. We say he
cannot be tried until 12 men of the grand jury indict
him, and then he cannot be convicted until 12 men of
the petit jury have proved him guilty beyond a reason­
able doubt, and then when we have done all that, what
do we do to him? Put him before a policeman and have a
policeman shoot him? Of course not. We give him three
years in a penitentiary . It cannot be then that we allow
the officer to kill him because he stole the automobile,
because the statute provides only three years in a peni­
tentiary for that.... Is it for fleeing that we kill him?
Fleeing from arrest is also a common law offense and is
punishable by a light penalty, a penalty much less than
that for stealing the automobile. If we are not killing
him for stealing the automobile and are not killing him
for fleeing , what are we killing him for? 13

And also from Professor Wechsler:

... [T]he preservation of life has such moral and ethical
standing m our culture and society, that the deliberate
sacrifice of life merely for the protection of property
ought not to be sanctioned by law. 14

As can be readily seen, valid points are to be made on both

sides of the argument. It is also clear that the debate deals with
competing interests of society at the highest rank-interests in
protecting human life against unwarranted invasion and in pro­
moting peaceable surrender to the exertion of law enforcement
authority. Yet, the balance that has been struck to date is very
likely not the best one that can be. In the area where any
balance is imperfect, there must be some room for different
views to prevail .15
The American Law Institute's almost 50 years of consider­
ation of the problem demonstrates that the area in which we
are treading is one still characterized by "shifting sands and
obscured pathways.''

70
RESTRICTIONS UPON THE USE OF DEADLY FORCE
THROUGH LEGISLATIVE REFORM
Most states have justification statutes dealing with the use
of deadly force by law enforcement officers to effect arrest.
They may be divided into three groups; namely, those that fol­
low the common law rule , those that have modified the rule and
mandate that only "forcible" felonies justify the use of deadly
force, and those that have adopted the Model Penal Code. Each
approach will be discussed.

Codification of the "Any Felony" Rule

At least 24 states currently have codifications of the com­
mon law fleeing felon rule. 16 It would not be accurate to as­
sume, however , that states with such statutes are relying on
archaic law or that the respective state legislative bodies have
not considered different versions of the rule. At least 17 of the
24 states have revised and updated their penal codes since 1970
and have preserved the rule in legislative recodifications. The
Missouri House of Representatives, for example, rejected an
attempt to amend their statute in June 1975. 17
Under the provisions of a typical state statute, four require­
ments must be present to justify deadly force: (1) The officer
must have probable cause to believe that a felony has been
committed and that the person to be arrested committed it; (2)
the arresting officer must give the defendant notice of his in­
tention to arrest; (3) the defendant either flees or forcibly re­
sists; and (4) whatever force the officer uses must be necessary
to effect the arrest. 18 While an officer cannot use deadly force
to apprehend a fleeing misdemeanant, he is privileged to use
such force regardless of the felony that is committed.
The significance of " necessity" as a limitation upon the use
of deadly force was illustrated recently in a civil rights case. At
approximately 10 p.m. two officers received a radio dispatch
indicating an "entry in progress, three Negro males on the
scene ...." Armed with a 12-gauge shotgun, one officer posi ­
tioned himself at the scene on a well-lighted public sidewalk,

71
 while his partner circled around to the rear of the homes lo­
cated on the block. The officer on the sidewalk was five to ten
yards away when he observed three black males emerge from
the gangway located between two houses. Each one was of
junior high school age and approximately 5 feet 6 inches tall.
The plaintiff, one of the three boys, had in his hand a thin, 12­
inch-steel-blue file. He turned and faced the officer for an in­
stant after he was ordered to halt. All three then retreated into
the gangway of an adjacent house. The officer fired his shotgun
over their heads as they ran. The officer ran up the sidewalk,
parallel to their path of retreat, and positioned himself directly
in front of that gangway. The plaintiff was now facing the
officer again. He was approximately 45 feet from the officer.
The officer fired a second shot directly at the plaintiff, hitting
him in the head. Another boy was also hit. A civil rights action
against the police officer claiming money damages for the use
of excessive force in connection with the arrest was commenced
in federal district court by the two injured.
The trial was a bench trial, which means simply that the
trial judge, in addition to deciding questions of law, also makes
the necessary factual determinations. He decided that the offi­
cer's version of the events, related above, represented the fac­
tual backdrop against which the liability issue would be deter­
mined. The officer testified that he believed the plaintiff
wielding the file had a long-barreled revolver and that he feared
for his life. He also testified that he believed that mere flight by
one suspected of burglary justified the use of his shotgun. Un­
der state cases, the test for liability was whether the amount of
force used by the arresting officer was reasonable under the
existing circumstances. While an officer may use deadly force
to apprehend any fleeing felon, he must reasonably believe it
necessary to prevent escape.
The trial judge held that the defendant's second shot, aimed
directly at the suspects, was unreasonable and unjustified. The
judge took into account not only the officer's frightened state
of mind , but also the lighting conditions, the proximity of the
boys to the officers, the physical appearance of the file, the

72
suspects' retreat, and the defendant's awareness that his part­
ner covered the only available avenue of escape, and deter­
mined the officer used excessive force in effecting the arrest.
While the defendant may have actually feared for his life, he
said, a defense is still not established. The belief must also be
reasonable under all the existing circumstances. Judgment was
for the plaintiff. 19 The judge's decision was affirmed on appeaU0

Statutes Limiting Use ofDeadly Force to "Forcible"

Felonies
Seven states have justification statutes which specify the
felonies for which deadly force may be used. 21 These statutes
permit such force only for "forcible" felonies. The force used
must be necessary to effect arrest. In addition, if a person is
attempting to escape by use of a deadly weapon, or otherwise
indicates that he will endanger human life or inflict great bodily
harm unless arrested without delay, then deadly force is per­
mitted.
The officer knows which felonies are "forcible" because
they are defined by statute. In states which have such statutes,
the law enforcement officers memorize those felonies so that
there is no misunderstanding as to what constitutes a forcible
felony. Generally, training academies use the acronym MA M.
BARKER to teach what felonies are forcible-M is for murder,
A for arson, M for mayhem, B for burglary, A for aggravated
battery, R for rape, K for kidnapping, E for extortion, and R
for robbery.
The officer, in deciding whether or not to use deadly force,
asks himself two questions: (1) Has the person to be arrested
committed a forcible felony? and (2) Is it necessary to use
deadly force? If the answer is "no" to either question, then he
may not use such force. Take this illustration. A thief steals a
$500 diamond ring from the counter of a jewelry store. A dep­
uty is attempting to arrest the thief, but he flees. The thief does
not have a weapon. The deputy is not justified in using deadly
force, since theft, although a felony under these circumstances,
is not a forcible felony. 22

73
 Model Penal Code
The Model Penal Code proposes that the use of force be
justifiable only where the arresting officer believes that (1) the
crime for which the arrest is made involved conduct including
the use or threatened use of deadly force, or (2) there is a
substantial risk that the person to be arrested will cause death
or serious bodily harm if his apprehension is delayed. 23 The
philosophy of this approach is to ignore the technical classifica­
tion of a crime as a felony or misdemeanor and to focus instead
on a balance of interests-the need to apprehend suspects and
preserve the safety of the arresting officers as against the
value of human life.
Seven states have justification statutes which have adopted
the Model Penal Code. 24 New York adopted the Model Penal
Code approach in 1965, but returned to the forcible felony rule
in 1967. Idaho adopted it in 1971, but repealed it three months
after its effective date in 1972.25

TITLE 42 U.S.C. 1983 SUITS

The essential elements of a section 1983 case are (1) con­
duct of some person, (2) acting under color of state law, and (3)
which deprives another of rights, privileges, or immunities se­
cured by the Constitution or laws of the United States. The
essence of the action is a claim to recover damages for injury
wrongfully done to another person. The liability is personal.
Allegations of misconduct in 1983 suits are drawn from a
broad spectrum of rights, privileges, and immunities afforded
protection by the federal Constitution and laws of the United
States. The approach is for the complainant to allege a violation
of the fourteenth amendment, section 1, which contains the
following language: "No State shall make or enforce any law
which shall abridge the privileges or immunities of citizens of
the United States; nor shall any State deprive any person of
life, liberty, or property, without due process of law; nor deny
to any person within its jurisdiction the equal protection of the
laws."

74
 The key phrases, "privileges and immunities," and "due
process of law," and "equal protection of the laws" are the
vehicles by which 1983 protections are usually identified. For
example, the guarantee against unreasonable searches and sei­
zures contained in the fourth amendment is applicable to state
officers by reason of the "due process" language of the four­
teenth amendment. Thus, an officer acting contrary to the
fourth amendment might be held liable for denying a citizen his
constitutional right to due process.
Practically all routine law enforcement work has the poten­
tial of becoming the subject of complaint by an irate citizen who
demands satisfaction by way of a civil suit under this statute.
Therefore, one of the heavy responsibilities of each law enforce­
ment officer is to recognize and protect the rights, privileges,
and immunities of persons within the jurisdiction he serves.
Section 1983 crystallizes the officer's duty in this respect where
constitutional or federal rights are concerned. Thus, the statute
implies that an officer has a specific duty to avoid depriving
others of the enjoyment of these guarantees and that, by his
failure to comply with that duty, he may incur personal liability
for the resulting injuries.
Does this mean that an officer, who is negligent in the use
of his firearm, may be sued in federal court under 1983 for the
violation of a constitutional right?
Section 1983 was not intended to be a substitute for state
tort action, nor grant a federal forum for every citizen's claim
or injury by a state official. Negligence, as such, is not action­
able as a civil rights complaint. The official conduct must de­
prive another of a constitutional right. 26 Yet, conduct that a
state court would classify as negligence has formed the basis of
a 1983 suit. Let us look at some examples of constitutional
classifications and see how plaintiffs have fashioned their com­
plaints so as to bring their case into federal court as a 1983 case
of action.

THE FOURTH AMENDMENT

The fourth amendment declares in part: "The right of the

75
people to be secure in their persons . . . against unreasonable
searches and seizures, shall not be violated ...." This constitu­
tional provision has long been interpreted to embrace security
from arbitrary intrusion by the police. The following case illus­
trates how one federal court applied this language to facts that
sound of negligence. An officer, after reporting to the scene of
a disturbance, observed a young boy leave the scene. The offi­
cer pursued, thinking the boy had a gun. The boy carried a tire
tool in his hand, which he dropped when the officer yelled for
him to "halt." All the witnesses, including the officer, heard
the tool drop. The officer testified that as he lowered his gun he
accidentally pulled the trigger, putting a hole through the boy's
thigh. The district judge found the officer' s use of force
amounted to gross or culpable negligence; however, he was of
the opinion that the plaintiff could not prevail under federal law
since 1983 was not intended as a means of recoupment for
injuries caused by the negligence of a state officer acting in the
course of his duty. With this the appellate court disagreed. The
appeals court reasoned that gross or culpable conduct was the
equivalent of arbitrary action; that is, the officer's action was
more than just simple negligence. "Our concern here is with
the raw abuse of power by a police officer . . . and not with
simple negligence on the part of a policeman or any other offi­
cial."27 Such arbitrary action is a constitutional violation.

CRUEL AND UNUSUAL PUNISHMENT-THE EIGHTH

AMENDMENT
Plaintiffs have also contended that the use of deadly force
against a nonviolent fleeing felon is cruel and unusual punish­
ment in violation of the eighth amendment. In a recent case,
officers investigating a burglary attempt killed the plaintiff's
son as he was fleeing from an arrest. The plaintiff contended
that the state statute, which followed the common law "any
felony" rule, was unconstitutional on its face and as it was
applied because it permitted the administration of cruel and
unusual punishment in violation of the eighth amendment.
Deadly force can be constitutionally authorized only when nee­

76
essary to protect " one's own life or safety, or the life and
safety of others."
The three-judge court, convened to determine the constitu­
tionality of the state statute permitting the use of deadly force
to arrest any felon, held that the statute was not in violation of
the eighth amendment. The amendment deals with punishment,
and the short answer to the plaintiff's contention was that the
state statutes simply were not dealing with punishment. An
officer in effecting an arrest cannot use any force for the pur­
pose of punishing a person and to do so is a crime under title
18, United States Code, section 242. It may be better as a value
judgment to allow nonviolent felons to escape rather than incur
the risk of killing them. But that is a policy question for the
state legislature, not for the federal courts to decide in the
guise of constitutional adjudication, the court said. The panel
went on to hold that the state statute was not unconstitution­
ally overbroad or vague and was not violative of the equal pro­
tection clause of the fourteenth amendment. 28
The use of deadly force by law enforcement officers
in effecting an arrest is a well-recognized ground for a
1983 case. Yet, the exact place in the Constitution of a
right to be free from such force is not clear and has
been the subject of disagreement in the decisions of the
Federal courts of appeal.

DUE PROCESS
The fifth amendment to the U.S. Constitution provides in
part: "No person shall deprived of life, liberty, or property,
without due process of law . . . ." The fourteenth amendment
applies the same limitation on the states: " . .. nor shall any
State deprive any person of life, liberty, or property, without
due process of law.. . ."
The use of deadly force by law enforcement officer s in ef­
fecting an arrest is a well-recognized ground for a 1983 case.
Yet, the exact place in the Constitution of a right to be free
from such force is not clear and has been the subject of dis­
agreement in the decisions of the federal courts of appeal. Sev­
77
eral opinions have expressed the thought that the right arises
from the due process clause of the fourteenth amendment; that
is, the right to be secure in one's person, a right to life itself,
which stands separate and apart from any specific right found
in the Bill of Rights. Such a right is fundamental and basic to
an ordered society and is inherent in the Constitution. It is thus
protected by the due process clause. The claim is, therefore,
that the state fleeing felon statute violates the due process
clause of the fourteenth amendment because, procedurally, it
permits the arbitrary imposition of death by the officer, violates
the presumption of innocence, and denies the suspect a right to
trial by a jury. Of course, the arguments would apply as well to
the use of deadly force against the violent, dangerous felon.
Courts, in applying a due process analysis, attempt to balance
the interests of society in guaranteeing the right to life of an
individual against the interest of society in insuring public
safety. They have not agreed on where the balance should be
struck.
Two cases illustrate the conflict. Both are from states
which follow the common law "any felony" rule, and perhaps
best illustrate the constitutional challenge made against the
rule. One case is from Connecticut; the other is from Missouri. 29

Connecticut Case
An officer, while cruising in his patrol car in the ordinary
course of his duties, observed an automobile occupied by three
young males. Both cars proceeded for several blocks at a lawful
rate of speed. Through radio contact, the officer determined
the vehicle had been reported stolen. The boys became aware
they were being followed and accelerated to about 80 miles per
hour. The officer followed in hot pursuit. After traveling sev­
eral blocks, they reached the end of the road. Both the stolen
vehicle and the patrol car slid to a stop, causing a large cloud of
dust. Since the occupants of the car were not immediately visi­
ble, the officer climbed to the top of a nearby embankment. He
observed two men running across a nearby field and called for

78
them to halt. They momentarily turned to face him, but then
began to run away. The officer fired his gun at the leg of one of
the fleeing suspects, but struck him in the left buttock, causing
internal injuries which resulted in his death. It was stipulated
that none of the occupants had threatened physical injury to the
officer in any manner.
The rule in Connecticut is that an arresting officer may use
deadly force if he reasonably believes it necessary to effect an
arrest, or to prevent the escape from custody of a person whom
he reasonably believes has committed or attempted to commit a
felony. 30

Missouri Case
Two young boys entered the office of a golf driving range
at night by means of an unlocked window for the purpose of
stealing money. As they departed through a back window they
were intercepted by a policeman. He ordered them to stop, but
rather than submit to arrest, they fled in different directions.
As another officer, who had just arrived on the scene, rounded
the building, he collided with one of the boys. They both fell to
the pavement. The officer grabbed the boy's leg, but he broke
from the officer's grasp and ran. The officer pursued, but was
losing the race. He shouted: "Stop, or I'll shoot," but the boy
did not stop. Believing that it was necessary to take further
action to prevent escape, the officer fired a warning shot. The
bullet, however, struck the youth in the head, causing his death.
It was stipulated by the parties that the officer's use of his gun
was "reasonably necessary under the circumstances and was
authorized by the statutes of the State of Missouri."
The pertinent Missouri statutes read as follows:
Justifiable Homicide
Homicide shall be deemed justifiable when commit­

ted by any person in either of the following cases:

(3) When necessarily committed in attempting by

lawful ways and means to apprehend any person for

any felony committed, or in lawfully . . . keeping or

preserving the peace.

79
 R ights ofOfficer in Making Arrests
If, after notice of the intention to arrest the defend­
ant, he either flees or forcibly resists, the officer may
use all necessary means to effect the arrest.~ 1
A civil rights action was instituted in each case under title
42, United States Code, section 1983, alleging that the individ­
ual officers, acting under color of state law, deprived the flee­
ing persons of their lives without due process of law. The offi­
cers' answers were the same; namely, they acted in good faith
plus they had a reasonable basis to believe their conduct was
lawful. In each case, the arresting officer simply relied upon
validity of the state statute, which permits a law enforcement
official to use deadly force in apprehending a person who has
committed a felony.
The plaintiffs' contention was that such statutes as these
are unconstitutional, and they should be declared so by the
federal courts. While such declarations may not affect the liabil­
ity of the current defendants, it would remove the defense of
good faith in future damage actions of this kind. They asked the
courts in each case to fashion a constitutional standard which
would restrict the use of deadly force in effecting an arrest to
violent felonies or circumstances where there is substantial risk
that the person to be arrested will cause death or serious bodily
harm if his apprehension is delayed.
In the Connecticut case, the federal appellate court rejected
the plaintiff's argument: " ... [S]tates must be given some lee­
way in the administration of their systems of justice, at least
insofar as determining the scope of such an unsettled rule as an
arresting officer' s privilege for the use of deadly force. Fur­
ther, in the light of the shifting history of the privilege, we
cannot conclude that the Connecticut rule is fundamentally un­
fair."32
In the Missouri case, the federal district court held that a
defense of good faith had been established and therefore denied
an award of damages. The court concluded there was no longer
a controversy between the parties which would permit the
granting of declaratory relief; therefore, the court declined to

80
rule on whether the Missouri statutes were unconstitutional.
Even if the statutes were unconstitutional, the court reasoned,
the defense was still available to the officer, since he reasonably
believed in their constitutionality at the time. No appeal was
taken from the denial of damages, but the plaintiff appealed the
court's denial of declaratory relief. The federal appellate court
disagreed with the district court and remanded the case for
consideration on the merits of the constitutional issue. The
good faith defense cannot serve as a reason for denying equita­
ble relief. Furthermore, the appellate court disagreed that the
parties lacked sufficient adverse interest. The result of a declar­
atory judgment in favor of the plaintiff would be to remove the
defense of good faith in future damage actions. "Those who
would use a statute as a shield must be prepared to defend the
constitutional validity of that shield." 33
On remand, the district court held the Missouri statutes did
not violate the U.S. Constitution. To abolish the use of deadly
force would deprive the state and its citizens of their rights to
security, safety, and a feeling of protection. To pick and choose
those crimes warranting the application of deadly force is the
duty of the legislature. "It is not the role of a federal judge to
legislate for the people of a state. " 34
On the second appeal, the federal appellate court again re­
versed and held the Missouri statutes unconstitutional. Statutes
as broad as these deny due process in that they create a conclu­
sive presumption that all fleeing felons pose a danger to the
bodily security of the arresting officers and the general public.
The court reasoned:
The police officer cannot be constitutionally vested
with the power and authority to kill any and all escap­
ing felons, including the thief who steals an ear of corn,
as well as one who kills and ravishes at will. For the
reasons we have outlined, the officer is required to use
a reasonable and informed professional judgment, re­
maining constantly aware that death is the ultimate
weapon of last resort, to be employed only in situations
presenting the gravest threat to either the officer or
the public at large. Thus we have no alternative but to
find [the statutes] unconstitutional in that they permit

81
 police officers to use deadly force to apprehend a flee­
mg felon who has used no violence in the commission of
the felony and who does not threaten the lives of either
the arresting officers or others.35
On May 16, 1977, the U.S . Supreme Court vacated the
judgment of the Court of Appeals and remanded the case with
instructions to dismiss the complaint. For a declaratory judg­
ment to issue, there must be a dispute which calls for an adjudi­
cation of adverse interest. There was no such dispute in this
case. The plaintiff's claim of a present interest was twofold: (I)
that he would gain emotional satisfaction from a ruling that his
son's death was wrongful ; and (2) he has another son, who if
ever arrested on suspicion of a felony, might flee or give the
appearance of fleeing, and would therefore be in danger of
being killed by defendant or other police officers. As to the first
claim, the Court stated that emotional involvement in a lawsuit
is not enough to meet the case or controversy requirement, and
were the law otherwise, few cases could ever become moot. As
to the second claim, the Court stated that such speculation is
insufficient to establish the existence of a present, live contro­
versy.36
In disposing of the case in the manner described above , the
Supreme Court emphasized it was not considering the merits of
the Court of Appeals' opinion. Therefore, the question whether
the use of deadly force to apprehend a nondangerous fleeing
felon constitutes a violation of the U.S. Constitution remains
open. The Missouri case represents the only federal appellate
court opinion which, on the merits, has indicated that it does.

CONCLUSIONS
Critics of the common law rule claim the use of deadly force
against a nondangerous fleeing felon is an abuse of deadly
force. The possible remedies against such abuse-namely, civil
liability or criminal prosecution, or both-are ineffective deter­
rents. Where the state has a justifiable homicide statute which
codifies the common law "any felony" rule, it operates to form
a shield for the officer, not only against criminal liability but

82
 also against civil liability. Thus, civil courts, while not techni­
cally bound to do so, usually recognize in the state statutes a
legislative policy toward which they will defer in defining tort
liability. Even while doing so one court pointed out: " ... the
preferable rule would limit the privilege to the situation where
the crime involved causes or threatens death or serious bodily
harm, or where there is a substantial risk that the person to be
arrested will cause death or serious bodily harm if his apprehen­
sion is delayed. " 37
Every modern law enforcement executive knows well his
duty to insure efficient and effective firearms training before
an officer is assigned a weapon. Yet, the executive's responsi­
bility does not rest there. He realizes, in addition, that the
officers under his command are entitled to clear and specific
instruction on the circumstances under which the use of a fire­
arm is permissible. This takes form in written departmental
policy.
One law enforcement executive has remarked that "a policy
without teeth is just about as effective as a patrol car with four
flat tires." Policy must be reinforced by effective instruction
from recruit training at the academy through advanced inser­
vice or firearms training throughout an officer's career.
Notwithstanding departmental policy and excellent instruc­
tion in both the skill and proper use of a sidearm, the final
decision to use it must rest with the individual officer. That
decision will be formed in some measure by his own moral and
ethical judgment concerning the use of deadly force. The ad­
ministrator should be as concerned with an officer who is afraid
to use his sidearm when the situation requires its use as he is
with the officer' s reckless and unjustified use . He fulfills his
administrative duty when he addresses both issues. A recent
Police Foundation report38 makes the point that many depart­
ments lack adequate recordkeeping procedures designed to
identify and monitor officers' conduct involving the use of ex­
cessive force and repeated involvement in shooting incidents.
The authors point out " . .. the lack of systematic centralized
data collection in many departments inhibits the rational devel­

83
 opment of new policies, training programs, and enforcement
procedures." 39
One important misconception about deadly force that be­
came evident in the several cases reviewed in this article is that
officers think they have the ability to shoot to wound when the
person shot at is fleeing the scene. In case after case, the testi­
mony of the officer was to the effect that he actually shot at an
arm or leg, but the bullet struck the head, the neck , or the back.
One coroner's report stated: "Given a moving target, in a
range of seventy-five yards, or less, the target will probably be
hit, but not where the gun was aimed. Therefore, the police
officer should not think he is going to inflict a nonfatal wound
by shooting at an arm or leg. He should fully expect the shot to
be fatal. " 40
Contrary to the popular image of police work, a decision to
use deadly force against a fleeing suspect is a rare one for most
law enforcement officers. Yet, of all the decisions an officer is
called upon to make in emergency arrest situations, whether to
use deadly force can turn out to be the most agonizing and
tormenting of all. Officer Marshall's testimony about his deci­
sion to shoot at a fleeing felon, which led to the Connecticut
case of J01U3s v. Marshall, is a powerful example of the conflict­
ing emotions affecting an officer faced with a decision whether
to use deadly force. 41 In another case, the permanent paralysis
of a 15-year-old boy who was caught with a stolen car and the
distressed emotions of the defendant police officer following
the shooting emphasize the tragedy of the legal, but unwise,
use of deadly force. 42
Law enforcement personnel everywhere have a vital inter­
est in what constitutes the legal use of deadly force. Especially
is this true of administrators. They should follow any effort to
restrict its legal use, whether that restriction comes through
legislative reform, their own state court decisions, or continued
constitutional attack in federal courts. Beyond this, the admin­
istrator has a more difficult responsibility. He must decide
when the use of deadly· force is wise and prudent and support
that decision with clear policy and effective training.

84
 Notes
1. "Deadly force" as a term of legal art means force by an officer for
the purpose of causing, or which he knows creates a substantial risk of
causing, death or serious bodily injury. See Schumann v. McGinn, 240
N.W. 2d 525 n. 1 (Minn. 1976).
2. Plucknett, A Concise History of the Common Law, 424-54 (5th Ed.
1956).
3. Comment, Deadly Force to Arrest: Triggering Constitutional Re­
view, 11 Harv. Civ. Rights-Civ. Lib. L. Rev. 361 (1976) (hereinafter
cited as Comment, 11 Harv. Civ. Rights-Civ. Lib. L. Rev. 361).
4. Petrie v. Cartwright, 70 S.W. 297, 299 (Ky. 1902). See Note, The
Use of Deadly Force in Arizona by Police Officers, 1973 L . & Soc.
Order 481, 482.
5. Head v. Martin, 3 S.W. 622, 623 (Ky. Ct. App. 1887). See Pearson,
The Right to Kill in Making Arrests, 28 Mich. L. Rev. 957, 964 (1930).
6. Mattis v. Schnarr, 547 F. 2d 1007 (8th Cir. 1976),cert. filed 45
U.S.L.W. 3669.
7. Ashcroft v. Mattis, 45 U.S.L.W. 3751 (1977).
8. Note, The Use of Deadly Force in the Apprehension of Fugitives
From Arrest, 14 McGill L. J. 293 (1968). See also, Comments of Pro­
fessor John Barker Waite reported in Model Penal Code, Tent. Draft
No. 8 (1958), §3.07, p. 60.
9. Mattis v. Schnarr, supra note 6, at 1023 (Judge Gibson, dissent­
ing).
10. Jones v. Marshall, 528 F. 2d 132 (2nd Cir. 1975).
11. N.Y. Pen. Law §55.05(1) (1973).
12. Jones v. Marshall, supra note 10, at 142. Chief Justice Burger's
comment in Bivens v. Six Unknown Named Agents of Federal Bureau
ofNarcotics, 403 U.S. 388, 419 (1971) (dissenting opinion): "I wonder
what would be the judicial response to a police order authorizing ' shoot
to kill' with respect to every fugitive. It is easy to predict our collec­
tive wrath and outrage. We, in common with all rational minds, would
say that the police response must relate to the gravity and need; that a
'shoot' order might conceivably be tolerable to prevent the escape of a
convicted killer but surely not for a car thief, a pickpocket or a shop­
lifter."
13. 9 American Law Institute Proceedings 186-187 (1931), quoted in
Mattis v. Schnarr, supra note 6, at 1014.
14. ALI Proceeding (1958), p. 285; Schumann v. McGinn, supra note
1, at 541.

85
15. Jones v. Marshall, supra note 10, at 142.
16. The 24 states, according to Comment, 11 Harv. Civ . Lib. L. Rev.
at 368 n. 30, are: Alaska Stat. §11.15.090 (1970); Ariz. Rev. Stat. Ann.
§13-461 (Supp. 1972); Ark. Stat. Ann. §41-510(2)(a) (Spec. Pamphlet
1976); Cal. Penal Code §196 (West 1970); Colo. Rev. Stat. Ann. §18-1­
707(2)(b) (1973); Conn. Gen. Stat. §53a-22 (cX2) (1975); Fla. Stat. Ann.
§766.05 (Supp. 1975); Idaho Code §19-610 (1970); Ind. Code §35-1-19-3
(Burns 1975); Iowa Code §755.8 (1971); Kan. Stat. Ann. §21-3215(1)
(1974); Minn. Stat. §609-065(3) (1974); Miss. Code. Ann. §97-3-15
(1972); Mo. Rev. Stat. §559.040 (Vernon 1969); Mont. Rev. Code. Ann.
§94-2512 (Spec. Supp. 1973); Nev. Rev. Stat. §200-140(3)(b) (1973);
N.H. Rev. Stat. Ann. §627-5(11)(b)(1) (Supp. 1973); N.M. Stat. Ann .
§40A-2-7 (1963); Okla. Stat. Ann. tit. 21, §732 (1951); R. I. Gen. Laws
§12-7-9 (1969); S.D. Comp. Laws Ann. §22-16-32 (1967); Tenn. Code
Ann. §40-808 (1956); Wash. Rev. Code Ann. §9.48.160 (1961),
§9A.16.040(3) (1975) (effective July 1, 1976); Wis. Stat. §939.45(4)
(1973).
17. Mattis v. Schnarr, supra note 6, at 1022.
18. Mo. Rev. Stat. §554,190, §559,040 (1969).
19. Clark v. Ziedonis, 368 F. Supp. 544 (E.D. Wis. 1973).
20. Clark v. Ziedonis, 513 F.2d 79 (7th Cir. 1975).
21. According to Comment, 11 Harv. Civ. Rights-Civ. Lib. L. Rev.,
at 368 n. 31, they are: Ga. Code Ann. §26-902 (1972); Ill. Rev. Stat. ch.
38, §7-5(aX2) (1973); N.Y. Penal Law §35.30(1XaXii) (McKinney Supp.
1971); N.D. Cent. Code §12.1-05-07(2)(d) 1975; Ore. Rev. Stat. §161­
239 (1973); Pa. Stat. Ann. tit. 18, §508(aX1Xii) (1973); Utah Code Ann.
§76-2-404(2)(b) (Supp. 1975).
22. Harvie, The Police Officer's Use of Force: Law and Liability 10,
published by Police Training Institute, University of Ill. (undated).
23. Model Penal Code, §§3.04-3.11 (Proposed Official Draft, 1962).
24. According to Comment, 11 Harv. Civ. Rights-Civ. Lib. L. Rev.
at 369 n. 32, these states are: Del. Code Ann. tit. 11 §467(c) (1974);
Hawaii Laws, Act 9, ch. 3 (1972) (effective 1973) §307(3); Ky. Rev.
Stat. §503.090(2) (1975); Maine Rev. Stat. Ann. tit. 17A §107-2(B)
(1975) (effective March 2, 1976); Neb. Rev. Stat. §28-839(3) (Supp.
1974); Tex. Penal Code art. 2, §9.51(c) (1974); North Carolina allows
the use of deadly force to arrest one fleeing from a felony with a
deadly weapon in addition to those situations in which the Model Penal
Code formulation authorizes deadly force. N.C. Gen. Stat. §15A­
401(d)(2)(b) (1973).

86
25. Note, Justification: The Impact of the Model Penal Code on Statu­
tory Reform, 75 Column. L. Rev. 914, 955 (1975).
26. Paul v. Davis, 47 L. Ed. 2d 405 (1976).
27. Jenkins v. A verett, 424 F. 2d 1228, 1232 (4th Cir. 1970). See also
Reed v. Philadelphia Housing Authority, 372 F. Supp. 686 (E.D. Pa.
1974).
28. Cunningham v. Ellington, supra note 39.
29. Jones v. Marshall, supra note 10; Mattis v. Schnarr, supra note 6.
30. Conn. Gen. Stat., §53a-22.
31. Mo. Rev. Stat., supra note 18.
32. Jones v. Marshall, supra note 10, at 142. See also Wiley v. Mem­
phis Police Department, 548 F. 2d 1247 (6th Cir. 1977}; Wolfer v.
Thaler, 525 F. 2d 977 (5th Cir. 1976), cert. denied 425 U.S. 975 (1976);
Hilton v. State, 348 A. 2d 242 (Me. 1975).
33. Mattis v. Schnarr, supra note 6, at 1020.
34. Mattis v. Schnarr, 404 F . Supp. 643, 651 (E.D.) Mo. 1975).
35. Mattis v. Schnarr, supra note 6, at 1020.
36. A shcroft v. Mattis, supra note 7.
37. Jones v. Marshall, supra note 10, at 140.
38. Police Foundation, "Police Use of Deadly Force" (1977).
39. ld. at 141.
40. Coroner's Report in Jones v. Marshall, 383 F. Supp. 358 (D. Conn.
1974), affd 528 F. 2d 132 (2nd Cir. 1975), reported in Goldstein,
Dershowitz, & Schwartz, Criminal Law: Theory and Process 331
(1974).
41. Goldstein, Dershowitz, and Schwartz, Criminal Law: Theory and
Process 327-30 (1974).
4~. Schumann v. McGinn, supra note 2, at 541.

87
 CHAPTER 4

EXECUTION WITHOUT TRIAL: POLICE

HOMICIDE AND THE CONSTITUTION*

LAWRENCE W. SHERMAN
The national debate over the State's right to take life has
been sidetracked, in a sense, on the issue of "capital punish­
ment," or more precisely, execution after trial. Far more
deadly in impact is the body of law permitting execution with­
out trial through justified homicide by police officers. In 1976,
for example, no one was executed and 233 persons were sen­
tenced to death after trial, yet an estimated 590 persons were
killed by police officers justifiably without trial. 1 Even in the
1950s, when an average of seventy-two persons were executed
after trial each year/ the average number of police homicides
was 240 a year, according to official statistics,3 and 480 a year
according to one unofficial estimate. 4 Since record keeping be­
gan in 1949, police actions have been by far the most frequent
method with which our government has intentionally taken the
lives of its own citizens.
The significance of police homicide is not, however, derived
solely from its frequency. Equally important is the nature of

*This writing was supported in part by the National Institute of

Mental Health, Center for Studies in Crime and Delinquency. Fred
Cohen, David Wukitsch, Michael Gottfredson, Eva Sherman, Mark
Blumberg, and Robert Langworthy contributed ideas and assistance.

SOURCE: "Execution Without Trial: Police Homicide and the Constitution,"

Vanderbilt Law Review (January 1980), 33:71:71-100. Copyright 1980 by Van­
derbilt Law Review.

88
the crimes that justify police use of deadly force. Unlike execu­
tions after trial, executions before trial are not limited to ex­
tremely serious crimes such as murder, rape, and treason.
Twenty-four states follow what is thought to be the traditional
common-law doctrine, which permits the use of deadly force
whenever necessary to prevent a felony or to arrest someone
whom an officer has reasonable grounds to believe has commit­
ted a felony5 -any felony , including, in at least one state, spit­
ting on a policeman.6 Eight states have adopted the more re­
stricted version of this common-law doctrine proposed by the
Model Penal Code; 7 ten other states have adopted statutes
allowing police to use deadly force to arrest suspects of "vio­
lent" or "forcible" felonies, 8 which in some states may include
burglary.9 Even under these relatively recent restrictions, most
police officers are still legally empowered to shoot unarmed
fleeing burglary suspects in the back.
The available evidence suggests that when the police do use
deadly force, their targets are often suspects of less serious
crimes.10 Approximately half of the people at whom police shots
were fired in the several cities studied have not carried guns,
and the proportion of those shot while fleeing is substantial. 11
To be sure, many police homicides occur in defense of life,
although the data are not precise enough to determine how
many. There is no doubt, however, that many executions with­
out trial occur in response to crimes against property without
any defense justification.
A review of the legal history of police homicide shows that
the rule that any felony warrants the use of deadly force is a
common law anachronism to which our courts and legislatures
continue to cling long after the Crown Courts have treated the
doctrine as dead and Parliament has laid it to rest through
criminal law reform. More important, an analysis of the consti­
tutional status of the any-felony rule shows that it should be
held to violate the due process clause of the fifth amendment,
the ban on cruel and unusual punishment of the eighth amend­
ment, and the equal protection clause of the fourteenth amend­
ment. Both the historical and constitutional lines of inquiry sug­

89
gest that only the defense-of-life doctrine is appropriate to
govern police use of deadly force.

I. THE ANY-FELONY RULE: AN HISTORIC

ANACHRONISM
The original meaning of the common-law justification for
homicide to effect a felony arrest was very different from its
current meaning. A barbaric legal doctrine 12 transplanted to
England before the common law began, 13 the justification arose
at a time when (1) there were no accurate and reliable weapons
available that could kill at any distance, (2) the label "felony"
was reserved for only the most serious crimes, all of which
were punishable by death, and (3) there was virtually no com­
munication among law enforcement officers in different com­
munities. Each of these three elements of the historical context
has changed drastically over the centuries, and with it the prac­
tical meaning of the doctrine.
The medieval weaponry used in "hue and cry" 14 during the
early years of the any-felony rule was apparently limited to
knives, swords, farm tools, and halberds. The longbow was not
introduced until 1415, 16 and in 1504 the Tudors restricted the
crossbow to lords and large landowners.15 Henry VIII allowed
noblemen and wealthy commoners to own guns, 17 but "[t]he
musket of Shakespeare's time could not reach an enemy
thoughtless enough to stand farther than eighty or ninety yards
away." 18 A "typical" London street brawl in the reign of Henry
VIII was put down by a band of constables, none of whom were
armed with any weapons other than those used in hand to hand
combat . 19 In this technological context, then, the practical
meaning of the deadly force doctrine was that suspects could be
killed if they resisted in a hand to hand struggle, but it did not
mean that they could be killed from a distance behind while
they were in flight.
That meaning changed in the nineteenth century with the
invention of the revolver. Police officers in large American cit­
ies, who had been disarmed since the decline of Indian attacks
before the Revolutionary War, began to carry revolvers in the

90
1850s after criminals used revolvers to shoot and kill their col­
leagues.20 The dumping of thousands of army revolvers on the
surplus market after the Civil War speeded the general rearma­
ment of an increasingly violent urban society21 .and led to official
acceptance of police use of revolvers. 22 The immediate effect of
this change was that the police could, and did, shoot fleeing
suspects who were posing no immediate threat to anyone.
The effect of the revolution in weaponry on police homicide
was compounded by the expansion in the scope of felonies .
Originally reserved under the common law for felonious homi­
cide, mayhem, arson, rape, robbery, burglary, larceny, prison
breach, and rescue of a felon, all punishable by death,23 the
felony label was attached to many more crimes after the advent
of the revolver. 24 Moreover, while the scope of felonies was
expanding, the scope of capital felonies contracted, leaving the
death penalty in most states only applicable to treason and
crimes endangering life or bodily security. 25 These changes in
the legal context of police homicide significantly altered the
meaning of the common-law any-felony doctrine. The changes
greatly expanded the number of situations in which the police
could kill without trial, and they created a gross difference in
proportion between the severity of the post-trial penalty and
the severity of the penalty for attempting to escape arrest.
While advances in weapon technology and changes in the
criminal law were expanding the scope and potency of the any­
felony rule, one of the primary reasons for its existence was
fading. By the late nineteenth century, the rise of bureaucratic
police agencies with the capacity to communicate information
about suspects at large was undermining the necessity for the
use of deadly force in the apprehension of felons. The escaping
suspect of eleventh-century England might establish a new life
in another community with little fear of eventual capture, and
the social goal of retribution was thus easily frustrated by a
fleeing felon. By the eighteenth century, however, Justice
Fielding was circulating descriptions of wanted criminals out­
side of London,Z6 and by the early twentieth century American
detectives consulted their colleagues in other cities about vari­

91
ous thieves and their whereabouts. 27 The effect of the increas­
ingly sophisticated apprehension techniques meant that it was
no longer absolutely necessary to kill a suspect, if his identity
were known , in order to insure his eventual capture.
These changes in the scope and impact of the any-felony
doctrine did not escape public notice and criticism. An 1858
New York Times editorial questioned one of the first police
shootings there, making a value judgment supported by the
constitutional analysis below. The Times suggested, "if a police­
man needed to defend his life , the use of force was permissible,
but if he was chasing a suspect, he had no right to shoot the
man. A policeman either had to be swift enough to catch the
suspect or justice must be lost. " 28 Another Times editorial the
same year expressed grave concern about a possible future in
which "[e)very policeman is to be an absolute monarch, within
his beat, with complete power of life and death over all within
his range , and armed with revolvers to execute his decrees on
the instant, without even the forms of trial or legal inquiry of
any kind ," 29 a future that, to a large extent, has been realized.
These changes did not escape the notice of the courts. As
early as 1888 the Supreme Court of Alabama, observing the
legislative inflation of crimes to felony status, pronounced that
"the preservation of human life is of more importance than the
protection of property." The court restricted the common-law
rule by disallowing deadly force in the prevention of secret
felonies not accompanied by force. 30 Several other decisions
grappled with the obsolete common-law standard / 1 but gener­
ally the courts were, as one commentator noted, "reluctant to
abandon a convenient pigeon-hole disposal of cases on the basis
of whether the crime was a felony or a misdemeanor."32
Meanwhile, the English common law had already effec­
tively abandoned the absolute right to kill to prevent felonies or
apprehend felons. It replaced the any-felony doctrine with a
balancing test emphasizing necessity and proportion:
The circumstances in which it can be considered reason­
able to kill another in the prevention of crime must be
of an extreme kind; they could probably arise only in

92
 the case of an attack against a person which is likely to
cause death or serious bodil:y injury and where killing
the attacker is the only practicable means of preventing
the harm. It cannot be reasonable to kill another merely
to prevent a crime, which is directed only against prop­
erty.33
This principle was so well established in case Jaw that by 1879
the Criminal Code Bill Commission took it as a "great principle
of the common law" that the "mischief done by [the use of
force to prevent crimes should not be] disproportioned to the
injury or mischief which it is intended to prevent." 3' Moreover,
a close reading of the original common-law codifiers Foster,
Blackstone, Hawkins, and East reveals so many internal con­
tradictions and exceptions to the right to kill all felons35 that
one may question whether there ever was such a rule. Thus, in
1965 the Criminal Law Revision Committee reported to Parlia­
ment that despite "old authority" for the right to kill all felons,
"the matter is very obscure; . . . owing no doubt to the re­
straint of the police there is a dearth of modern authority on
it;" and concluded that their central proposal to reclassify
crimes would have no effect on police powers since "the likeli­
hood that anything would turn nowadays on the distinction be­
tween felony and misdemeanor is very slight. " 36
In this country, however, the use of the distinction re­
mained anything but slight. As recently as 1977 the Sixth Cir­
cuit upheld a Tennessee statute under which the Memphis po­
lice shot and killed a sixteen-year-old burglary suspect fleeing
from a hardware store. 37 Noting that "the legislative bodies
have a clear state interest in enacting laws to protect their own
citizens against felons," and that the statute "merely embodied
the common law which has been in force for centuries and has
been universally recognized" 38 (something that we have seen is
clearly not the case in English common law), the court rejected
a broad constitutional challenge to the statute. An argument
that the statute violated the eighth amendment's ban on cruel
and unusual punishment was rejected on the grounds that po­
lice homicide is not "punishment. " 39 The assertion that the stat­
ute violated due process protections was rejected on the

93
grounds that state interests served by police homicide were
more important that an individual's right to trial before being
killed by police. 40 While recognizing that the Eighth Circuit had
recently held that a similar Missouri statute did violate fifth
and fourteenth amendment due process guarantees, 41 the Sixth
Circuit criticized that decision for intruding into legislative mat­
ters.42Finally, the Sixth Circuit case dismissed a claim of racial
discrimination in violation of the fourteenth amendment be­
cause "both white and black fleeing felons ... have been fired
upon or shot by Memphis police."' 3 The Supreme Court denied
certiorari."
The Sixth Circuit's cursory treatment of the threshold issue
of whether police homicide constitutes punishment, however, is
hardly definitive. Measured against well established Supreme
Court standards, police homicide clearly constitutes punish­
ment. When police homicide is viewed as punishment, the fifth,
fourteenth, and eighth amendment arguments that all present
police homicide statutes and case law are constitutionally un­
sound are much more compelling.

II. CONSTITUTIONAL ANALYSIS

A. The Characterization ofPolice Homicide as

Punishment

The often elusive definition of punishment in philosophy

and jurisprudence has been a "major obsession with the Eng­
lish linguistic philosophers of this century. " 45 The definitions
vary sharply, with distinctions focusing upon the intent of the
putative punisher, or the purpose of inflicting pain or suffering.<6
As the recent ruling in Bell v. Wolfi sh47 reveals, the issue of
intent has likewise proved to be divisive in the Supreme Court's
efforts to define deprivations that constitute punishment. Jus­
tice Rehnquist, delivering the opinion of the Court, held that in
determining whether particular conditions accompanying pre­
trial detention amount to punishment in the constitutional
sense a "court must decide whether the disability is imposed for

94
the purpose of punishment or whether it is but an incident of
some other legitimate governmental purpose. " 48 "Absent a
showing of an expressed intent to punish," Justice Rehnquist
continued, "that determination will turn on 'whether an alter­
native purpose to which [the restriction] may rationally be con­
nected is assignable for it, and whether it appears excessive in
relation to the alternative purpose assigned [to it],' " 49 (quoting
Kennedy v. Mendoza-Martinez,S0 apparently as the controlling
case on the subject). Justice Stevens, however, pointed out in
his dissent that the Mendoza Court also recognized that evi­
dence of intent would sometimes be "unavailable or untrust­
worthy."51 "In such cases," Justice Stevens said "the [Mendoza]
Court stated that certain other 'criteria' must be applied 'to the
face' of the official action to determine if it is punitive." 52 Even
Justice Rehnquist, whose opinion in Bell v. Wolfish reveals a
very restrictive conception of what constitutes punishment,
cited the seven Mendoza criteria approvingly. Although he did
not, as Justice Marshall pointed out,53 make full use of them, he
nonetheless refers to them as "useful guideposts in determin­
ing" what is punishment, calling them "the tests traditionally
applied to determine whether a governmental act is punitive in
nature. " 54
With the original intent of the Gothic chieftains in estab­
lishing the kill-to-arrest rule lost in history, and determination
of the subjective intent of police officers acting within the rule
vulnerable to "hypocrisy and unconscious self-deception," 55 it is
necessary to turn to the criteria used in Mendoza and apply
them "to the face" of police homicide to determine whether
that action constitutes punishment. The decision offered seven
criteria:
[1] Whether the sanction involves an affirmative disability
or restraint,
[2] whether it has historically been regarded as a punish­
ment,
[3] whether it comes into play only on a finding of scienter,
[4] whether its operation will promote the traditional aims
of punishment- retribution and deterrence,

95
 [5] whether the behavior to which it applies is already a
crime,
[6] whether an alternative purpose to which it may ration­
ally be connected is assignable for it, and
[7] whether it appears excessive in relation to the alterna­
tive purpose assigned ....56
The Mendoza Court noted that all of these criteria are relevant
to the inquiry, although they "may often point in differing di­
rections. " 57 All seven criteria, however, suggest that police
homicide constitutes punishment, as is clear when each crite­
rion is examined.
(1) Whether the sanction involves an affirmative disability
or restraint. Recent pronouncements by the Court leave no
doubt that the sanction of police homicide constitutes "an af­
firmative disability or restraint." It is not only a deprivation of
rights, but a deprivation of "the right to have rights, " !\8 not
only a sanction, but a "unique" sanction. As Justice Brennan
stated, "[i]n a society that so strongly affirms the sanctity of
life, ... the common view is that death is the ultimate sanc­
tion."59 Five members of the present Court have "expressly
recognized that death is a different kind of punishment from
any other which may be imposed in this country" and stated
that "[f]rom the point of view of the defendant, it is different in
both its severity and finality. From the point of view of society,
the action of the sovereign in taking the life of one of its citi­
zens also differs dramatically from any other legitimate state
action. " 60 The right to life has consistently been held fundamen­
tal and preeminent. 61 Its deprivation has the same effect no
matter what the expressed purpose may be.
(2) Whether it has historically been regarded as punish­
ment. The historical record clearly demonstrates that execu­
tions without trial, including the kill-to-arrest doctrine, were
generally viewed as punishment. Thieves were often killed out­
right during the hue and cry, even after they had been cap­
tured. "Let all go forth where God may direct them to go,"
urged the tenth-century laws of Edgar; "Let them do justice on
the thief. " 62 Suspicion sufficed to convict thieves without any

96
 trial at all, and "execution in such cases often followed immedi­
ately on arrest. " 63 According to the preamble to Act 24 of
Henry VIII, it appears that the common law authorized the
victims of crimes and attempted crimes to kill the criminal,
regardless of whether it was necessary to prevent the felony. 6'
In the twelfth and thirteenth centuries "outlaws could be be­
headed by anyone, and a reward was paid for their heads under
Richard I. " 65 Abjurors of the realm (felons who had escaped
into religious sanctuary and agreed to leave the country for­
ever) who strayed from the highway on their journey to the sea
could also be beheaded by anyone .66 In the context of the times
in which the kill-to-arrest doctrine evolved, it was clearly linked
to a philosophy of summary justice that can only be viewed as
punishment.
Modern commentators have taken the same view of the
historical status of the doctrine. Professor Perkins notes that
"as the felon had forfeited his life by the perpetration of his
crime, it was quite logical to authorize the use of deadly force. " 67
Another commentator on killing fleeing felons described "the
extirpation [as] but a premature execution of the inevitable
judgment" in the era of capital punishment for all felonies. 68
With the passing of that era, premature execution is of course
more severe than the "inevitable judgment." The historical
character of police homicide as punishment, however, is not
altered by the modern disproportion between pretrial and post­
trial sanctions.
(3) Whether it comes into play only on a finding of scien­
ter. The basis and parameters of the Mendoza Court's "scien­
ter" criterion are unclear. Of the two cases cited to support the
relevance of scienter to a punishment characterization,69 one in
fact holds that penalties may constitute punishment regardless
of scienter, apparently contradicting the point for which it was
cited. The holding stated that, regardless of scienter, any fine
imposed on an import merchant for underestimating the value
of certain goods was "still punishment and nothing else." 70 The
other case cited in Mendoza only mentions in passing that the
exemption from a federal child labor "tax" of employers who

97
do not know that their workers are underage suggests that the
tax is really a penalty. The Court in that case opined that,
"[S]cienter is associated with penalties, not with taxes." 71 Nei­
ther case actually holds that punishment is only imposed after
finding a scienter.
The apparent contradictions notwithstanding, the Supreme
Court has held that "the general rule at common law was that
scienter was a necessary element . . . of every crime ." 72Regard­
less of criticisms of this usage, 73 one may proceed from it to
infer that when an officer finds sufficient cause to believe some­
one is a felon and thus has met a requisite justification for
killing him, the officer finds scienter at the same time. If the
officer does not have probable cause to believe that scienter is
present, then he does not have probable cause to believe the
person is a felon, and killing is not justified. Justified police
homicide therefore historically presumes scienter, and satisfies
the apparent meaning of this Mendoza criterion of punishment.
(4) Whether its operation will promote traditional aims of
punishment-retribution and deterrence. Police homicide
clearly promotes retribution, the first of the two "traditional
aims of punishment" named by the Mendoza Court. As the
dissent in Mattis v. Schnarr, 74 a recent Eighth Circuit decision
argued in support of the any-felony rule, which the court had
found unconstitutional, "[t]here is no constitutional right to
commit felonious offenses and to escape the consequences of
those offenses." In that context, "consequences" strongly im­
plies "just desserts, " or retribution.
Whether police homicide, or indeed any punishment, actu­
ally promotes deterrence, the second of the two traditional
aims named, may be an impossible question to answer.75 If un­
disputed empirical evidence of a deterrent effect is required to
evaluate whether a sanction is a punishment, then many social
scientists would argue that few sanctions qualify. If, on the
other hand, a deterrent effect need only be hypothesized for the
sanction to be a punishment, then police homicide passes the
test. The assumption by legal scholars that police homicide has
a deterrent effect is reflected in the American Law Institute's

98
 debates over the issue. The deterrence of flight from arrest7o;
and the deterrence of robbery77 were both specifically men­
tioned, albeit with differences of opinion. The deterrence hy­
pothesis is also implied in recent federal cases, such as Jones v.
Marshall, 18 a Second Circuit opinion in which a three-judge
panel upheld Connecticut's common law permitting police to
kill fleeing felons, observing that the states had the right to
place a higher value on order than on the rights of suspects.
The only way such a homicide could achieve order is through
deterrence .
(5) Whether the behavior to which it applies is already a
crime. All of the behavior to which police homicide applies is
already a crime, or the officer must reasonably believe it to be a
crime. There is however, some question about which crime po­
lice homicide is punishing. As Professor Mikell asked in his
often quoted statement to the American Law Institute:
May I ask what we are killing [the suspect] for when he
steals an automobile and runs off with it? Are we killing
him for stealing the automobile? . . . It cannot be ...
that we allow the officer to kill him because he stole the
automobile, because the statute provides only three
years in a penitentiary for that. Is it then ... for flee­
mg that we kill him? Fleeing from arrest . .. is punish­
able by a light penalty a penalty much less than that
for stealing the automobile. If we are not killing him for
stealing the automobile and not killing him for fleeing,
what are we killing him for?79
No matter how little sense it makes in relation to the post-trial
penalty, we are in fact killing the auto thief for the volatile
combination of felony and flight, both of which are crimes.
(6) Whether an alternative purpose to which it may ration­
ally be connected is assignable for it. The purposes of capture
and crime prevention, rather than punishment, may no doubt
be rationally connected to police homicide as alternatives to the
purpose of punishment. Just as the Wolfish Court held that
overcrowding and other disabilities. imposed on pretrial detain­
ees in a federal jail did not constitute punishment because they
were merely an "inherent incident" of t he objective of insuring

99
detainee's presence at trial, it could be argued that death is
merely an inherent incident to insuring that felony suspects are
captured and that felonies are prevented. By this logic, death
from police homicide is not a punishment if the expressed in­
tent of the officers using deadly force is to apprehend felony
suspects.
An equally strong case, however, could be made that the
presence of multiple purposes in a governmental action does
not automatically grant preeminence to the non-punitive pur­
pose . One purpose of prison systems in some states is the man­
ufacture of license plates, but a penitentiary sentence could
hardly be described as merely an inherent incident of a legiti­
mate state interest in manufacturing license plates. Implicit in
the Wolfish Court's reasoning is a judgment about the primary
purpose of any governmental action that has more than one
purpose. Punishment rather than apprehension can be judged
the primary purpose of police homicide. As one court once
noted, "[t]he reason for ... killing felons .. . in attempts to ar­
rest them ... is obvious ... [T]he safety and security of society
require the speedy arrest and punishment of a felon." 81
Unlike the other Mendoza criteria, this one is explicitly
qualified by the succeeding criterion, which questions whether
the possible alternative purpose to punishment appears exces­
sive. No matter what the primary purpose of police homicide is
judged to be, then, if it appears excessive in relation to a non­
punitive purpose, it must be defined as punishment. As Justice
Stevens interprets Mendoza in his Wolfish dissent, "when there
is a significant and unnecessary disparity between the severity
of the harm to the individual and the demonstrated importance
of the regulatory objective, ... courts must be justified in
drawing an inference of punishment. " 82
(7) Whether it appears excessive in relation to the alterna­
tive purpose assigned to it. The disparity between the death of a
suspect and the purposes of prevention (of nonviolent crimes)
and capture is both significant and unnecessary, and therefore
excessive in relation to those purposes. It is significant in the
case of capture because, once again, the means used to prevent

100
the suspect's escape is far more severe than the maximum pen­
alty that would be imposed upon sentencing for all crimes (de­
pending on the jurisdiction) except murder, treason, and rape.
It is significant in the case or prevention of nonviolent crimes
because the evil imposed is greater than the evil presented. It is
unnecessary in the case of capture because most suspects can
eventually be recaptured, and in the case of prevention because
nonlethal intervention is usually possible. A sanction that takes
a life to prevent the theft of an ear of corn83 or a chicken84
cannot, in a society that values life, be other than excessive.
Each of the Mendoza criteria point to the conclusion that
the use of deadly force to capture felons and prevent felonies
constitutes punishment, and is therefore subject to the constitu­
tional restraints on the use of punishment. Even if it were ruled
not to be punishment, however, it is still a deprivation of rights
subject to the due process requirements of the fifth and four­
teenth amendments . Although a ruling that police homicide
constitutes punishment has the added advantage of subjecting
it to eighth amendment review, that review is generally
reached only after due process guaranties have been satisfied. 85
In the case of police homicide, the due process guaranties are
anything but satisfied.
B. Due Process Requirements
Although police homicide raises serious due process ques­
tions if viewed merely as a deprivation of rights, when rec­
ognized as ·punishment its apparent violation of due process
guaranties is striking. The framers "intended to safeguard the
people of this country from punishment without trial by duly
constituted courts, " 86 and "under the due process clause, a de­
tainee may not be punished prior to an adjudication of guilt in
accordance with due process of law." 8 7 The limitation on impos­
ing death, under the fifth amendment, is particularly strict. It
requires that "[n]o person shall be held to answer for a capital,
or otherwise infamous crime unless on a presentment or indict­
ment of a grand jury." Indeed, the Eighth Circuit observed that
a literal reading of the due process clause would mean that " life

101
could never be taken without a trial. " 88 And that is precisely
what it should mean, with respect to life taken under the au­
thority exercised on behalf of the State. A less rigid standard,
however, must be applied when deadly force is used by an indi­
vidual in accordance with the self-defense doctrine. 89 In addition
to personal defense, this doctrine includes the defense of "an­
other person against what is reasonably perceived as an imme­
diate danger of death or grievous bodily harm to that person
from his assailant." 90
The Eighth Circuit, the only circuit to hold that the any­
felony rule violates the fourteenth amendment, finds this inter­
pretation too extreme. "Such a literal reading," it stated,
"would fail to recognize the interests of the state in protecting
the lives and safety of its citizens," and therefore the court held
that the situations in which the State can take a life without
according a trial to the person whose life is taken are to be
determined by balancing society's interest in public safety
against the right to life of an individual. 91 Irrespective of their
conclusion, the use of the balancing test is a fundamentally
flawed procedure for determining whether the right to a form
of due process specified in the Constitution is applicable. The
fifth amendment does not depend upon a showing that it is in
the community's best interests that the procedures be ac­
corded.92 As Professor Dworkin has observed, "a right against
the Government must be a right to do something even when the
majority would be worse off for having it done." 93 The majority
is no doubt worse off whenever a fleeing felon escapes, but that
should not alter the felon's fifth amendment right to grand jury
review and trial before he is executed.
The balancing test is, however, the prevailing method of
determining how much process is due once it is determined that
due process applies. 94 Although the severity of individual depri­
vation and the relative importance of governmental interest in
summary action is arguably incommensurable, 95 even a balanc­
ing procedure should lead reasonable men and women to a
more restrictive scope of executions without trial. Both the
fifth and fourteenth amendments specifically forbid deprivation

102
of life without due process of law, so there is no question that
some process is due. The issue of when to allow executions
without the due process of trial must then balance the individ­
ual's fundamental right to life96 and the right not to be deprived
of life without the due process of trial 97 against the state's inter­
est, not just the interest in general public safety, but its narrow
interest in protecting the property and Jives of other specific
individuals. We have long since decided that life is more impor­
tant than property, and that no property offender, no matter
how serious or recidivistic, may be executed after trial for his
offenses. It should follow that the state's interest in protecting
the property of others is not compelling enough to allow execu­
tion without the due process of trial.
The state's interest in protecting the lives and bodies of
other individuals is, however, far more compelling, and much
more appropriate for a balancing test. 98 When someone poses
an immediate threat of grievous injury to another, the use of a
balancing test would lead to the conclusion that the state's
interest in protecting the other person allows it to commit an
execution without the due process of a trial. It is not necessary,
however, to adopt the balancing test procedure in order to con­
clude that police officers may kill in defense of life. The self­
defense doctrine gives them that power as individuals irrespec­
tive of their association with the state . Th e police can kill those
posing an immediate threat of violence without violating the
fifth amendment rights of those killed, just as any citizen can .
The legitimate concern some courts have shown with police
officer's safety99 can, acco rdingly, be satisfied without a fleeing­
felon or any-felony rule. If a fleeing felon whom the officer
reasonably believes to be armed turns toward the pursuing offi­
cer with reasonably apparent intent to shoot the officer, the
officer may kill him under the self-defense doctrine. The
fleeing-felon rule in no way increases the officer's safety be­
yond the safeguard of the self-defense rule.
If a balancing test is used, however, the final and most
difficult problem is to assess the state's interest in insuring
public safety. An escaped felony suspect is certainly free to

103
commit other crimes, but that should not be a compelling justi­
fication for the use of deadly force. A released convict who has
served a full penitentiary sentence may be equally likely to
commit more crimes, but that justifies neither his execution nor
his incarceration beyond the end of his sentence. Far more
compelling is the deterrence argument that the failure to kill
fleeing felons will encourage more felonies. No empirical at­
tempt to evaluate this argument has been made to date, but let
us assume, arguendo, that each police homicide prevents eight
or even eighty, robberies. Are we to measure the value of life in
such utilitarian terms? Is it a lesser evil that a life be lost than
several hundreds or thousands of dollars be stolen? In a society
that punishes million-dollar white-collar frauds with a four­
month prison term, it seems difficult to answer that question
affirmatively.
Our primary concern, however, is with the Constitution,
and not with the greatest good for the greatest number. Even
if a balancing test determined that the state's interest in main­
taining public safety allowed it to execute fleeing and in­
progress felons without trial under the due process require­
ments of the fifth and fourteenth amendments, those
executions could still be ruled unconstitutional as either cruel
and unusual punishment under the eighth amendment, or a
denial of equal protection under the fourteenth amendment.

C. Police Homicide as Cruel and Unusual Punishment

The lack of guidance on the framers' intent in banning cruel
and unusual punishment makes that phrase difficult to define
precisely. 101 Nonetheless, four criteria for judging whether a
given punishment is cruel and unusual can be clearly discerned
in Furman v. Georgia102 and its predecessor cases. The criteria
are whether the penalty is (1) inherently cruel, 103 (2) dispropor­
tionately severe to the offense it punishes, 10' (3) unacceptable to
contemporary society, 105 or (4) inflicted arbitrarily. 106 None of
the four seems to have been overruled in the death penalty
cases since Furman, and all but the third are specifically ad­
dressed in the opinion of the Court-a consensus the Furman

104
Court lacked-in Gregg v. Georgia. 101 Any of the four criteria
can make a punishment cruel and unusual. Police homicide sat­
isfies at least three, and on occasion all four criteria.
(1) Inherent cruelty. The present Court has consistently
held that death is not, per se, an unconstitutional punishment.10!!
Previous courts have, however, considered whether particular
modes of inflicting death are unconstitutionally cruel. 109 Shoot­
ing and electrocution have both withstood challenges, but it is
doubtful that any court would uphold death inflicted by a sus­
tained beating after a suspect has been subdued, 110 or by a
drowning or a choke-hold. 111 Nonetheless, police have used all
three methods to kill suspects in cases that have received wide­
spread attention, and have sometimes received light penalties
for doing so. Yet most police homicides do not receive much
attention or review. 11 2 Under the present any-felony rule, prose­
cutors are on firm ground for declining to prosecute police offi­
cers who beat felony suspects to death when the beating is
necessary to effect an arrest. Unless such action can be justi­
fied by the self-defense doctrine, it would seem to be an inher­
ently cruel and unusual form of punishment.
(2) Disproportionate severity. The determination whether
a punishment is proportionately severe to the crime it punishes
is essentially a moral judgment, not based on objective assess­
ments of the necessity or efficacy of the penalty imposed. 11 3
When judged in accord with contemporary standards, police
homicide is "grossly out of proportion to the severity" 114 of
most of the crimes it punishes. 115 As a former Oakland, Califor­
nia, police chief graphically explained when restricting his offi­
cers' right to shoot fleeing burglars beyond the state law's limi­
tations:
Considering that only 7.65 percent of all adult burglars
arrested and only .28 percent of all juvenile burglars
arrested are eventually incarcerated, it is difficult to
resist the conclusion that the use of deadly force to
aiJprehend burg-lars cannot conceivably be justified. For
adUlts, the pohce would have to shoot 100 burglars in
order to have captured the eight who would have gone
to prison. For juveniles, the police would have to shoot

105
 1,000 burglars in order to have captured the three who
would have gone to the Youth Authority. 116
Comparisons to actual punishments typically imposed after
trial would probably show that killing a fleeing suspect of any
crime, even murder, would impose a more severe punishment
without trial than could be expected after conviction . In the
case of murder, treason, and rape, a state's decision to make
available the death penalty for post-trial punishment might
mean that pretrial execution would not be disproportionately
severe. But murder and rape do not even appear as categories
in most studies of police use of deadly force, since they com­
prise such a small percentage of all crimes punished by police
homicide . Under the proportional severity test used for the past
century in English law, which embodies social values quite simi­
lar to our own, even fleeing murderers could probably not be
killed justifiably in order to arrest them once they no longer
posed an immediate threat of violence. 117
When analyzed from a utilitarian perspective, police homi­
cide is as disproportionately severe as it is when evaluated by
moral standards as a punishment. 118 Assuming that prevention
of escape is the utilitarian goal served by police homicide, the
fact that modern apprehension techniques have diminished con­
siderably the importance of immediate capture leaves police
homicide disproportionately severe in relation to the utilitarian
purposes it might serve. Whether viewed as a punishment or a
method of capture, the severity of police homicide is dispropor­
tionate to its objective.
(3) Lack ofacceptability in contemporary ·Society. Although
police homicide in arresting serious felons did not shock the
conscienceu9 of medieval England, the eighth amendment must
be interpreted in light of the evolving standards of a maturing
society. 120 Three of four available objective indicators/21 police
department administrative policies, scholarly opinion, and mass
public protests, show a considerable evolution in the attitudes
toward police homicide in recent years. A fourth indicator, leg­
islative authorization, lags behind the others, but that alone
does not demonstrate the acceptability of police homicide to

106
 society. Moreover, even the legislative arena has markedly
changed its approach toward police homicide over the past dec­
ade.
Until quite recently, police department policies were either
vague or silent on the use of deadly force, 122 but that is rapidly
changing. Since 1977, police policies in Los Angeles, Birming­
ham, and Houston, among others, have restricted the use of
deadly force far beyond the limits of state law. Los Angeles
adopted a modified defense-of-life policy after officers shot and
killed a naked chemist. 123 Houston reportedly adopted a
defense-of-life policy in the wake of the beating and drowning
of a young Chicano male. 124 Birmingham adopted a more restric­
tive policy after a Police Foundation study of seven cities
showed Birmingham to have the highest police shooting
rate 125 -the public outcry over which lends some support to Jus­
tice Marshall's hypothesis that the public is more likely to find a
punishment unacceptable when it knows the full facts. 126
Police policies more restrictive than state law are far from
new, however. A 1974 study of the Boston Police Department
found that the majority of the large cities surveyed permitted
their officers to use deadly force only to apprehend suspects
who present a threat of serious injury or death to someone. 127 In
1975 the California Peace Officer's Association and the Califor­
nia Police Chiefs' Association jointly adopted a similar policy. 128
The policy of the Federal Bureau of Investigation since at least
1972 has been "that an agent is not to shoot any person except,
when necessary, in self-defense, that is, when he reasonably
believes that he or another is in danger of death or grievous
bodily harm ." 129 The federal Bureau of Narcotics and Danger­
ous Drugs, which operates one of the most hazardous types of
law enforcement programs,t30 adopted a similar policy in 1971. 131
These policies were preceded by some fifty years of nearly
unanimous scholarly criticism of the any-felony rule. Law re­
views, 132 professional police publications, 133 and a Presidential
commission 134 all lobbied for a change in the rule. A more pow­
erful force for change, however, has been the long series of
public protests-often violent-over police use of deadly force

107
in minority communities. In the 1960s, several race riots were
precipitated by police shootings. 136 In the 1970s, police homi­
cides have produced more limited protests with less violence,
but with a clear focus on the problem of police homicide. New
York, Houston, Los Angeles, Dallas and other cities repeatedly
felt such protests throughout the late 1970s. 136 In the South­
west, minority groups even managed to enlist President Cart­
er's concern for the problem, 137 leading to an intensified effort
at federal prosecution of police for civil rights violations. 138 Yet
as long as the any-felony rule survives, many of the incidents
that stir public outrage will remain legal and beyond prosecu­
tion.
Although state legislatures appear less vulnerable to such
protests than police chiefs and mayors, a steadily growing num­
ber of legislatures have nonetheless reflected the apparent
change in public sentiment toward police homicide. Since 1973,
at least eight states139 have adopted the Model Penal Code limi­
tations on the use of deadly force to arrest. Minnesota has even
required that all police shootings be reported to the state gov­
ernment, in part for monitoring purposes. 140 Taken in conjunc­
tion with the developments in police policy, scholarly opinion,
and public protests, the state legislative actions are consistent
with the general trend toward restricting executions without
trial as unacceptable to society.
(4) Arbitrary infliction. Relative to the total number of
arrests and police-citizen encounters, police homicide is in­
flicted so rarely and with such arbitrariness as to be wanton
and freakish. ••• It can be likened to a virtual lottery system in
which there are no safeguards for the capricious selection of
criminals for the punishment of death. 142 Even in police depart­
ments with comparatively restrictive deadly force policies, the
discretion that even those policies allow officers in the use of
deadly force is so uncontrolled that people literally "live or die,
dependent on the whim of one man." 14 3 The available evidence
strongly suggests that police homicide is inflicted in a trivial
number of the cases in which it is legally available, through
procedures that give room for the play of racial and other prej­
udices. Unlike convictions for capital offenses, there are no
108
records kept of the number of felony suspects whose actions
make them legally vulnerable to execution without trial. The
fact that the rate of police homicide was only one per 6,822
Part I Index 144 arrests in 1975, however , provides a reasonable
inference that the sanction is rarely used even when it is availa­
ble, since the rate of flight per attempted arrest seems likely to
be much larger. Moreover, the extreme rarity of occurrence
alone raises a strong inference of arbitrariness.145
Despite the progressive policies of many police depart­
ments, many other departments still allow their officers total
discretion to use their legal power to kill.146 Even the depart­
ments with restrictive policies typically say when officers may
use their weapons, and not when they must. Noninvocation of
available legal penalties is the common practice in American
policing, as extensive research has shown, and police homicide
is no exception. 14 7 As a Kansas City, Missouri, police officer
recently said about the control of firearms discretion in that
department (one of the best managed police agencies in the
country), "they pretty much leave it up to your own conscience
to decide" whether or not to shoot someone when their restric­
tive policy allows it. 148 Many police officers are punished for
using their guns when they should not have, but recent re­
search149 has found no case in which an officer was punished for
not using force when he or she could have.
The inconsistency among police officers in deciding when to
use force is further demonstrated by a recent experimental
study of twenty-five randomly selected Connecticut police offi­
cers who were given identical information about three arrest
situations. When asked if they would be likely to use deadly
force, their responses were almost evenly split, even though
they were all making decisions under Connecticut common
law. 150
In comparison to the vigorous controls on the post-trial
death penalty described and approved in Gregg v. Georgia, 151the
use of deadly force by police is virtually uncontrolled. The trier
of fact, without any information from a record keeper about
what the typical police action has been in previous situations
similar to an instant case must also determine the sentence. If
109
decision making without access to that information is an uncon­
stitutionally arbitrary way to impose the death penalty after
the careful finding of facts at trial, then surely it must be so
without a trial.

D. Police Homicide and Equal Protection

A final argument against the use of deadly force to arrest is
that present practices deny equal protection to blacks. The ar­
gument is not without its weaker points, for discrimination in
the use of deadly force is methodologically difficult to prove .
Nonetheless, the extremely disproportionate impact of execu­
tions without trial on blacks compels consideration of the argu­
ment.
According to official statistics, blacks constituted forty-six
percent of the people killed by official police action in 1975/52
while they only constituted 11.5 percent of the population. 153
The national death rate from police homicide of black males
over age ten in a recent ten-year period was nine to ten times
higher than the rate for white males. 154 Studies in specific cities
have found even greater racial disparities in the rate of police
homicides. 155 There have been some attempts to explain the
disparity using arrest rates for FBI Part I Index crimes, 156 but
that approach has several limitations. First, the power to use
deadly force under the common-law rule is not limited to ar­
rests for .~'index" crimes. Indeed, as the empirical studies 157
show, most police shooting incidents arise out of situations in
which the initial criminal offense is clearly not an Index crime.
Second, in many police shooting situations there is no offense
recorded unless the police intervention precipitates more vio­
lence. Many violent family fights, for example, are not reported
as crimes, 158 although they are reported if a police officer is
assaulted. Third, the evidence of racial discrimination in arrests
undermines any use of arrest rates to show an absence of dis­
crimination in police homicide. 159
Even if arrest rates by race were an appropriate means of
showing that the disparity in police homicide rates is not dis­
criminatory, they do not always match the police homicide

110
 rates. In Philadelphia from 1950 to 1960, for example, where
eighty-seven percent of the police homicide victims but only
twenty-two percent of the city's population were black, only
thirty-one percent of the arrest population was black. '60 More
recently, a study of the Chicago police found the police homi­
cide rate per 10,000 arrests (for all charges) in 1969-70 to be
1.00 for whites and 2.01 for blacks.' 61 Nationally, in 1975 blacks
accounted for forty-six percent of the police homicide victims
and only thirty-three percent of the Part I FBI Index offense
arrests. 162
The existence of racial discrimination in police homicides
can be neither proved nor disproved with the available evi­
dence. Resolution of the issue would require data on the num­
ber of blacks and whites who committed acts that made them
legally vulnerable to police homicide: assaulting or threatening
to assault police or others, fleeing from arrest for felonies,
participating in a riot, or engaging in other specifically covered
behavior.163 Short of a mammoth systematic observation studt64
costing millions of dollars, there is no reliable way to obtain
such data. A sample of the narrative accounts found in arrest
reports, somewhat less expensive, would be the next best mea­
sure of legal vulnerability of whites and blacks, but no such
study has yet been done.
In the absence of more conclusive evidence, the demonstra­
bly higher rates of police homicide for blacks strongly suggests 165
racial discrimination on a national basis. Although such pat­
terns are quite likely to vary from one city to the next, such a
variation would support the argument that present procedures
allow police homicide to be administered in a discriminatory
fashion .

III. SUMMARY AND CONCLUSION

This analysis of police homicide and the Constitution leads
to the conclusion that the present state laws are unconstitu­
tional , not just in the common-law states, but in the Model
Penal Code and "forcible felony" states as well. 166 The present
laws of every state in the union deny police homicide victims

111
fifth and fourteenth amendment rights to due process, allow
the punishment of death to be imposed in a cruel and unusual
fashion, and appear to deny equal protection to blacks. The only
constitutional alternative apparent is to remove police homicide
from the realm of punishment and confine justification for it to
the self-defense doctrine, more properly called a defense-of-life
doctrine. In short, the conclusion is that the police throughout
the country should adopt the first section of the firearms policy
of the Federal Bureau of Investigation. 167
The defense-of-life policy has the virtue of being both con­
stitutional and highly practical. It is constitutional, first, be­
cause it demonstrably does not constitute punishment. Since
self-defense is an individual action rather than a state action, it
is not subject to evaluation by the Mendoza criteria. The right
to life is fundamental, and so the right to defend life need not
be granted by the State; it is, rather, something the State may
not restrict. Police and other citizens may kill under self­
defense on the same evidentiary basis-eyewitnessing an imme­
diate threat to life. If police were not granted special powers,
police killings in self defense could be distinguished from pun­
ishment administered by the state. The adoption of such an
approach would signal a return to the English tradition of
citizen-police officers, whose only special power is to arrest on
probable cause (as citizens could only do during the hue and
cry), and a rejection of the Continental tradition of soldier­
police that we have unconsciously adopted by giving the police
special powers to kill. 168 Police homicide in defense of life is
nonpunitive by its very nature. It is inherently preventive. It
uses an overt act-such as refusing to drop a gun on demand­
as the evidentiary basis for taking preventive action. By pre­
venting the consummation of a violent crime threatened by an
overt act, the defense-of-life killing looks toward the offender's
behavior in the future . Present police homicide rules all look
primarily toward the offender's past behavior, and therefore
constitute punishment.
Moreover, the defense-of-life policy is constitutional be­
cause it does not violate due process. As a solely individual

112
 action, police killings in defense of life do not deprive citizens of
rights on behalf of the state, but merely on behalf of protecting
their own rights. Finally, the defense-of-life policy does not
constitute cruel and unusual punishment. It is neither inher­
ently cruel, nor disproportionate to the conduct to which it
responds, nor unacceptable to society, nor imposed in an arbi­
trary and capricious manner. The defense-of-life policy would
still leave room, hypothetically, for racial discrimination, but it
seems most unlikely that police would grant preferential treat­
ment to whites who pose immediate threats to life and limb.
The defense-of-life policy would also be more practical to
implement than any of the other attempts to create a policy
more restrictive than the common-law doctrine. The Model Pe­
nal Code exemplifies the practical problems. As the dissent
observed in Mattis v. Schnarr, 169 a policy that allows police to
kill someone who the officer reasonably believed "would use
deadly force against the officers or others if not immediately
apprehended" requires too much guessing and analysis for an
emergency situation. This language differs sufficiently from the
"immediate danger" language of the FBI's policy to include the
killing of a fleeing felon merely because he is labeled "armed
and dangerous," (as opposed to someone who is actually com­
mitting an overt act such as pointing a gun at someone else).
The police are not armed with a crystal ball. Predicting that a
fleeing felon is likely to kill someone is no more possible than
predicting that a paroled felon is likely to kill someone. Such a
policy places an undue burden on the police officer. When peo­
ple commit overt threatening acts, however, there is much less
ambiguity.
A self-defense policy avoids the Model Penal Code's prob­
lems in allowing police officers to shoot fleeing felons only
when they have used or threatened to use deadly force. Profes­
sor Perkins argues that this provision of the Code "goes too
far" because officers making split-second decisions will find it
difficult to evaluate all the details of the suspect's conduct. 170 On
the contrary, for precisely that reason the Model Penal Code
does not go far enough.

113
 The self-defense policy also avoids the practical problems of
allowing officers to shoot fleeing suspects of specified "for­
cible" felonies, the approach used in ten states. As a former
Los Angeles Police Department policy observed, "[it] is not
practical to enumerate specific felonies." 171 An informal survey
of police officers from three New York state police depart­
ments found that none of them could remember the types of
felonies which warranted the use of deadly force under New
York state law. 172 With a self-defense policy, there is nothing
complex to remember, and no need to consider prior events; the
officer need only evaluate the information he observes to assess
whether someone is committing an overt act signaling an imme­
diate threat to the officer or someone else.
It is not the practicality of the defense-of-life rule that
makes it constitutional, however; that is merely a fortunate
byproduct. Rights cannot depend on administrative conven­
ience, especially not the right to life. The defense-of-life rule is
necessary for the simple reason that anything else constitutes
execution without trial, in violation of the Constitution.

Notes

1. The official death records of the National Center for Health Statis­
tics, preserved on tape, show a total of 295 deaths by legal interven­
tion of police for 1976. Independent tests of the death record data,
however, reveal that they are rather consistently under-reporting po­
lice homicides by about 50%. Sherman & Langworthy, Measuring
Homicide by Police Officers 70 J. Crim. L . & Criminology 546 (1979).
On the number of post-trial death sentences, see U.S. Dep't of Justice,
Law Enforcement Assistance Administration, National Criminal Jus­
tice Information and Statis tics Service, Capital Punishment, 1976; Na­
tional Prisoner Statistics Bulletin SD-NPS·CP5 at 3 (1977) [hereinaf­
ter cited as Capital Punishment Statistics].
2. Capital Punishment Statistics, supra note 1, at 13.
3. Vital Statistics of the United States, 1950-1959 (Annual).
4. See note 1 supra.
5. Comment, Deadly Force to Arrest: Triggering Constitutional Re­
view, 11 Harv. C.R.-C.L.. Rev. 361 , 368 (1976); Note, Justifiable Use

114
ofDeadly Force by the Police: A Statutory Survey, 12 Wm & Mary L.
Rev. 67 (1970). On the common law, see e.g., 2 Hale's P.C., 76-77.
6. Comment, Policeman's Use of Deadly Force in Illinois, 48 Chi.­

Kent. L. Rev. 252, 252 (1971).

7. The Code provides, in part:

The use of deadly force is not justifiable under this Section unless:

(i) the arrest is for a felony; and

(ii) the person effecting the arrest is authorized to act as a peace

officer or is assisting a person whom he believes to be authorized to

act as a peace officer; and

(iii) the actor believes that the force employed creates no substantial

risk of injury to innocent persons; and

(iv) the actor believes that:

(1) the crime for which the arrest is made involved conduct including

the use or threatened use of deadly force; or

(2) there is a substantial risk that the person to be arrested will cause

death or serious bodily harm if his apprehension is delayed.

Model Penal Code, §3.07(2)(b) (1962).

8. Sherman, Restricting the License to Kill-Recent Developments in

Police Use ofDeadly Force, 14 Crim. L. Bull. 577, 581 (1978).
9. Comment, supra note 5, at 365 n. 34.
10. The following table is constructed from four empirical studies of
police use of deadly force: (1) a study of the 32 persons killed by
Philadelphia police officers in 1950-1960. See Robin, Justifiable Homi­
cides by Police, 54 J. Crim. L.C. & P.S. 225 (1963), (2) A study of 911
police killings reported in newspapers around the country in 1965-69.
See Kobler, Figures (and Perhaps Some Facts) on Police Killings of
Civilians in the United States, 1965-1969, 31 J. Soc. Issues 185 (1975).
(3) A study of police department records, producing pooled data for
1973 and 1974 in Birmingham, Alabama; Oakland, California; Port­
land, Oregon; Kansas City, Missouri; Indianapolis, Indiana; and Wash­
ington, D.C., and in Detroit for all of 1973 and part of 1974, on 320
police firearms discharges in which a bullet wounded or killed some­
one. See C. Milton, J. Halleck, J. Lardner, G. Abrecht, Police Use of
Deadly Force (1977) [hereinafter cited as C. Milton]. (4) A study of
2,926 incidents in which New York City Police Department officers
discharged their weapons, regardless of impact, during 1971-1975. See
J. Fyfe, Shots Fired: A Typological Examination of New York City
Police Firearms Discharges (1978) (unpublished Ph.D. disser tation,
School of Criminal Justice, State University of New York at Albany).

11 5
 Events Preceding Police Use of Deadly Force
,_.
,_. Event Type
0)
STUDY FINDINGS·
Robin, 1963 Kobler, 1975b Milton ,et a/, 1977 Fyfe, 1978

(N=32) (N =911) (N=320) (N=2,926)

% Rank % Rank % Rank OJb Rank
Disturbance Calls 31 (2) 17 (4) 32 (1) 12 (2)
Family Quarrels
Disturbed Persons
Fights
Assaults
"Man with a gun"
Robbery: 28 (3) 20 (3) 21 (2) 31 (1)
In Progress
Pursuit of Suspect
Burglary: 37 (1) 27 (2) 20 (3) 8 (5)
In Progress
Larceny
Tampering with Auto
Pursuit of Suspects
Traffic Offenses: 3 (4) 3o· · (1) 8 (5) 11 (4)
Pursuits
Vehicle Stops
Officer Personal
Business: ? - ? - 4 (6.5) ?
Dispute
Horseplay
Acc ident
Stakeout/Decoy ? - ? - 4 (6.5) ?
Other 0 (5) 6 (5) 11 (4) 14 (3)
•percentages may not total 100 due to rounding
..Includes other misdemeanors not listed above
11. Of the studies cited in note 10, supra, Kobler, at 188, found 50%
of those shot by the police to have carried guns at the time and 25% to
have been completely unarmed. Milton, at 22, found 45% to have had
guns and 43% to have been unarmed. Fyfe, at IV-30, found 54% to
have had guns, and 30% to have lacked a gun or a knife . Another
study found 53% of the 1969-70 police homicide victims in Chicago to
have carried a gun, and 23o/o to have lacked any weapon. Harding &
Fahey, Killings by Chicago Police, 1969-70: An Empirical Study, 46 S.
Cal. L. Rev. 284, 292-93 (1973).
Kobler, at 165, also found that, measured by a defense-of-life stan­
dard, only 40% of the killings would have been justified; the rest were
either killings of suspects in flight or to prevent a nonviolent crime. In
contrast, Fyfe, at 279, found that 71.5 % of the police firearms inci­
dents in his New York sample were reportedly in defense of life, a
finding consistent with the tradition of relative restraint in that de·
partment. Other cities are quite different. A study of Philadelphia
police use of deadly force in 1970-74 found that approximately 45% of
those people shot had been fleeing at the time, and in approximately
25% of the incidents the shooting victim was both fleeing and un­
armed. A study by the Boston Police Department found that 102 of
the 210 targets of Boston police firearms discharges in 1970-73 were
fleeing at the time, and 80 of the 102 were unarmed. See Mattis v.
Schnarr, 547 F. 2d 1007, 1019-20 n. 30 (8th Cir. 1976).
12. See 4 W. Blackstone, Commentaries 180 (1800) (citing Von
Stiernhook, Treatise on Gothic Law).
13. W. Melville-Lee, A History of Police in England 35 (1901}.
14. "Hue and cry," under old English law, refers to the loud outcry
with which robbers, burglars, and murders were pursued. All who
heard the outcry were obliged to join in pursuit of the felon. See 4 W.
Blackstone, supra note 12, at 293.
15. L. Smith, This Realm of England 1399-1688, at 15 (1966).
16. L. Kennett and J. Anderson, The Gun in America 22 (1975).
17. Id. at 23.
18. R. Sherrill, The Saturday Night Special 4 (1973).
19 . G. Elton, Policy and Police 4, 5 (1972).
20. L. Kennett & J. Anderson , supra note 16, at 151; R. Lane, Polic­
ing The City 103-04 (1967); J. Richardson, The New York Police 113
(1970).
21. L. Kennett and J. Anderson, supra note 16, at 91.
22. This did not occur without the strenuous objections of some police

117
commanders who thought the use of revolvers was cowardly. See W.
Miller, Cops and Bobbies, 51-53 (1977).
23. R. Perkins, Criminal Law 10-11 (2d ed. 1969). As Blackstone
noted, "The idea of felony is indeed so generally connected with that
of capital punishment, that we find it hard to separate them . ..." 4
W. Blackstone, supra note 12, at 98.
24. Comment, Use of Deadly Force in the Arrest Process, 31 La. L.
Rev. 131, 132-33 (1970); see 4 W. Blackstone, supra note 23.
25. Furman v. Georgia, 408 U.S. 238, 333-41 (1972) (Marshall, J.,
concurring).
26. P. Pringle, Hue and Cry 133 (1955).
27. The Professional Thief, by a Professional Thief (E.H. Sutherland,
ed.) 112 (1937).
28. Quoted in L. Kennett and J. Anderson, supra note 15, at 150.
29. Quoted in Miller, supra note 21, at 146.
30. Storey v. State, 71 Ala. 329, 340 (1882) (involving the theft of a
horse).
31. E.g., United States v. Clark, 31 F. 710, 713 (8th Cir. 1887);
Reneau v. State, 70 Tenn. 720 (1879).
32. Pearson, The Right to Kill in Making Arrests, 28 Mich. L. Rev.
957, 976 (1930).
33. Regina v. McKay [1957] V.R. 560, 572-73 (Smith, J., dissenting);
11 Halsbury's Laws of England §1179 (4th ed. 1976) (emphasis added).
The question of deadly force to prevent flight is either implied in this
formulation, or so far beyond the pale that the current formulations
make no mention of it. See also Lanham, Killing the Fleeing Offender,
1 Crim. L.J. (Australia) 16, 17-18 (1977).
34. Quoted in Regina v. McKay, [1957] V.R. 560, 572-73 (Smith, J.,
dissenting).
35. Id. at 572.
36. Criminal Law Revision Committee, Seventh Report: Felonies and
Misdemeanours 7 (1965); 18 Parliamentary Papers (House of Com­
mons and Command) (1964-65).
37. Wiley v. Memphis Police Dep't, 548 F.2d 1247 (6th Cir.), cert.
denied 434 U.S. 822 (1977).
38. Id. at 1252.
39. Id. at 1251.
40. Id. at 1252.

118
41. Mattis v. Schnarr, 547 F.2d 1007 (8th Cir. 1976).
42. Wiley v. Memphis Police Dep't, 548 F.2d 1247, 1252-53 (6th Cir.),
cert. denied 434 U.S. 822 (1977).
43. Id. at 1254.
44. Wiley v. Memphis Police Dep't, 434 U.S. 822 (1977).
45. G. Newman, The Punishment Response 7 (1978).
46. Professor Hart, for example, suggests five defining characteristics
of punishment:
(1) It must involve pain or other consequences normally considered
unpleasant
(2) It must be for an offense against legal rules
(3) It must be imposed on an actual or supposed offender for his
offense
(4) It must be intentionally administered by human beings other than
the offender
(5) It must be imposed and administered by an authority constituted by
a legal system against which the offense is committed.
H. Hart, Punishment and Responsibility 4, 5 (1968).
Professor Packer, in contrast, finds that definition insufficiently
clear as to the distinction between the purposes and effects of punish­
ment, and proposes a sixth defining characteristic of punishment: "It
must be imposed for the dominant purpose of preventing offense
against legal rules or of exacting retribution from offenders, or both."
H. Packer, The Limits of the Criminal Sanction 21-23, 31 (1969).
47. 99 S. Ct. 1861 (1979).
48. Id. at 1873.
49. Id. at 1873-74.
50. 372 u.s. 144 (1963).
51. 99 S. Ct. at 1899.
52. Id.
53. Id. at 1887 (Marshall, J., dissenting).
54. Id. at 1873.
55. Id. at 1898 (Stevens, J., dissenting); H. Packer, supra note 45, at
33.
56. 372 U.S. at 168-69.
57. Id. at 169.

119
58. Furman v. Georgia, 408 U.S., 238, 290 (1972) (Brennan, J., con­
curring).
59. !d. at 286.
60. Gardner v. Florida, 430 U.S. 349, 357-58 (1977) (Stevens, J., con­
curring).
61. Roe v. Wade, 410 U.S. 113, 157 (1973); Screws v. United States,
325 U.S. 91, 123 (1945); Johnson v. Zerbst, 304 U.S. 458, 463 (1938);
Yick Wo v. Hopkins, 118 U.S. 356, 370 (1886); Mattis v. Schnarr, 547
F.2d 1007, 1018 (1976).
62. Quoted in T. Critchley, A History of Police in England and Wales
(2d ed. 1972).
63. J. Bellamy, Crime and Public Order in England in the Later Mid­
dle Ages 134 (1973).
64. Cited in Regina v. McKay, [1957] V.R. 560, 571-72 (Smith, J.,
dissenting).
65. R. Hunisett, The Medieval Coroner 67 (1961).
66. !d. at 49.
67. R. Perkins, supra note 23, at 985.
68. Note, Legalized Murder of a Fleeing Felon, 15 Va. L. Rev. 582,
583 (1929).
69. Helwig v. United States, 188 U.S. 605 (1902).
70. !d. at 612.
71. Child Labor Tax Case, 259 U.S. 20 (1921).
72. United States v. Balint, 258 U.S. 250, 251 (1922).
73. See e.g., R. Perkins, supra note 23, at 771.
74. 547 F.2d 1007, 1023 (8th Cir. 1976).
75. J. Gibbs, Crime, Punishment and Deterrence (1975).
76. Professor Waite argued for extending the right to kill to arrest for
all offenses in order to deter flight, for otherwise "we say to the
criminal, 'You are foolish ... if you submit to arrest. The officer dare
not take the risk of shooting at you. If you can outrun him, outrun him
and you are safe .. . .If you are faster than he is you are free and God
bless you.' I feel entirely unwilling to give that benediction to the
modern criminal." 9 ALI Proceedings 195 (1931), quoted in J. Michael
& H. Wechsler, Criminal Law and its Administration 81-82 n. 3 (1940).
77. Judge Learned Hand once commented that "It has been con­
stantly supposed here that if you are able to shoot a robber you are

120
less likely to have a robber. I question that. I challenge it altogether. I
don't believe that possibility figures at all in the commission of crime."
35 ALI Proceedings 258-334 (1958), quoted in Mattis v. Schnarr, 547
F.2d 1007, 1015 (8th Cir. 1976). While Judge Hand's remarks were
directed specifically towards private citizens' rights to defend prop­
erty, the Mattis court observed that he was spea king to the larger
problem of justification to use deadly force in general. Id. at 1015 n.
17.
78. 528 F.2d 132, 142 (2d Cir. 1975).
79. ALI Proceedings, 186-87, quoted in J. Michael and H. Wechsler,
supra note 76.
80. 99 S. Ct. at 1873.
81. Holloway v. Moser, 193 N.C. 185, 136 S.E. 375, (1927), quoted in
Pearson, supra note 32, at 964.
82. 99 S. Ct. at 1899 (Stevens, J., dissenting).
83. Storey v. State, 71 Ala. 329, 341 (1882).
84. Regina v. McKay, [1957] V.R. 560.
85. Ingraham v. Wright, 430 U.S. 651, 671-72, n.40 (1977).
86. United States v. Lovett, 328 U.S. 303, 317 (1946).
87. 99 S. Ct. at 1872.
88. Mattis v. Schnarr, 547 F.2d 1007, 1018-19 (8th Cir. 1976).
89. Brown v. United States, 256 U.S . 335(1921).
90. 40 Am. Jur. 2d Homicide §§170-71, quoted in Mattis v. Schnarr,
547 F.2d 1007, 1015 (1976).
91. Mattis v. Schnarr, 547 F.2d 1007, 1019 (1976).
92. Note, Specifying the Procedures Required by Due Process: To­
wards Limits on the Use ofInterest Balancing, 88 Harv. L. Rev . 1510,
1524 (1975).
93. Dworkin, Taking Rights Seriously, in Oxford Essays in Jurispru­
dence 202, 214 (2d Series 1973), quoted in Note, supra note 92, at
1527 n.76.
94. Morrissey v. Brewer, 408 U.S. 471, 481 (1972); Cafeteria & Res­
taurant Workers Union v. McElroy, 367 U.S. 886, 895 (1961).
95. Note, supra note 92, at 1519.
96. See note 61 supra and accompanying text.
97. See Palko v. Connecticut, 302 U.S. 319, 327 (1937); Comment,
supra note 5, at 378.

121
98. Note, supra note 92, at 1528-29.
99. Wiley v. Memphis Police Dep't, 548 F.2d 1247, 1251-52 (8th Cir.
1976). See also Terry v. Ohio, 392 U.S. 1, 23 (1968).
100. For an example of such a sentence, see the case of nursing home
operator Bernard Bergman , reported in N.Y. Times, June 18, 1976,
§A, at 1, col. 7.
101. Furman v. Georgia, 408 U.S. 238, 258 (1972) (Brennan J., con­
curring).
102. 408 u.s. 238 (1972).
103. Robinson v. California, 370 U.S. 660 (1962); Louisiana v. Reswe­
ber, 329 U.S. 459 {1947); In re Kemmler , 136 U.S. 436 (1890); Wilker­
son v. Utah, 99 U.S. 130 (1878).
104. Robinson v. California, 370 U.S. 660, 676 (1962) (Douglas J.,
concurring); O'Neil v. Vermont, 144 U.S. 323, 339 (1892) (Field, J.,
dissenting).
105. Trop v. Dulles, 356 U.S. 86 (1958).
106. 408 U.S. at 256 (Douglas, J., dissenting).
107. 428 u.s. 153 (1976).
108. Id. at 169; see Coker v. Georgia, 433 U.S. 584 (1977); Roberts v.
Louisiana, 428 U.S. 325 (1976); Woodson v. North Carolina, 428 U.S.
280 (1976); Jurek v. Texas, 428 U.S. 262 (1976); Proffitt v. Florida,
428 U.S. 242 (1976).
109. See Louisiana v. Resweber , 329 U.S. 459 (1947); Wilkerson v.
Utah, 99 U.S. 130 (1878).
110. See Screws v. United States, 325 U.S. 91 (1945).
111. See Sherman, The Breakdown of the Police Code of Silence, 14
Crim. L. Bull. 149, 150-51 (1978) (discussing the Joe Campos Torres
beating and drowning case). At least four southern California men
died from police choke-holds in one recent year. See Cory, Deadly
Force, Police Magazine, Nov. 1978, at 5, 6.
112. One study found that police homicide cases are typically not re­
ferred to a grand jury, and that only three cases in some 1,500 led to
police officers being criminally punished. Kobler, Police Homicide in a
Democracy, 31 J. Soc. Issues 163 (1975). A study of police use of
deadly force in 49 Los Angeles county police agencies found that of 18
incidents officially designated as having been in violation of the de­
partment's firearms policies, only one was referred for criminal prose­
cution; only two led to dismissals, two led to suspensions, and 13 (72o/o)
led to either a reprimand or no punishment at all. Delman, Varieties of

122
Police Policy: A Study of Police Policy Regarding the Use of Deadly
Force in Los Angeles County, 6 L.A.L. Rev. 1, 40 (1973). A study of
police records in six cities found that of the eight percent of shooting
incidents judged improper by administrative reviews punishment
"generally consisted of a reprimand rather than suspension or termi­
nation." Milton, supra note 10, at 28.
113. 408 U.S. at 394 (Burger, C.J., dissenting).
114. Id. at 393 (Burger, C.J., dissenting).
115. See note 10 supra.
116. Milton, supra note 10, at 46.

117_ The justification, however, is up to the jury to determine in light

of all the circumstances of a particular case. See 11 Halsbury's Laws,

supra note 33, §1180.
118. See 408 U.S. at 279-80 (Brennan, J., concurring).
119. "[The Court], before it reduces a sentence as 'cruel and unusual,'
must have reasonably good assurances that the sentence offends the
'common conscience,' " which not even opinion polls can measure.
United States v. Rosenberg, 195 F.2d 583, 608 (2d Cir. 1952), quoted
in Furman v. Georgia, 408 U.S. 238, 360 (1972) (Marshall, J., concur­
ring).
120. Trop v. Dulles, 356 U.S. 86, 100-01 (1958).
121. 408 U.S. at 278 (Brennan, J., concurring).
122. Milton, supra note 10, at 45-49.
123. Gun Rules Tightened, L.A. Times, Sept. 9, 1977, at 1.
124. Cory, Police on Trial in Houston, Police Magazine, July 1978, at
33, 40.
125. Findings ofPolice Deadly Force Study Spark Three-Way Contro­
versy in Birmingham, Law Enf. News, June 21, 1977, at 1, col. 1;
Personal Communication with B. R. Myers, Police Chief, Birmingham,
Alabama. (November 1978).
126. See also Sarat and Vidmar, Public Opinion, The Death Penalty,
and the Eighth Amendment: Testing the Marshall Hypothesis, 1976
Wis. L. Rev. 171, 179.
127. Planning and Research Division, Boston Police Department, the
use of Deadly Force by Boston Police Personnel, (May 3, 1974), cited
in Mattis v. Schnarr, 547 F.2d 1007, 1016 n.19. See also Glance, Police
Tell Firearm Policies, San Diego Union, Oct. 16, 1975, (nine of ten
cities in San Diego County employ a defense-of-life police firearms

123
policy). Contra, Leeds & Lowe, Survey Finds Fe:w Rules on Police Use
of Guns, Chicago Tribune, Dec. 6, 1977.
128. Baker, Model Firearms Policy for California Law Enforcement,
10 J . Cal. L. Enforcement 5 (1975).
129. FBI, Memorandum 31-72 (Nov. 21, 1972), quoted in Mattis v.
Schnarr, 547 F .2d 1007, 1015 (8th Cir. 1976). Policies more restrictive
than state law are also reported in Comment, The Use ofDeadly Force
in Arizona by Police Officers, 1973 L. and Soc. Order 481.
130. In the 40-year history of federal narcotics enforcement, 17
agents have been killed by assault in the line of duty, almost as many
as in the FBI which has had at least four times as many agents and a
longer history. J. Wilson, The Investigators 48 (1978).
131. Mattis v. Schnarr, 547 F.2d 1007, 1015 (8th Cir. 1976). Even
these policies, however, may be ambiguous. The FBI policy reportedly
goes on to allow the use of any force necessary to effect an arrest.
Personal communication with Dr. Charles Wellford, Office of the
United States Attorney General (Dec. 7, 1979).
132. See, e.g., Pearson, supra note 32; Safer, Deadly Weapons in the
Hands of Police Officers, On Duty and Off Duty, 49 J. Urb. L. 565
(1972); Note, supra note 68; Note, supra note 5; Comment, supra note
24; Comment, supra note 6; Comment, supra note 23. But see Miller,
The Law Enforcement Officer's Use ofDeadly Force: Two Approaches,
8 Am. Crim. L.Q. 27 (1969).
133. See, e.g., Police Policy on the Use ofFirearms, The Police Chief,
July 1967, at 16.
134. President's Commission on Law Enforcement and Administra­
tion of Justice, Task Force Report: The Police 189-90 (1967).
135. For example, such riots occurred in San Francisco, St. Louis and
Los Angeles in 1966. See id. at 189.
136. Killings of Chicanos by Police Protested, N.Y. Times, Oct. 12,
1977, §A, at 17 col. 1; Houston Quiet After Violence Hospitalizes Over
12, N.Y. Times, May 9, 1978, at 22 col. 1; 2,000 Assail Police at Black
Rally As Off-Duty Officers Meet Nearby, N.Y. Times, July 17, 1978,
§B, at 3, col. 1; Los Angeles Police Scored on Shooting, N.Y. Times,
Aug. 15, 1977, at 13 col. 1.
137. Gilman, In Washington, A Ne:w Zeal for Prosecuting Police, Po­
lice Magazine, Nov. 1978 , at 15, 18.
138. Id. Measured by the number of cases in which the victim died,
however, Justice Department prosecutions of police officers have actu­
ally declined under the Carter administration. From 1970 through

124
1976, the average number of federal civil rights prosecutions for police
homicide was four per year; in 1977, and 1978 it was only two per
year. Personal communication from Daniel F. Rinzel, Civil Rights Divi·
sion, U.S. Department of Justice (November 30, 1978).
139. See Minn. Stat. §609.066 (1976); Comment, supra note 5, at 368·
69.
140. Minn. Stat. §626-553 (1976).
141. Furman v. Georgia, 408 U.S. 238, 310 (Stewart, J., concurring).
142. Id. at 293 (Brennan, J., concurring).
143. Id. at 253 (Douglas, J., concurring).
144. Computed from National Center for Health Statistics, Public
Health Service, Department of Health, Education and Welfare, Vital
Statistics of the United States 1975 II Mortality Part A 1·168; FBI,
Crime in the United States 1975-1979. Using the unofficial estimated
number of police homicides, the rate was one per 3,411 Part I Index
arrests.
145 . Goldberg and Dershowitz, Declaring the Death Penalty Unconsti­
tutional, 83 Harv. L. Rev. 1773, 1790 (1970), quoted in Furman v.
Georgia, 408 U.S. 238, 249 (1972) (Douglas, J., concurring).
146. Until 1968, one large southwestern department employed the
following policy on the use of a firearm, quoted in its entirety: "Never
take me out in anger; never put me back in disgrace." Milton, supra
note 10, at 47. Other "policies" have included "Leave the gun in the
holster until you intend to use it," and "It is left to the discretion of
each individual officer when and how to shoot." Id. at 47-48.
147. K. Davis, Police Discretion (1975); National Institute of Law En­
forcement and Criminal Justice, Police Discretion: A Selected Bibliog·
raphy (1978); Black, The Social Organization of Arrest, 23 Stan. L.
Rev. 1087 (1971); Goldstein, Police Discretion Not to Invoke the Crimi­
nal Process: Law Visibility Decisions in the Administration ofJustice,
69 Yale L.J. 543 (1960).
148. Personal interview (January, 1979).
149. The Project on Homicide by Police Officers, Crim inal Justi ce
Research Center, State University of New York at Albany, has stud·
ied this area.
150. G. Hayden, Police Discretion in the Use of Deadly Force: An
Empirical Study of Information Usage in Deadly Force Decision Mak·
ing (1979) (unpublished paper, University of New Haven).
151. 428 u.s. 153 (1976).

125
 152. National Center for Health Statistics, supra note 144. The total
figure for all minority group members is probably somewhat higher,
but no official statistics for other nonwhites are reported.
153. Bureau of the Census, Department of Commerce, Statistical Ab­
stract of the United States 25 (1976).
154. P. Takagi, A Garrison State in a "Democratic" Society, in Police
Community Relations 358 (A Cohn & E. Viano eds.).
155. See note 10 supra.
156. E.g., Milton, supra note 10, at 19; Burnham, 3 of 5 Slain by
Police Here are Black, Same as the Arrest Rate, N.Y. Times, Aug. 26,
1973, at 50, col. 3. See also The Management of Police Killings, Crime
& Soc. Just., Fall-Winter 1977, at 34; Goldkamp, Minorities as Vic­
tims of Police Shootings: Interpretations of Racial Disproportionality
and Police Use ofDeadly Force, 1 Just. Sys. J. 169 (1977).
157. See note 5 supra and accompanying text.
158. Parnas, The Police Response to the Domestic Disturbance, 1967
Wis. L. Rev. 914.
159. See Black, supra note 147. The fact that the greater likelihood of
police to arrest black suspects can be largely attributed to (a) the
greater tendency of blacks to be antagonistic to the police and (b) the
greater tendency of black complainants-who do almost all of the
accusing of black suspects during street encounters with the police-to
demand an arrest does not remove discrimination in a legal sense.
Neither suspect's attitudes nor a complainant's preference constitute
proper grounds for enforcement decisions. ld. at 1097-1107.
160. Robin, supra note 10.
161. R. Knoohuizen, R. Fahey, & D. Palmer, The Police and Their Use
of Fatal Force in Chicago 21 (1972) (unpublished study).
162. National Center for Health Statistics, supra note 144; FBI, su­
pra note 144.
163. Comment, supra note 5.
164. See, e.g., A. J. Reiss, the Police and the Public (1971); Reiss ,
Systematic Observation of Natural Social Phenomena, in Sociological
Methodology 3-33 (H. Costner, ed. 1971). Since police only draw their
weapons once in every hundred citizen encounters (and patrol cars in
many large cities ave rage no more than ten encounters in eight hours),
it could typically require two weeks of observation in order to capture
one drawing of a weapon. See Cruse & Rubin, Determinants of Police
Behavior, in Project Report to National Institute of Law Enforcement
194 (1972).

126
165. Other equal protection arguments can be made in addition to
those concerning race. See Comment, supra note 5, at 375-80.
166. For a survey of the differing state approaches, see materials
cited in note 5 supra.
167. See notes 129 & 131 supra and accompanying text.
168. See generally B. Chapman, Police State (1970); R. Fosdick, Euro­
pean Police Systems 17-20 (1915); Bayley, The Police and Political
Development in Europe in The Formation of National States in West­
ern Europe 328-79 (C. Tilly, ed. 1975).
169. See Mattis v. Schnarr, 547 F .2d 1007, 1023 (8th Cir. 1976) (Gib­
son, C.J., dissenting).
170. Perkins, supra note 23, at 986.
171. Milton, supra note 10, at 48.
172. This survey was conducted by the Project on Homicide by Police
Officers, Criminal Justice Research Center, State University of New
York at Albany.

127
 CHAPTER 5

MINORITIES AS VICTIMS OF POLICE

SHOOTINGS: INTERPRETATIONS OF RACIAL

DISPROPORTIONALITY AND POLICE USE OF

DEADLY FORCE*

JOHNS. GOLDKAMP
The power to take life exists not only at the final stage of
the criminal process where the state may execute prisoners
under sentence of death but also at the earliest stage where
deadly force may be used by police in the apprehension of sus­
pected lawbreakers. The controversy surrounding the use of
capital punishment continues to this day, but what distin­
guishes a death sentence from the taking of life by police deadly
force is the availability of due process safeguards. Although
the use of deadly force by police is often defined by statute and
restricted by departmental policy, 1 it remains a decision guided
mainly by the judgment of individual officers in pressure situa­
tions.
For more than a decade studies have shown that racial
minorities-principally black Americans-number dispropor­
tionately among persons killed by police through the use of
deadly force. 2 In 1972 the Supreme Court found that the death

*I would like to express my appreciation to Professor Hans Toch and

Michael Hindelang for permission to cite unpublished manuscripts in
this article.
SOURCE "Minorities as Victims of Police Shootings: Interpretations of Racial
Disproportionality and Police Use of Deadly Force," Justice System Journal
(Winter, 1976) 2:169-183.

128
 penalty discriminated against minorities and considered this a
justification for its suspension.a Yet findings of racial dispropor­
tionality in killings by police have not warranted similar atten­
tion.
A useful first step in consideration of racial disproportional­
ity and police killing is examination of how these death rates
are presently viewed . How the "problem" is defined and under­
stood has ramifications for the formulation of future social pol­
icy; especially as it pertains to law enforcement. The present
discussion is primarily concerned with interpretation of the
phenomenon of racial disproportionality in civilian death rates
resulting from the use of deadly force by police. Consideration
of the most common interpretations is important because they
are inexorably linked to beliefs about race and crime which
cannot help but influence social policy in criminal justice
decision-making.
Some writers suggest that the disproportionately high
death rates of minorities at the hands of the police can be
explained by the disproportionately high arrest rates of minori­
ties for crimes of violence,4 or by assumptions concerning the
suspect's responsibility for his/her own death in violent police­
suspect interactions. 5 Others see disproportionate minority
deaths as resulting from both irresponsible use of deadly force
by a small minority of police officers and differential adminis­
tration of Jaw enforcement toward minority citizenry (which in
effect produces disproportionately high arrest and death rates
for minorities in general). Kobler6 and Knoohuizen, Fahey, and
Palmer' stress the possibility that police misconduct may play a
considerable role in generating these civilian deaths. Takagi
ascribes disproportionality to the simple fact that "police have
one trigger finger for whites and another for blacks."8
In the analysis that follows, explanations of minority death
rates are aligned with either of two schools of thought. The
first school encompasses a broad range of writings that address
themselves to the policing of minority groups - from strong
rhetoric that speaks in terms of the "racist machinery of polic­
ing,"9 to hypotheses of criminologists who ponder the possible

129
effects of prejudice or discrimination.10 The second school of
thought typically relates disproportionately high minority death
rates to high arrest rates for crimes of violence and seeks to
explain the violent propensities of racial minorities.
In subsequent sections, these two positions or belief per­
spectives will be elaborated by presenting the ideas of a number
of writers whose views these positions embody. 11 After discern­
ing how disproportionality in minority death rates is viewed,
the perspectives will be discussed in terms of their
plausibility-by consulting some recent data and/or posing al­
ternative theorizationY Finally, the question of whether the
two positions go far enough in explaining the high death rates
of racial minorities in situations where deadly force is used
must be addressed.

POSITION 1: DISPROPORTIONALITY-A
QUASI-LABELING VIEW
As is typical of most of the writers who deal with racial
disproportionality, Forslund acknowledges the higher arrest
rates of blacks for crimes of violence. In fact, he cites arrest
rates for rape and homicide (from UCR, 1967), which are 12
and 17 times higher for blacks than for whites. However, it is
his interpretation of these rates which aligns him with the first
school of thought:
When comparing the crime rates of whites and Negroes
in the U.S. , the answer to the question is complicated
by the possibility that pre_iudice and discrimination
may at least subtly, if not blatantly, affect the Negro's
crime rates ...
. . . For example, as a consequence of decades of dis­
crimination the Negro, to a greater extent than the
white, is concentrated in the lower socioeconomic strata
of American society. If the agents of criminal justice
act more quickly toward those at the bottom of the
socioeconomic heirarchy than toward those at the top,
and there is some evidence to suggest that this is the
case, then the Negro's crime rate would be affected by
his disproportionate concentration among the bottom
strata of society.13

130
Forslund suggests that black arrest rates for violent crimes
may be substantially inflated due to some mechanism which
encourages law enforcement to apply itself more rigorously
against individuals associated with the "bottom strata of soci­
ety." Clark stresses a similar theme when he remarks that "the
lowest status of minorities (including the Irish and Italian
Americans of the earlier twentieth century) always have higher
street crime rates." 14 Disproportionately high arrest rates, he
implies, are in part the result of the mobilization of law enforce­
ment in such a manner that more arrests are produced from
settings whose residents are characterized as among the lowest
socioeconomic strata.
In a standard criminology text, Sutherland and Cressey
take much the same position when they state that "the proce­
dures used in the administration of criminal justice are biased
against minority groups, especially blacks" and argue that any
crime rate index (such as arrests) will exaggerate the amount of
black crime. 15 Geis claims that "arrest statistics do not tell us
very much about the criminal activity among minority groups,"
and further characterizes them as "misleading" and "subject to
misinterpretation.' ' 16
Writers grouped under the first perspective, it becomes
clear, characterize arrest statistics for minorities as distorted
and of minimal value since they are generated by differential
deployment of law enforcement against lower status racial mi­
norities. Such statistics are not seen as true indicators of mi­
nority crime. On these grounds, Position 1 incorporates the
view that the disproportionate death rates expressed by racial
minorities are also artifacts of differential policing, and not
because of the greater criminality of minorities.
A number of perspectives exist which attempt to explain
how differential policing develops. One writer, Swett, relates
differential policing of minorities to the ethnocentrism of the
police officers by arguing that an individual who enters police
work has a certain status-quo linked (middle-class) view of
society-which he affirms by choosing a law enforcement ca­
reer. When the citizenry he serves deviates culturally from his

131
middle-class, ethnocentric orientation, the officer is more likely
to become suspicious of behavior and is more likely to intercede
in the affairs of citizens. In this way, more arrests are likely to
be generated by police officers who operate in lower socioeco­
nomic status areas, because it is there the greatest cultural
differences will be perceived. Arrest statistics, which are gener­
ated as a result, create stereotypes that reinforce the officer's
view of racial minorities as more criminal. 17 Since police/citizen
interaction will be more frequent and more hostile, racial mi­
norities will find themselves involved more frequently in violent
interactions with police where deadly force will be more readily
deployed.
When taken together as one school of thought, these views
represent an approach to disproportionate arrest and death
rates which amounts to an invocation of labeling theory; that is,
racial minorities are labeled by majority society as highly crime
producing. These minorites are segregated or contained by one
means or another, singled out as specially deserving of extra­
substantial attention from law enforcement agencies, more fre­
quently considered dangerous, and consequently more fre­
quently subdued by means of deadly police force. From this
point of view, arrest rates and death rates in police shootings
can be seen as outcomes of this labeling process. A logical ex­
tension of this view of disproportionality pictures the police as
labelers actively and consciously participating in the oppression
of minorities. 18

POSITION II: DISPROPORTIONATE RATES DUE TO

DISPROPORTIONATE PARTICIPATION BY RACIAL
MINORITIES IN CRIMES OF VIOLENCE
The second school of thought represented here can be char­
acterized in a manner quite different from the quasi-labeling
perspective of Position I. To begin with, this second position
holds that a disproportionate number of minority "suspects"
are killed by police because a disproportionate number of mi­
nority group members are arrested for violent crimes. Position
II advocates do not dismiss the disproportionate arrest rates of

132
minorities for crimes of violence, but instead consider them to
be good reflections of minority participation in violent crimes;
in fact, arrest rates are pointed to as indications that minority
members are actually involved in the crimes to a disproportion­
ate degree . From the point of view of this position, prejudice
and discrimination are not considered significant factors in the
generation of disproportionate arrest figures. IH
In addition, not only does Position II assume that racial
minorites are disproportionately participating in crimes as of­
fenders, it further assumes that by learning about actors in
violent encounters more will be learned about the possible
sources of racial disproportionality in violent crime rates. In
contrast to the labeling stance of Position I, Position II imputes
an essentially active role to racial minorities in bringing about
high crime rates resulting in high death rates where confronta­
tions with police occur.
Wolfgang and Ferracuti provide the principal thrust for
this position. While they do not address the question of police
killings of minorities, they do interest themselves in the dispro­
portionate arrest rates of minorities for crimes of violence by
assuming that these rates reflect an actual disproportionate
participation by minorities in violent crimes. They thus attempt
to explain this disproportionate participation in terms of subcul­
tural differences which characterize minorities. For example,
they explicitly delcare the following:
Statistics on homicide and other assaultive crimes in
the U.S. consistently show that Negroes have rates be­
tween four to ten times higher than whites. Aside from
a critique of official arrest statistics that raises serious
questions about the amount of Negro crime, there is no
real evidence to deny the greater mvolvement that Ne­
groes have in assaultive cnmes ....
There is reason to agree, however, that whatever may
be the learned responses and social conditions contrib­
uting to criminality, persons visibly identified and so­
cially labeled as Negroes in the U.S. appear to possess
them in considerably higher proportions than do per­
sons labeled white. Our subculture of violence thesis
would therefore expect to find a large spread to the

133
 learning of, reso:t t<?, and criminal display of violence
values among mtnonty groups such as Negroes. 20
It was in large part while attempting to explain the dispro­
portionate arrest rates of blacks and other minorities for crimes
of violence that Wolfgang and Ferracuti were led to conceptual­
ize the existence of "subcultures of violence." From the per­
spective of this theory, subcultures are found to exist-mainly
in the ghettoes, mainly among blacks-that are characterized
by the presence of norms which differ considerably from the
"larger society" to the extent that assaultiveness is not seri­
ously discouraged, but is even permitted as normal. In these
subcultures, violence may not be uncommon in certain situa­
tions, may be expected, and in fact may even be called for. 2 ' By
means of this view, consequently, racial disproportionality in
arrest rates for violent crimes is made considerably easier to
comprehend: one finds in the higher violent crime rates for
blacks the expression of "subcultural themes." Violence for
blacks is more "normal" than it is for whites.
In sum, Wolfgang and Ferracuti's contribution to Position
II's perspective is two-fold: blacks are relatively different from
and more violent than whites. Whites, as members of "the
larger society" are not "different," and are less violent. Conse­
quently, one would expect that the "agents of the larger soci­
ety" (the police) will be confronting violent blacks-and other
minorities (as subcultural theory is extended to include His­
panic populations) much more frequently than they will be con­
fronting violent members of the white population. In this sense,
the subculture of violence theory is essentially racial, at least as
applied to the United States. Through it, disproportionate ar­
rest rates are interpreted, and disproportionate probabilities
that blacks will be involved in fatality-producing police-minority
encounters can be predicted.
Another major contribution to the Position II perspective­
which explains disproportionate rates by disproportionate par­
ticipation in violent criminal acts-is found in some of the work
of Toch. As a student of violent interpersonal interactions,
"violence-prone" individuals, and police violence in particular.

134
Toch's approach resembles that of Wolfgang and Ferracuti in
the sense that he accepts the disproportionality in crime rates
as actual and seeks to explain the actors who tend to become
disproportionately involved in violence. Toch exhibits the think­
ing characteristic of the Position II perspective, for example, in
a passage which discusses police killings of minorities:
... [A] disproportionate number of civilians involved in
police violence are non-white .... Nationally, 60 per­
cent of persons shot by officers are non-white. In New
York City, where blacks make up about 19 percent of
the populations, they account for 59 percent of fatal
police victims. In Detroit, only one of the various per­
sons killed by STRESS was white . .. .
. . . [Dlistributions of known offenses and of arrests
show ethnic disproportions similar to those of the vio­
lence data. In New York, for example, during the per­
iod corresponding to the 59 percent police victims fig­
ure, 62 percent of the persons arrested for violent
crimes were black. 22
It should be emphasized that Toch approaches the racial
disproportionality issue only incidentally, through a larger con­
cern with violent interactions, "violence prone" individuals,
and police violence. Race-or "ethnicity" as it is referred to by
Toch-is more of a complicating factor, to be considered only
after understanding some of the other ingredients which seem
to characterize interactions which become violent.
In Toch's scheme, violence-prone individuals contribute dis­
proportionately to the total of all violent interpersonal inci­
dents. This is partly because, in Toch's words:
The violence-prone person invites violence-prone inter­
actions with other people. These interactions follow a
pattern, in that they arise under repeatedly occurring
circumstances, and in that they serve equivalent ends. 23
In another passage from Violent Men, he suggests that:
. . . two types of orientations are especially likely to
produce violence: one of these is that of the person who
sees other people as tools designed to serve his needs;
the second IS that of the individual who feels vulnerable

135
 to manipulation. These two perspectives, when we ex­
amine them more closely, become faces of the same
coin; both rest on the premise that human relationships
are power-centered, one-way affairs; both involve ef­
forts at self-assertion with a desperate feverish quality
that suggests self-doubt. 24

Toch's understanding of violence-prone individuals and the

themes that influence their involvement in violent incidents can
be seen to merge rather easily with the subculture of violence
concept propounded by Wolfgang and Ferracuti. It appears log­
ical to Toch that violence-prone themes might be prevalent in
certain segments of society, while not in others . For instance,
Toch points to the "well-known machismo syndrome" which he
relates to the mother-dominated family configuration of Mexi­
can slums-and which he associates with the result that males
in these segments are prone to asserting their masculinity. 25
Moreover, it is clear that, like Wolfgang and Ferracuti, Toch
sees a definite role for the subculture of violence:
Whatever the origins of the subculture of violence, it
exists in the form of values, beliefs, and attitudes held
by its members. These may relate to all manner of situ­
ations, and may prescribe appropriate conduct for
them. Violence-proneness is restricted to the select mi­
nority within the subculture who have assimilated its
violence-prone teachings and who live by them. 26

Toch's use of the subculture of violence , however, is quali­

fied; and with it, he moves beyond Wolfgang and Ferracuti's
preoccupation with violent "suspects" or criminals. His study
of violence-prone individuals has included violence-prone police
officers as well as violence-prone offenders. He has extended
the subcultural concept to include police subcultural themes.
Thus, violence-prone individuals - who are characterized as
sharing essentially similar problems of self-assertion in the face
of self-doubt differ mostly in the way that their personalities
intersect with their respective group norms. In other words,
blacks in encounters with white police may be playing out cer­
tain themes in relation to black subcultural norms, while white

136
police officers may be dealing with similar feelings in the con­
text of norms prevalent in the police subculture. 27 By consider­
ing the intersection of personality needs, insecurity and self­
doubt with subcultural themes, Toch sees an opportunity for
understanding the inclinations of both actors (potentially, at
least) in police-minority confrontations. 28
Thus, against the background of these two general
themes-actors attempting to assert themselves in the face of
self-doubt and the availability of certain violence-prone group
norms-Toch explains the disproportionate death rates of mi­
norities at the hands of the police. As noted earlier, in line with
Position II thinking, his explanation centers on the dispropor­
tionate arrest rates of minorities for crimes of violence. While
Toch stresses the similarity of minority and police subcultures,
his emphasis on their similarity noticeably withdraws when dis­
proportionate arrest and death rates are concerned. It is clear
that he would not extend the concept of police violence­
proneness to explain either of the high minority rates in ques­
tion. The question that is not answered, then, is: What mecha­
nism is it that attracts police and minorities to interact with
each other to the extent that the death rates in question are the
end product?
Toch clearly contends that minority death data are not
"transferable into inferences about discrimination." 29 Conse­
quently, it can be assumed that he would not consider racism­
as implicitly charged by Position I -a possible explanation.
Race (or "ethnicity" as it is alluded to by Toch) only enters the
violence equation incidentally: "ethnicity is a contributing vari­
able to violence , because a disproportionate number of civilians
involved in police violence are non-white." 30 That is, race enters
the police violence picture, quite simply, because more minority
group members are actually engaging in more violent crime.'11
From Toch's analysis, it is possible to understand why
police-minority confrontations might be especially violent-prone
and embittered. However, in a contradictory fashion, Toch sub­
sequently lays his elaborate analysis aside when approaching
minority fatalities and takes the position that more minorities

137
are killed by police because more minorities are involved in
violence-and that neither prejudice nor discrimination figures
in. This is the essence of the Position II perspective.
Finally, as if to underline this subcultural difference, or at
least to dismiss the possibility of reading racism or discrimina­
tion into the explanation of racial disproportionality, Toch
seeks to bolster his argument by noting that black officers are
as a group disproportionately responsible for fatalities resulting
from the police use of deadly force-fatalities where minorities
are the most frequent victims. 32 The New York Times study (to
which he makes reference) showed that while only one of 250
white officers killed suspects and one of 58 Hispanic officers
killed suspects, as many as one of 38 black officers killed sus­
pects.33 The validity of Toch's argument, perhaps a logical out­
growth of the subculture of violence theory, will be examined in
a subsequent section.
Overall, the Position II perspective differs from the Posi­
tion I perspective by the manner in which it assigns a very
active role to one or both of the participants in police-minority
violent encounters. Wolfgang and Ferracuti have theorized that
subcultural themes may be influential in producing dispropor­
tionate arrest rates for blacks and other minorities for crimes
of violence. It follows from their reasoning that blacks commit
more crimes of violence, and will therefore be more frequently
involved in violent confrontations with police. Toch has ex­
tended consideration of violence-proneness to encompass per­
sonal themes which intersect with subcultural norms-for both
minority groups and police alike. Yet his use of arrest statistics
to explain death rates for victims of police violence tends to
associate violence-proneness disproportionately with minority
actors.

ARE THESE POSITIONS PLAUSIBLE?

Both positions-the quasi-labeling perspective of Position I
and the violent racial minority stance of Position II-make as­
sertions that deserve empirical testing. Comprehensive data on
incidents where deadly force has been deployed by police are

138
not currently available nor easily obtained. However, for the
purposes of this discussion, three of the more important points
of contention will be examined in light of recent victim survey
data or in light of alternative support or theorization.

Point A: The Meaning ofArrest Rates

Perhaps the most striking point of contention between the
two positions centers around the question: Are disproportion­
ate arrest rates manufactured by some mechanism of control
which is differentially directed at minorities, or are they rea­
sonable reflections of disproportionate participation by minori­
ties in crimes of violence?
Recent data from a victimization survey made available by
the National Crime Panel and reported by Hindelang provide
an opportunity to examine this question from a source other
than official statistics. 34 It is especially noteworthy that consid­
erable discrepancies were found to exist between the amount of
crime reported to interviewers by victims and the amount of
crime officially reported to the police and reflected in official
statistics. For various reasons, considerably less crime was re­
ported to authorities than was experienced by victims among
the persons interviewed. 35
In addition to the hope that the victim survey could provide
a more accurate picture of the amount and kind of crime occur­
ring in American society, another hope was that a great deal
more might be learned about the characteristics of criminal
victimization. One particular variable, which bears directly on
the present point, recorded the race of the offenders involved in
personal victimizations as perceived by the victims.36
Perhaps the simplest victimization category relevant to the
present discussion of disproportionality and crimes of violence
is "assaultive violence without theft. " This victimization cate­
gory includes rape, aggravated and simple assault where theft
was not also involved.) When data in this category are exam­
ined, it is learned that 45 percent of the offenders (an aggre­
gate all-Impact City figure) were perceived to be black/others. 37
Only about 35 percent of the aggregate population for the cities

139
surveyed was composed of black/others . When each of the cities
were examined separately, the finding of moderately dispropor­
tionate participation by minorities in violent offenses appeared
to be upheld.
Yet, in addition, it should also be noted that blacks were
found to be much more disproportionately responsible for theft­
related personal victimizations. In these victimizations, nearly
three quarters of all offenders were perceived to be black/
others-in sharper contrast to their actual representation in
the total population, or 35 percent. 38
These general victimization findings would appear to
weaken considerably the claims advanced by Position I which
interpret disproportionality as entirely "manufactured" by ar­
rest policy, rather than as reflecting actual minority participa­
tion in crime. Concomitantly, some support would seem to be
garnered for the Position II stance which relates the dispropor­
tionately high arrest rates of blacks for crimes of violence to
actual disproportionate involvement by blacks in violent acts.
Black/others are seen to be somewhat disproportionately in­
volved in assaultive victimizations by the victims themselves;
however, black/others are considerably more disproportionately
involved in personal theft victimizations.

Point B: The Race and Socioeconomic Status of

the Offender

A second point involves the role of race and the role of

socioeconomic status in relation to high arrest rates for violent
crimes, a relationship which underpins the arguments of both
positions when they address the issue of high minority death
rates at the hands of the police.
Position I would argue that both high arrest rates and high
death rates are correlates of low socioeconomic status-to the
extent that law enforcement is differentially directed toward
the lowest income groups in contemporary American society.
Since blacks are considerably overrepresented among these
strata of society, blacks are disproportionately affected by law

140
enforcement policy. The Position I argument becomes less sim­
ple to characterize on this issue because of the diverse views
which were grouped together in forming it. (For example, cer­
tain adherents of this position would insist that the relationship
between low socioeconomic status, race and differential polic­
ing cannot be adequately explored without considering how ra­
cial minorities came to be overrepresented among the lowest
income groups.) Nonetheless, Position I proponents would es­
sentially agree that the high arrest rates and high death rates
of minority suspects testify to the degree to which differential
policing is operative.
Position II would not dispute the general relationship be­
tween high crime rates, low socioeconomic status, and race. On
the contrary, Position II theorists would readily concede that
subcultures of violence characteristically find roots in low in­
come areas. However, Position II theorists might wish to con­
tend that, while low socioeconomic status might be an anteced­
ent of subcultural norms, the facts are that subcultures exist
among racial minorities and racial minorities participate dispro­
portionately in violent crimes.
Resolution of this point of contention through victimization
data is not so clear. One finding which addresses the question
of differential policing and race, and differential policing and
socioeconomic status, is the following: The survey found that on
the whole the most highly victimized segments of society (for
any type of personal victimization) were families whose incomes
were less than $7,500 annually. 39 Those persons whose incomes
fell below $3,000 annually were especially frequently victimized
both by violence- and theft-related victimization - regardless of
race.
From these findings, Position II theorists would find fuel to
argue that what is called "differential policing" (by Position I)
may well be an honest attempt to provide adequate police pro­
tection to the lowest income groups -since these strata experi­
ence the most victimization. Unfortunately, what these data
cannot demonstrate is the extent to which lower income blacks
and lower income whites (as the groups which are most highly
victimized) receive varying amounts of police attention.

141
 Since data pertaining to the socioeconomic status of per­
ceived offenders are unfortunately not available, no new deter­
minations concerning the role of race and socioeconomic status
of offenders can be made here. It is not possible to ascertain
whether or not poor whites might not also be overrepresented
among perceived offenders. Thus, in this way the question
raised, but not resolved, is: Would it be just as fruitful to dis­
cuss economic disproportionality as racial disproportionality?
If, based on the findings presented under Point A above
("The Meaning of Arrest Rates"), Position I theorists were to
concede that a great deal of minority participation in offense
activity was real, they might then insist on characterizing mi­
nority involvement in violence as stemming from economic sub­
cultural themes, rather than the essentially racial subcultural
themes which are ascribed to minority participation in crime by
Position II. This economic approach would be intended to sub­
sume minority death rates in police shootings, as well as minor­
ity arrest rates.
Implied in this line of investigation would be an examina­
tion of victims of police violence for socioeconomic status, in
addition to race. At the same time, it would be interesting to
learn what proportion of non-minority police victims are very
low socioeconomic status whites: Do poor whites also become
disproportionately involved in crimes of violence when com­
pared to their share of the total white population?

Point C: Prejudice and Discrimination

Is it a logical necessity that discrimination or prejudice be
ruled out as considerations in police violence because black offi­
cers (like black offenders) appear to participate disproportion­
ately in violent incidents and are in fact disproportionately re­
sponsible for the deaths of alleged minority suspects?
Alternative interpretations have been suggested to explain
statistics which purport to demonstrate disproportionate use of
deadly force by black officers. For example, a former New York
police commissioner, Donald Cawley, offered the following in­
terpretation in an interview:

142
 . . . the reason a larg-er number of black policemen
were involved in fatal mcidents was that a larger pro­
portion of black policemen were assigned to dangerous
JObs such as servmg as detective and undercover agents
and working in precincts with the highest crime rates.~"
Thus it is questionable whether simple statistics showing
higher black officer involvement in suspect fatalities can be so
easily read in terms of racial propensities toward violence.
Other variables, such as differential assignment to hazardous
duty, may well contribute to the high rates of participation that
characterize black officers. It is especially questionable when
these statistics are used to dismiss the argument that the dis­
proportionate number of minority deaths might be linked to
policies of differential enforcement, discrimination, the dispens­
ing of "summary justice," or other charges involving racism.
Differential assignment of blacks to dangerous police duty in no
way precludes the criticism that policing may be unfairly de­
ployed against racial minorities, thus generating higher arrest
and death rates. On the contrary, some would see differential
assignment of black officers to dangerous duty as one compo­
nent of an overall racist policing policy.
Even without differential assignment, however, there are
other interpretations which might be made to explain why black
officers might become more frequently involved in violent inci­
dents with minority suspects-interpretations that do not pre­
clude the existence of a discriminatory or racist policing mecha­
nism. These particular interpretations would describe the
dilemma of the black officer who is caught between the minor­
ity status which he is seeking to escape and the police culture
which he is seeking to join.
One such interpretation is offered by Swett:
Observational experience indicates that as a result of
selection procedures, upward social mobility, the encul­
turation aspects of police training, and internalization
of the police cultures value system, such officers iden­
tify more with the police culture than with their ethnic
mmority culture. At the same time the position of the
ethnic minority officer within the police culture is in
some respects marginal.
143
 ... In consequence, the ethnic minority officer displays
a greater degree of suspicion of the culturally different,
particularly members of his own ethnic group, than the
non-minority officer. Whereas suspicion of the non­
minority officer is reciprocated by a culturally different
minority so that he IS regarded as the enemy, the
greater suspicion of the ethnic minority officer is recip­
rocated to a greater degree by his own ethnic group, so
that he is not only considered an enemy but also a trai­
tor. This often impairs his effectiveness in dealing with
members of his own group :11
What Swett describes is not very different from what
Bettelheim described in his discussion of life in a concentration
camp, where the individual identities of particular members of
the oppressed group become so entangled with the identity of
the oppressor/aggressor to that they begin to emulate the
oppressor/aggressor to the extent of participating in their own
oppression and the oppression of their groupY Thus, seen in
this light, the fact that black officers participate to a dispropor­
tionate extent in the killing of minority suspects does not neces­
sarily mean that one may dismiss the Position I charges which
at least imply a role for racism in accounting for the dispropor­
tionate number of minority "suspect" deaths. In this regard,
whether black police officers kill more or less frequently is be­
side the point.

SUMMARY AND CONCLUSIONS: DO THESE POSITIONS

OFFER ENOUGH OF AN EXPLANATION?
This discussion began by attempting to examine two per­
spectives which are frequently invoked when racial dispropor­
tionality is being interpreted-either in the realm of police kill­
ings or in the realm of arrest rates for crimes of violence. The
first perspective, characterized as a quasi-labeling stance,
linked the disproportionately high death rates of minority sus­
pects to the impact of some differential policing mechanism
which seemed to be operating in American society. The second
perspective linked disproportionate minority death rates to dis­
proportionate minority arrest rates for crimes of violence. Ar­

144
 rest rates were assumed to reflect disproportionate actual par­
ticipation by minorities in violence, and were thereby used to
explain why minorities might be found to a disproportionate
extent among fatalities resulting from violent encounters with
the police. Higher actual participation in violent incidents by
minorities was related to some differences which were appar­
ently linked with minority status (principally race or ethnicity).
Have these positions explained enough about racial dispro­
portionality? Perhaps not. In effect, more questions seem to be
raised than can be easily answered by these positions.
A look at recent victimization (and census)•a data revealed
that disproportionate minority arrests did appear to reflect a
disproportionate actual participation by minorities in crimes of
violence to a moderate degree-based on the victim's percep­
tion of the race of the offender. Position l's quasi-labeling, dif­
ferential enforcement theory is not greatly supported.
An additional finding showed that the lowest income
groups (regardless of race) were the most highly victimized
segments of society. Although this finding might serve as a
rationale for providing greater police protection for lower in­
come persons in general-that is, a form of differential
policing-it is not ascertainable from these data whether such
differential policing actually exists, or whether the degree of
such policing differs among white and black low income groups.
Nor is it possible to surmise whether poor whites might also be
generating a disproportionate share of offenders. If this last
possibility were discovered to be true to some extent, then the
race-linked theorizations of Position II would prove inadequate.
Further, even if disproportionate participation by minori­
ties in violent crimes is granted, is it entirely sufficient to as­
sume, as Position II implicitly does, that "pure risk" can ex­
plain the greater death rate of minorities? That is, merely
because minorities are seen to be involved more frequently in
violent crimes than whites, is it logical to assume that minori­
ties will appear among victims of police shootings much more
frequently than whites? Without a comparative analysis of vio­
lent interactions between police and "suspects" (white and

145
black) to determine when, how frequently, and for whom deadly
force is invoked as a response by police, it is difficult to feel
confident merely assuming that higher offense and arrest rates
automatically extrapolate into higher death rates.
In addition, even though minorities do appear in fact to
participate somewhat disproportionately in crimes of violence,
they participate much more disproportionately in theft-related
personal victimizations (economic rather than violent crime).
This is an attribute of disproportionality which has not been
addressed by either of the positions outlined above.
The utility of this article lies in the presentation and clarifi­
cation of two views of racially disproportionate death rates re­
sulting from the use of deadly force by police. Contemporary
discussion of these death rates is rare. When discussion occurs,
as in the two "belief perspectives" above, it is inadequate.
Elaboration of these perspectives may be useful since they re­
flect current thinking about race and crime in general, and
represent views most likely to influence formulation of social
policy in this area of criminal justice.
Analysis of racial disproportionality in police-shooting fatal­
ities in even the simplest terms has been lacking-with the
result that criminal justice agencies may feel free to adopt ei­
ther belief perspective without basing their decisions on empiri­
cal evidence. Consequently, depending on which view of racial
disproportionality is favored, different approaches might be
freely adopted in such areas as the deployment of police man­
power in minority neighborhoods, the role of minority officers,
or in establishing a policy toward incidents where the use of
deadly force by officers had resulted in fatalities. Certainly,
racial disproportionality in deaths resulting from police use of
deadly force deserves the kind of careful consideration given
race in the Supreme Court's deliberation on capital punish­
ment.

Notes
1. For a comprehensive discussion of the restrictions on the use of
deadly force at arrest and the constitutional question s at issue , see

146
 Comment, "Deadly Force to Arrest: Triggering Constitutional Re·
view," 11 Harvard Civil Rights-Civil Liberties Law Review 361-389
(1976).
2. Findings of racial disproportionality in civilian deaths by deadly
force are by now commonplace. On August 26, 1973, for example, the
New York Times reported that about 60 percent of all such deaths
recorded between 1970 and 1972 in New York City were black citi­
zens. Kenneth Clark reported (New York Times, September 18, 1974)
that three-quarters of all victims of police shootings between 1970 and
1973 were members of minority groups; about half were black and one
quarter were of Hispanic ethnicity, in sharp contrast to their actual
share in the local population.
Takagi reviewed civilian deaths resulting from police use of force in
California and reported that the incidence rate of such deaths in­
creased by two and one-half between 1962 and 1969 and that these
rates have remained constantly nine times higher for blacks than for
whites over the last 18 years (Paul Takagi, "A Garrison State in a
'Democratic' Society," Crime and Social Justice, p. 29 [Summer,
1974).) Kobler reports national statistics reflect disproportionality for
the years 1952 to 1969. Although blacks generally comprise about 50
percent of all victims in Kobler's data, it should be noted that blacks in
1970 comprised only from 12 to 14 percent of the nation's population.
(Arthur L. Kobler, "Police Homicide in a Democracy." 31 Journal of
Social Issues, 163 [1975] at p. 165.) A detailed analysis of deadly force
is found in Brief for Appellant at Appendix A. Wiley v. Memphis
Police Department (1975), noted in Comment, "Deadly Force to Ar·
rest: Triggering Constitutional Review," note 1 supra.
A decade earlier Robin investigated civilian deaths resulting from po­
lice force in Philadelphia which occurred between 1960 and 1970 and
found that nearly 90 percent of those killed were black- whereas the
black community during this time accounted for only about 22 percent
of the total local population. (Gerald Robin, "Justifiable Homicide by
Police Officers," 54 Journal of Criminal Law, Criminology, and Po­
lice Science 225 [1963) at pp. 226-27.) At the same time, Robin re­
viewed statistics concerning police killings of civilians in eight other
American cities (Akron, Chicago, Kansas City, Miami, Buffalo, Phila­
delphia, Boston, and Milwaukee) and found black victims included
among the dead more frequently than whites at ratios which ranged
from 6 to 1 (in Akron) to 30 to 1 (in Milwaukee).
3. Furman v. Georgia, 408 U.S. 238 (1972).
4. See Hans Toch's unpublished manuscript on police and prison vio­
lence (1976), and the New York Times, August 26,1973. Often the
news media serve as the principal sources of statistics and descriptions

147
of incidents involving police killings of civilians . In general, whether
from academic journals or from press releases, the information in this
area is rather incomplete and of questionable reliability. The intent of
this paper is not to review the methodology or to compare the findings
of different studies-there is wide agreement that minorities serve
disproportionately as victims of the use of deadly force by police. What
is of interest here is the logic that pervades the various explanations of
the disporportionate death rates of minorities whether that logic is
exhibited in academic journals or in the news media.
5. Robin, note 2 supra.
6. Kobler, note 2 supra.
7. Ralph Knoohuizen, Richard P. Fahey and Deborah J . Palmer, "The
Police and Their Use of Fatal Force in Chicago," Chicago Law En­
forcement Group, 1972 .
8. Takagi, note 1 supra, p. 30.
9. Angela Y. Davis, If They Come in the Morning (New York: New
American Library, 1971).
10. Morris A. Forslund, "A Comparison of Negro and White Crime
Rates," in Charles E. Reasons and Jack L. Kuykendall (eds.), Race,
Crime, and Society, (Pacific Palisades, Calif.: Goodyear, 1972), pp. 96­
102.
11. The treatment of these writers and their interpretations of racial
disproportionality is meant to be illustrative, not exhaustive.
12. The data used in the present analysis were not designed to ad­
dress the issue of racial disproportionality and police shootings. How·
ever, they do allow for a preliminary assessment of the soundness of
some of the points raised by the two positions.
13. Forslund, note 10 supra, p. 100. Emphasis added.
14. Kenneth Clark, as quoted in the New York Times, September 18,
1974.
15. Edwin H. Sutherland and Donald R. Cressey, Criminology (New
York: Lippincott, 1970), p. 133. In explaining racial disproportionality
in American prison populations, Charles Reasons and Russell Kaplan
(in "Tear Down the Walls? Some Functions of Prisons," Crime and
Delinquency, October, 1975, at pp. 369-70) state: "Cultural bias oper­
ating in law enforcement and the courts produces a prison population
that is largely non -white and poor."
16. Gilbert Geis, "Statistics Concerning Race and Crime," Crime and
Delinquency, (April 1965), pp. 144-149. Geis advocates discontinuing

148
the use of race categories in crime statistics, based on a fear that
criminal justice agencies would manipulate what he considered very
questionable figures for their own benefit and to the detriment of
minority groups.
17. Daniel Swett, "Cultural Bias in the American Legal System," in
Reasons and Kuykendall, Race, Crime and Society, note 10 supra, pp.
38-40.
18. Reasons and Kuykendall, for example, suggest that "police have
become committed to repressing change and have developed a political
philosophy to support their position" (in Race, Crime, and Society,
note 10 supra, pp. 142-43). This, in effect, is not a far cry from a more
radical characterization which would trace racial disproportionality to
a social policy of racial and economic genocide which is transmitted by
the present configuration of power and implemented through its polic­
ing apparatus. (See Bettina Ap theker, "The Social Functions of Prison
in the U.S.," in Davis, note 9 supra, pp. 51-56.)
19. Two studies report findings which lend support to this Position II
feature; both find little support for the prejudice argument. In sepa­
rate observational studies on police interactions with citizens and on
arrests of juveniles, the following conclusions were drawn: (a) "There
is a general paucity of evidence of discriminatory or prejudicial behav­
ior on the part of police officers in face-to-face encounters with Ne­
groes" (Donald J. Black and Albert J. Reiss, "Patterns of Behavior in
Police and Citizen Transactions," in Reasons and Kuykendall, note 10
supra, p. 203); and (b) where serious offenses were concerned (includ­
ing such violent crimes as robbery, homicide, rape and aggravated
assault), police discretion in effecting arrest was not influenced by
personal demeanor or social class characteristics. (Irvin Piliavin, and
Scott Briar, "Police Encounters with Juveniles," in Donald Cressey
and David Ward (eds.), Delinquency, Crime, and Social Process [New
York: Harper and Row, 1969], p. 159.)
20. Marvin E. Wolfgang and France Ferracuti, The Subculture of Vio­
lence (New York: Tavistock Publication, 1967), p. 265. Emphasis
added.
21. Ibid., pp. 158-65.
22. Toch, note 4 supra, ch. 2, p. 6.
23. Hans Toch, VioLent Men (Chicago: Aldine, 1969), p. 255.
24. Ibid., p. 183.
25. Ibid. , p. 190.
26. Ibid., p. 191.

149
27. Toch points to similarities between blacks and the police as two
minorities faced with uncertainty:
Militant spokesmen for the ghetto are prone to characterize the
police as an invading army representing the white establishment.
This characterization is ironic because the police, far from being
agents of the majority, are a minority themselves. The objective
situations of blacks and police are in many ways similar. The police
inhabit a ghetto of their own, and they are doomed to segregation.
They have little hope of man-to-man communication with civilians,
who-even if favorably disposed to law enforcement-tend to be
nervous and self-conscious in encounters with officers.
Quoted from Hans Toch, "Cops and Blacks: Warring Minorities," in
Gerald Leinwand (ed.), The Police (New York: Pocket Books, 1972),
pp. 196-97.
28. This does not mean to imply that both actors will always be
equally responsible for the denouement of violent interactions. While
violent interactions are interpersonal situations involving at least two
parties, the presence of one violence-prone individual may be enough
to explain the excalation into violent outcomes.
29. Toch , note 4 supra, ch. 2, p. 6.
30. Ibid.
31. It is true that Toch's analysis of police-minority violence extends
further than is described here; for example, he sees a particular role
for "fear" as it interacts with the ethnicity of the participants in police
violence to produce "polarization on a very large scale," Ibid. , p. 2.
32. Ibid. , p. 6.
33. New York Times, August 26, 1973.
34. Michael J. Hindelang, Criminal Victimization in Eight American
Cities (LEAA Draft Document, 1974). In this survey, representative
probability samples of approximately 10 ,000 households were selected
for study in eight Impact Cities (Atlanta, Baltimore, Cleveland, Dallas,
Denver, Newark, Portland, and St. Louis). Respondents 14 years of
age and older were asked to report on those criminal victimizations
affecting the entire household within the preceding 12-month period.
35. Concerning reasons for non-reporting, Hindelang reported the fol·
lowing:
For each category of victimization, the belief on the part of the
individual that nothing could be done about the victimization was
the reason most frequently given, followed by the belief that the
victimization was not sufficiently important to report.
See Hindelang, note 34 supra, p. 386.

150
36. The question of the extent to which racial perceptions might be
inaccurate has been raised, but even when allowances are made for
possible misperceptions, general findings concerning disproportional­
ity remain substantially unaffected.
37. Hindelang, note 34 supra, Table 3:21.
38. Ibid.
39. Ibid., Table 3:7.
40. New York Times, August 26, 1974.
41. Swett, note 17 supra, pp. 39-40.
42. Bruno Bettelheim, "Individual and Mass Behavior in Extreme Sit­
uations," in Harold Proshanksy and Bernard Seidenberg (eds.), Basic
Studies in Social Psychology (New York: Holt, Rinehart and Winston,
1965).
43. The following are examples of socioeconomic indicators which
characterized black populations according to the 1970 Census: Black
families earned median annual incomes which average about $3,500
less than all families in the total population; 25 percent of all black
families earned incomes below $3,000 annually-compared to only 12
percent of all American families (including blacks); blacks owned con­
siderably Jess of their own housing than did all families; blacks lived in
considerably more crowded housing conditions; fewer blacks on the
average completed high school than the general populations as a
whole; blacks were considerably more unemployed; more blacks
worked in menial occupations. (From U.S. Bureau of the Census, Sta­
tistical Abstract of the U.S., 1972; County and City Data Book, 1972;
1970 Census Tracts; General Social and Economic Characteristics,
1972.)

151
 CHAPTER 6

POLICE SHOOTINGS AT MINORITIES: THE

CASE OF LOS ANGELES

MARSHALL W. MEYER

The purpose of this article is to present basic data on the

involvement of minorities in Los Angeles Police Department
shooting incidents from 1974 to 1979. Few reliable statistics
concerning shootings at minorities now exist, although several
studies of the subject are now underway. As a consequence,
there has been little reasoned discussion of the numerous issues
arising in connection with minority involvement in police shoot­
ings.

The data in this article were originally released by the Los Angeles
Board of Police Commissioners in Part IV of their "Report of the
Board of Police Commissioners Concerning the Shooting of Eulia Love
and the Use of Deadly Force" for which the author served as the
commission's consultant. Any opinions or conclusions not stated in the
Police Commission's Report are solely those of the author and do not
necessarily reflect the views of the Los Angeles Board of Police Com­
missioners or of the Los Angeles Police Department. The cooperation
of Chief Daryl F. Gates and of numerous staff officers of the Los
Angeles Police Department throughout the course of this study is
gratefully acknowledged as are the comments and suggestions made
by the commission and the department.

SOURCE : "Police Shootings at Minorities: The Case of Los Angeles," The

Annals of the American Academy of Political and Social Science (November,
1980), 452:98·110. Reprinted with permission of t he American Academy of
Political and Social Science and the author.

152
 Another purpose of this article is to illustrate the limita­
tions of statistics on police shootings. The data presented in the
following pages show the involvement of blacks-but not of
Hispanics-in Los Angeles Police Department shooting inci­
dents to have been different in most respects from that of
whites. Many questions are raised by this pattern, but no defini­
tive explanation is provided by the data. Why this is so will be
discussed in the concluding section.
The data used in this article describe shooting incidents in
which Los Angeles police officers discharged firearms from 1
January 1974 to 31 December 1979. The analysis relies entirely
upon the Los Angeles Police Department's accounts of shoot­
ings presented in original investigative reports of shooting inci­
dents and other departmental documents. These investigations
of shootings are quite lengthy and detailed, perhaps more so
than in any other U.S. police agency. No attempt has been
made to reconcile these departmental records of shootings with
other accounts, such as those in the press or in court records,
and no independent investigation of shooting incidents was
made at the time of the incidents or in connection with the
preparation of this article.
Most of the text and all the tables in this article describe
Los Angeles Police Department officer-involved shooting inci­
dents in the 1974-78 interval. The text also makes reference to
shooting incidents occurring in 1979. The data are drawn from
files maintained by the Staff Research Section of the Personnel
and Training Bureau of the Los Angeles Police Department,
supplemented by information obtained from departmental per­
sonnel files and records of the Robbery-Homicide Division of
Detective Headquarters Bureau. Of the 913 incidents of shoot­
ing occurring from 1974 through 1978, all but one, the Sym­
bionese Liberation Army (SLA) shoot-out of May, 1974, are
included in our data files. 1
Two data files were constructed based on information made
available by the department. Records in the first, the "suspect"
file, describe the person-or object, if any-shot at. The date
and location of each shooting, a description of the person or

153
object shot at, the suspect's action prior to the shooting, weap­
ons, if any, possessed and/or used by the suspect, shots fired by
the Los Angeles Police officers, and the results of the shooting
review process are indicated for each person or object shot at.
Shootings of bystanders, hostages, animals, and accidental dis­
charges and other non-accidental shootings are included in the
suspect file, but are excluded from the present analysis. 2 One
entry is made in the suspect file for each person or object shot
at in an incident. There are 984 entries in the suspect file due to
the involvement of more than one civilian in some shooting
incidents, 605 of whom are suspects, that is, persons believed
by officers to have committed or to be engaged in criminal acts.
The second data file is our "officer" file. Records in this file
contain information on each Los Angeles Police officer involved
in a shooting in the 1974-78 interval. Up to six shootings are
coded for each officer.3 The location of the shooting incident,
the officer's assignment, shots fired, and the outcome of review
of each shooting are described in the officer file. Some 1,070
officers discharged their weapons in the shooting incidents re­
viewed for this study, excluding the SLA shoot-out.
One hundred forty-six officer-involved shooting incidents
occurring in 1979 have also been reviewed in connection with
this analysis, but have not been entered into our data files.
Certain data concerning 1979 shooting incidents, 101 of which
involved suspects, are presented in narrative discussion. The
1979 shooting incidents are not included in statistical tables
since investigations, reviews, and final adjudications of a num­
ber of the 1979 Los Angeles Police officer-involved shootings
were not completed in time to be included in this study.
Most of the items or variables used in this study are taken
directly from departmental accounts of shooting incidents.
Shooting investigations are quite detailed and include a chrono­
logical narrative of events preceding the shooting, a listing of
weapons used and shots discharged, listings of evidence and
witnesses, and descriptions of the suspect-including his or her
race or descent-gunshot wounds sustained, and the scene of
the shooting incident. Reports of shooting review boards are

154
 attached to departmental investigations and state the classifica­
tion of the shooting-in policy, in policy but below departmen­
tal standards, out of policy, or accidental-and the recom­
mended corrective or disciplinary action, if any. Shooting
review reports also include explanatory comments.
All but one of the items used in this study are taken directly
and unambiguously from Los Angeles Police Department shoot­
ing investigations and shooting review board reports. The de­
partment does not routinely classify the actions of suspects
shot at in tactical situations, but classification of suspects'
actions precipitating shootings was deemed necessary for pur­
poses of this study and was done for all shooting incidents
involving suspects occurring from 1974-78, except for the SLA
shoot-out. Seven categories were used to classify suspects'
actions prior to shooting incidents, based on the chronological
narrative of events preceding shootings. Using a weapon,
whether a gun, knife, automobile used for purposes of assault,
or any other potentially lethal or injurious object, is one such
category. Threatening the use of but not actually using a
weapon, whether by pointing or aiming it or by indicating ver­
bally that a weapon would be used, is a second category. Dis­
playing a weapon while not threatening its use, either verbally
or otherwise, is a third category. Assaulting an officer or civil­
ian where no weapon is used, threatened, or displayed is a
fourth category. Appearing to reach for a weapon when no
weapon is actually used, threatened, or displayed, and there is
no assault, is a fifth category! Disobeying an officer's order,
usually an order to freeze or halt, when no weapon is used,
threatened, or displayed, and there is no assault, in the sixth
category. A seventh category is other actions precipitating
shootings and includes accidental discharges at suspects.
In almost all instances, the suspect's act precipitating a
shooting incident is the final act that caused the officer to fire,
that is, that act but for which the shooting would not have
taken place. The exceptions are those occasional instances in
which two or more potentially precipitating acts occurred
within a very short period of time, for example, firing a weapon

155
and then disobeying a command to freeze, in which case only
the higher classification or most life-endangering act of the
suspect is the one coded. The categories of disobeying officers'
commands and of appearing to reach for weapons are thus
extremely restrictive and include only cases in which no more
threatening action of the suspect occurred within the period
immediately preceding the shooting.
The categories used to describe suspects' weapons are
straightforward, but the reader should note that the unarmed
category is quite restrictive. A suspect is considered to have
been unarmed only if he did not use a weapon, including a
vehicle for purposes of assault, and if he is found after the
shooting incident not to have possessed a weapon . In other
words, a suspect who did not use, threaten, or display a weapon
but is ultimately found to have been in possession of one is
classified as being armed.
One further introductory comment is required. Police fire­
arms discharges in Los Angeles decreased over the 1974-79
interval, especially those kinds of incidents specifically re­
stricted by a new shooting policy that took effect in late 1977.
Numbers of shooting incidents, suspects shot at, suspects shot
(hit), suspects shot fatally, and shots fired per incident fell sub­
stantially during this period. Shootings precipitated by suspects
disobeying officers' orders to halt or making furtive gestures as
if reaching for weapons and shootings at unarmed suspects
declined even more dramatically. 5 The declines of shootings of
all types, and especially of incidents requiring the most careful
review and evaluation, should be kept in mind in interpreting
data on minority involvement in Los Angeles Police Depart­
ment shootings.
LOS ANGELES POLICE SHOOTINGS INVOLVING
MINORITIES

Numbers ofShootings
A large number of blacks compared with Hispanics and
whites have been involved in police shootings in Los Angeles

156
(Table 1). Of the 584 suspects shot at from 1974-78 whose race
or descent is known, 321 (55 percent) were black, 126 (22 per­
cent) were Hispanic, 131 (22 percent) were white, and 6 (1
percent) were of other nonwhite origins. The race or descent of
21 suspects shot at from 1974-78 is unknown. In 1979, however,
of 101 suspects shot at whose race or descent is known, 46 (45
percent) were black, 32 (32 percent) were Hispanic, and 23 (23
percent) were white. The race or descent of one suspect shot at
in 1979 is not known.
The proportion of black suspects involved in Los Angeles
Police Department shooting incidents appears to have changed
little over the decade prior to 1979. During a three and a half
year period from 1968 to 1971, 57 percent of suspects shot at
by Los Angeles officers were black. 6 This proportion differs
insignificantly from the proportion of suspects shot at who
were black-55 percent-from 1974-78.
From 1974-78, blacks accounted for 36 percent of all ar­
rests and 46 percent of Part I-of FBI Index crime-arrests7 in
Los Angeles. From 1974 to 1978, blacks were reported to have
committed 44 percent of all attacks and 42 percent of assaults
with deadly weapons upon Los Angeles Police officers. Fifty­
five percent of the suspects shot at, 53 percent of those actually
hit, and 50 percent of suspects shot fatally by Los Angeles
Police officers in this period were black. In 1979, blacks ac­
counted for 36 percent of all arrests and 44 percent of Part I
arrests and were charged with 38 percent of all attacks and 41
percent of assaults with deadly weapons upon Los Angeles Po­
lice officers. Forty-five percent of the suspects shot at, 50 per­
cent of those actually hit, and 62 percent-8 of 13 suspects­
shot fatally by Los Angeles Police officers in 1979 were black.
From 1974 through 1978, Hispanics accounted for 27 per­
cent of all arrests and 24 percent of Part I-of FBI Index
crime - arrests in Los Angeles. From 1974 to 1978, Hispanics
were reported to have committed 24 percent of attacks and 25
percent of assaults with deadly weapons upon Los Angeles Po­
lice officers. Twenty-two percent of the suspects shot at, 22
percent of those actually hit, and 16 percent of suspects shot

157
1-'
01
00

TABLE 1 POPULATION, ARRESTS , ATIACKS ON OFFICERS, ADWs * UPON OFFICERS AND SUSPECTS SHOT AT,
HIT, AND SHOT FATALLY BY RACE OR DESCENT (IN PERCENT)

1977 1974-78 1974-78 1974-78 1974-78 1974-78 1974-78 1974-78

Popula- Total Part I Attacks On ADW's Upon Suspects Suspects Suspects
tiont Arrests Arrests Officers Officers Shot at Hit Shot Fatally

Black 18 36 46 44 42 55 53 50
Hispanic 24 27 24 24 25 22 22 16
White 52 35 28 28 26 22 23 33
Other
nonwhite 6 2 2 4 7 1 2
-­ - - -­ -­ -­ -­
100 100 100 100 100 100 100 100
Number
(not in
percent) 1,267,299 219,224 5976 2360 584 307 128

• ADW = Assaults with deadly weapon.

t Population on percentages are based on results of a 1977 sample survey conducted by the Los Angeles Community
Development Department and reported in "Population, Employment, and Housing Survey, 1977," volume Ill.
fatally by Los Angeles Police officers in the period were His­
panic. In 1979, Hispanics accounted for 31 percent of all arrests
and 30 percent of Part I arrests and were charged with 32
percent of all attacks and 34 percent of assaults with deadly
weapons upon Los Angeles Police officers. Thirty-one percent
of the suspects shot at, 33 percent of those actually hit, and 15
percent of those-2 of 13 suspects-shot fatally by Los Angeles
Police officers in 1979 were Hispanic.
Departmental records do not indicate the race or descent of
assailants involved in shootings of officers from 1974-78. How­
ever, a total of 19 officers who discharged their weapons were
shot-that is, hit-by suspects' bullets from 1974-78. Thirty­
seven percent-seven-of the suspects involved in these shoot­
ings were black, 37 percent-seven-were Hispanic, and 26
percent-five-were white. From 1974 through 1978, five Los
Angeles Police officers were shot fatally. Four blacks and one
Hispanic were apprehended in connection with these shootings;
the descent of the person responsible for one of the officer
fatalities is unknown.
A higher percentage of shootings by police officers than of
reported violent crimes has taken place in preponderantly black
communities in Los Angeles. From 1974-78, 26 percent of
homicides, forcible rapes, and robberies occurring in Los
Angeles took place in the Southwest, Seventy-seventh Street,
and Southeast Divisions of the Los Angeles Police Department.
Thirty-three percent of police shooting incidents involving sus­
pects within the city limits of Los Angeles occurred in these
three divisions, as did 31 percent of shootings in which a sus­
pect was hit and 34 percent of fatal shootings of suspects by the
Los Angeles Police Department~

Circumstances ofShootings

A greater proportion of shootings at blacks than at Hispan­

ics and whites followed suspects' disobeying officers' orders to

159
halt and suspects' appearing to reach for weapons. Table 2
shows that from 1974-78, 15 percent of shooting incidents in­
volving blacks were preceded by suspects' disobeying an offi­
cer's order to halt, and 12 percent were preceded by suspects'
appearing to reach for weapons. Nine percent of Hispanic sus­
pects were shot at after disobeying orders to halt and six per­
cent after appearing to reach for weapons; the corresponding
proportions for whites were nine percent following disobeying
orders to halt and nine percent after appearing to reach for
weapons. The proportion of black suspects shot at after display­
ing, threatening to use, or actually using a weapon was 66
percent, whereas 74 percent of Hispanics and 76 percent of
white suspects were shot at under these circumstances.
A greater proportion of blacks than of Hispanics or whites
shot at by the Los Angeles Police Department from 1974-78
were ultimately determined to have been unarmed . A some­
what higher percentage of blacks than of Hispanics or whites
were carrying guns when they were shot at, but a lower per­
centage of blacks than of Hispanics and whites had other weap­
ons, such as knives, blunt instruments , and so forth . Table 3
shows that of blacks involved in shooting incidents with the Los
Angeles Police Department, 28 percent in fact possessed no
weapon when they were shot at. Twenty-two percent of His­
panics and 20 percent of whites were ultimately determined to
be unarmed. Fifty-four percent of blacks shot at possessed guns
compared with 48 percent of Hispanics and 49 percent of
whites; 18 percent of blacks, 30 percent of Hispanics, and 31
percent of whites had other weapons.
Changes from 1977 to 1978, which reduced shootings at
suspects disobeying officers' orders to halt or appearing to
reach for weapons-where there was no assault and no use,
display, or threat of a weapon-and of unarmed suspects, di­
minished the frequency with which blacks and Hispanics were
involved in these kinds of shootings. Thus eight-of 57­
shootings at blacks in 1978 were precipitated by disobeying
officers' orders or by appearing to reach for weapons compared
with an average of 19.75 such shootings per year from 1974-77.

160
 TABLE 2 SUSPECTS' ACTIONS PRECIPITATING SHOOTINGS , BY RACE OR DESCENT, 1974-78 (IN
PERCENT)

Black Hispanic White

Using weapon 22 23 28
Threatening use of weapon 39 45 43
Displaying weapon 5 6 5
Assaulting officer or civilian 5 9 6
Appearing to reach for
weapon• 12 6 9
Disobeying command to
halt* 15 9 9
Other precipitating action,
including accidental
shootings at suspects 3

100 100 101

Number (not in percent) 321 126 131

• Disobeying command to halt or appearing to reach for weapon were coded only if no assault took place
and there was no use , threat, or d isplay of a weapon in the period immediately preceding the shooti ng. Assau lt
was coded only if there was no use , threat, or display of a weapon . For each person shot at, only one
precipitating event was coded-the most life endangering .
t-0
a>
t-0
.....
cr>
1:-:>

TABLE 3 SUSPECTS' WEAPONS, BY RACE OR DESCENT, 1974-78 (PERCENTAGES)

Black Hispanic White

No weapon 28 22 20
Gun 54 48 49
Other weapon , including automobile 18 30 31
- -
100 100 100
Number (not in percent) 321 126 131
 Eleven blacks shot at in 1978 were found to be unarmed com­
pared with an average of 20 from 1974-77. In 1978, one
Hispanic-of 20-was shot at following disobeying orders to
halt or by appearing to reach for a weapon-compared with an
average of 4.5 from 1974-77-and none was armed-compared
with the 1974-77 average of 4.5. Two whites-of 20-were also
shot at after disobeying orders to halt or appearing to reach for
a weapon - compared with an average of 5.5 from 1974-77­
and three white suspects were in fact unarmed-compared with
5.75 per year from 1974-77. 1

No significant difference has existed between blacks and

other suspects in the number of shots fired when other circum­
stances surrounding shooting incidents are controlled, although
under some circumstances fewer shots are fired at Hispanics
than at others. In shootings precipitated by disobeying an offi­
cer, appearing to reach for a weapon, or assault, blacks are
fired upon an average of 2.44 times, Hispanics 1. 73 times, and
whites 2.41 times. The mean number of shots fired when a
suspect displayed a weapon, threatened to use it, or actually
used it was 4.85 for blacks, and 4.78 for Hispanics, and 4.99 for
whites. The mean number of shots fired at blacks found to be
unarmed was 2.62 , unarmed Hispanics 1.50, and unarmed
whites 2.42 .

The Shooting Review Process

A brief discussion of the shooting review process is re­
quired. The reader is cautioned that the only information about
the review process we have is its result: the finding as to
whether or not a shooting was in policy, in policy but failed to
meet departmental standards, accidental, or out of policy and
the action, if any, taken against the involved officers in the
1974-78 interval. An in-policy finding, it should be noted, indi­
cates the department's judgment that the officer reasonably
believed himself or others to have been in danger of death or
serious bodily injury or that he knew a fleeing suspect to have
committed a felony. With the adoption in late 1977 of the cur·
rent policy restricting shootings at fleeing felons, the officer

163
was required to know that the suspect had committed a felony
involving death or serious bodily injury. We have no informa­
tion concerning informal discussions among review board mem­
bers or their interviews with investigators and witnesses that
could potentially yield evidence not in the written record, nor
do we have information about informal discussions that may
have entered into the final classification and the administrative
action taken, if any, against the officer. Prior to 28 November
1978, the classification of a shooting and administrative action
were under the jurisdiction of the director-assistant chief­
Office of Operations. The director-assistant chief-Office of
Special Services, had this responsibility for the following year. 9
When all shooting incidents that involved suspects occur­
ring from 1974-78 are considered, only small differences in
results from the shooting review process for blacks compared
with Hispanics and whites appear. Eighty-two percent of shoot­
ing incidents involving black suspects, 77 percent involving His­
panics, and 80 percent involving whites were determined to be
in policy. Seven percent of shootings at black suspects, nine
percent of shootings at Hispanics, and 11 percent of shootings
at whites were found out of policy. In 85 percent of shootings
involving blacks, 80 percent involving Hispanics, and 79 per­
cent involving whites, there was either no administrative action
or only training was recommended. For all 1974-78 shooting
incidents, there was administrative disapproval in 18 pe rcent of
the cases, and in 10 percent of incidents, an involved officer
was penalized by loss of days off, suspension, or termination. 10
While differences in outcomes from all shooting reviews are
small, larger percentage differences, which are not statistically
significant due to the small number of cases involved, appear
between suspects of different descent when suspects' most
threatening actions just prior to shootings are taken into ac­
count. It was shown previously that a higher percentage of
blacks than others are involved in shootings following suspects'
disobeying orders to halt or suspects' appearing to reach for
weapons when there was no assault and no display, threat, or
use of weapon immediately preceding the shooting. These kinds

164
 Overall, the statistical results indicate that shooting inci­
dents involving black suspects have differed in numbers, cir­
cumstances, and under some circumstances, in results of the
shooting review process from shootings involving others. There
has been no difference between blacks and other suspects in the
number of shots fired. Few differences have appeared between
shooting incidents involving Hispanic suspects and shootings
involving whites.
RESEARCH QUESTIONS
Many questions are raised by these data describing the in­
volvement of minorities in Los Angeles Police Department
shooting incidents, but none can be answered conclusively from
the data presented here. One question concerns whether these
statistical patterns are unique to the city of Los Angeles or to
the Los Angeles Police Department. Another is how the differ­
ences between patterns of shootings involving blacks and shoot­
ings involving others are to be explained. A third question con­
cerns why relatively few differences appear between the
involvement of Hispanics and of whites in Los Angeles Police
Department shooting incidents.
Few data that are presently available permit comparison of
Los Angeles Police shootings involving minorities and such
shootings elsewhere. Data provided by the P olice Foundation
show that Los Angeles differs little in the relationship of shoot­
ings to Part I arrests from the seven cities included in Police
Use ofDeadly Force. 11 Of the seven cities included in Police Use
of Deadly Force, the differences between percentages of per­
sons shot who were black and Part I arrestees were higher than
in Los Angeles in two cities but lower in five cities . Whereas 46
percent of Part I arrestees and 53 percent of persons shot in
Los Angeles from 1974-78 were black, the corresponding pro­
portions for blacks are 83 percent of Part I arrests and 80
percent of shootings in Birmingham; 76 percent of Part I ar­
rests and 76 percent of shootings in Oakland; 27 percent of
Part I arrests and 44 percent of shootings in Portland; 61 per­
cent of Part I arrests and 62 percent of shootings in Kansas
City; 53 percent of Part I arrests and 64 percent of shootings in

167
 Indianapolis; 94 percent of Part I arrests and 89 percent of
shootings in Washington, D.C.; and 83 percent of Part I arrests
and 80 percent of shootings in Detroit. Statistics concerning
the involvement of Hispanics in police shootings are presently
unavailable for other cities, as are data on events precipitating
shootings at minority suspects, the frequency with which these
suspects were armed, and administrative determinations con­
cerning police shootings involving minorities. In sum, data now
available neither show the Los Angeles pattern of shooting inci­
dents involving minorities to be unique nor do they indicate
that the same pattern holds nationwide. Much more evidence is
clearly needed.
A wholly different question is how one accounts for the
numbers and circumstances of shootings involving blacks com­
pared with shootings at other suspects. The empirical results
can lead quickly to inferences concerning attitudes and motiva­
tions of police officers generally. This could be mistaken, since
observers of police behavior, including the most critical, agree
that officers believe themselves to be unbiased and evenhanded
in extending justice, separating individual attitudes from offi­
cial conduct. 12 For this reason and in the absence of information
concerning individual attitudes and motivations, attention
should be directed toward the context or environment in which
the Los Angeles Police Department serves, the department's
understanding of this environment, and the consequences of
both for shootings at minorities. To do this requires a brief
excursion into organizational theory.
The link between environments and organizations remains
the central focus of organizational theory. However, there is a
sharp division in the field between work treating environments
as wholly external to organizations and work treating environ­
ments as external elements that are perceived or enacted by
organizations themselves. This disagreement as to what consti­
tutes organizational environments is much more than a matter
of semantics. Many investigators find little correspondence be­
tween perceptions of the environment and more objective mea­
sures. 13 Furthermore, perceived environmental elements may

168
 predict organizational actions as well as or better than the
more objective measures. 14
This study of Los Angeles Police Department officer­
involved shootings yields certain objective as well as perceptual
measures of the environment. Data on arrests, attacks on offi­
cers, and serious crimes constitute the objective measures. Out­
comes of shooting reviews may be taken as measures of the
perceived environment since the review process determines
whether or not an officer reasonably believed himself or others
to have been in danger or to have been confronted by a fleeing
felon whose escape would endanger others.
All of these measures of the environment have deficiencies.
Arrests, like shootings, are ultimately the result of police offi­
cers' decisions and may be due to the same causes as shootings.
Reported attacks and assaults with deadly weapons upon offi­
cers also reflect to some extent police discretion . Shooting re­
views, on the other hand, do not necessarily comport with what
officers actually believed at the time of shooting incidents. The
measures of the environment are also incomplete. We have no
count of the frequency with which situations potentially leading
to shootings occur. We know only the frequency of shootings .
We do not know how often officers actually confront dangerous
fleeing felons, but do not shoot. We do not know how often
suspects make furtive gestures appearing to officers as at­
tempts to reach for a weapon. We know only how often these
events are followed by shooting incidents. And we do not know
the frequency with which officers believe themselves or others
to be in imminent danger. Again, we know only how often this
has been the case in shooting incidents.
Despite deficiencies in the environmental measures, the
data at hand are sufficient to illustrate how profoundly differ­
ent conclusions may result from the use of different measures
of the environment. If one compares patterns of shooting inci­
dents with objective measures-arrest, attack, and crime
rates-then one would conclude that blacks have been dispro­
portionately involved in Los Angeles Police Department shoot­
ing incidents, especially those shooting incidents precipitated

169
by suspects' disobeying officers and appearing to reach for
weapons, as well as those incidents involving unarmed sus­
pects. But if one instead compares shooting incidents with mea­
sures of how the environment is perceived-the department's
judgment as to whether or not the officer reasonably believed
himself or others to have been in imminent danger or confront­
ing a dangerous fleeing felon-then one would conclude exactly
the opposite, namely, that no disproportion exists between
shooting incidents involving blacks and shootings involving oth­
ers and that large numbers of whites and Hispanics are in­
volved in those shooting incidents precipitated by disobeying
officers' orders or by appearing to reach for weapons as well as
those incidents involving unarmed suspects.
The lesson to be drawn from this exercise is clear. Con­
cretely, it has been shown that depending upon whether one
gauges the environment of police work using objective mea­
sures of arrests, attacks, and dangerous crimes as opposed to
the perceived measures of reasonableness of an officer's sense
of endangerment, one could conclude either that too many
blacks have been shot at by Los Angeles police officers or that
too many whites and Hispanics have been shot at under some
circumstances. A discrepancy exists, then, between the level of
endangerment indicated by the best objective indicators that
we could find and the department's judgments concerning the
degree of danger present when officers confront suspects of
different race or descent. Our objective indicators may be
faulty, but so may be the department's perceptions. More gen­
erally, it has been shown that entirely opposite inferences may
be drawn when perceptual measures of the environment are
used in place of objective measures, and vice versa. How one
gauges the environment, then, may be consequential for statis­
tical results as well as for implications drawn on the basis of
them. A compelling need for complete and reliable measures of
the environment of police work is indicated.
Finally, the absence of substantial differences between
shootings at Hispanic suspects and shootings at whites in Los
Angeles also requires exploration. It is of some importance to

170
know whether large numbers of minorities as opposed to large
numbers of blacks have been involved in police shootings in the
United States. If Los Angeles is representative of most U.S.
cities, then one would conclude the latter to be the case. The
need for reliable national data on police firearms discharges is
again indicated. At the same time, explanation of the similar
involvement of Hispanics and whites in Los Angeles Police
shooting incidents would be facilitated were there better mea­
sures of the environment than those now available. Substantial
research on police use of firearms as well as of other forms of
force remains to be done.
Notes
1. The SLA shoot-out was excluded from our computer files because it
would have distorted grossly certain shooting statistics. More than
5,000 rounds-plus 83 tear gas canisters-were fired by Los Angeles
Police officers in the SLA incident, more rounds than the total fired in
the remaining 912 officer-involved shootings analyzed here.
2. Bystanders and hostages include persons shot at whom officers
mistook for suspects when in fact a suspect was present or nearby, as
weB as persons hit unintentiona1ly by officers' shots aimed at suspects.
Accidental discharges include all incidents ruled accidental by shooting
review boards, except for those occurring in tactical situations in
which officers may have had cause to fire deliberately. Other nonacci­
dental shootings include shots fired at cars and at street lamps and
warning shots.
3. No officer was involved in more than six shooting incidents in the
1974·78 period.
4. Appearing to reach for a weapon is often called "furtive move­
ment" in departmental investigations and reports.
5. For example, 26 persons were shot fatally by the Los Angele s
Police Department in 1974, 30 in 1975, 30 in 1976, 33 in 1977, 20 in
1978, and 14 in 1979. The number of suspects shot at following their
disobeying an officer's command to halt or appearing to reach for a
weapon was 36 in 1974, 39 in 1975, 39 in 1976, 32 in 1977, and 11 in
1978. Of all suspects shot at, 39 were ultimately determined to have
been unarmed in 1974, 34 in 1975, 34 in 1976, 32 in 1977, but only 14
in 1978. The decline in Los Angeles Police shootings from 1974
through 1979 is reported in detail in Part IV of "The Report of the
Board of Police Commissioners" and will be the subject of a separate
article.

171
6. Descriptive data concerning 695 shooting incidents were included
as part of the "Enactment Development Plan" for the DEFT shooting
simulator, which is now in operation. Whether the 695 incidents in­
clude all shootings in the 42-month period covered is not stated
clearly. Data for Hispanics were not included in this document.
7. Part I offenses include violent and some nonviolent crimes: murder,
forcible rape, robbery, aggravated assault, burglary, larceny, theft,
and auto theft.
8. A similar comparison cannot be made for the Hispanic community
since the one preponderantly Hispanic police division in Los Angeles,
Hollenbeck, is small and accounts for only three percent of homicides,
forcible rapes, and robberies in the city.
9. In 1979, direct responsibility was transferred to the Board of Police
Commissioners and the Chief of Police.
10. These data also speak to the adequacy of the categories used to
evaluate shootings. Whereas less than 9 percent of shootings were
judged out of policy, 18 percent resulted in some form of administra­
tive disapproval. In other words, there were a fair number of shoot­
ings that were out of policy but were disapproved, or put somewhat
differently, not out of policy but not approved.
11. Catherine H. Milton et a!., Police Use of Deadly Force (Washing­
ton, D.C.: Police Foundation, 1977).
12. Jerome H. Skolnick, Justice Without Trial (New York: John Wiley
& Sons, 1966), ch. 4.
13. Henry Tosi et al., "On the Measurement of the Environment: An
Assessment of the Lawrence and Lorsch Environmental Sub-Scale,"
Administrative Science Quarterly 18:27-36 (1973); H. Kirk Downey et
al., "Environmental Uncertainty: The Construct and Its Application,"
Administrative Science Quarterly 20:613-29 (1975).
14. Cf. Paul R. Lawrence and Jay W. Lorsch, Organization and Envi­
ronment (Cambridge, Mass.: Harvard University Press, 1967) with
Tosi eta!.

172
 CHAPTER 7

RACE AND EXTREME POLICE-CITIZEN

VIOLENCE

JAMES J. FYFE
The disproportionate representation of blacks among the
clientele of the criminal justice system is a recurrent theme in
the literature (Wolfgang, 1964: 51). Research into police fire­
arms use (e.g., Clark, 1974: Milton et al. 1977) reports that
minority disproportionality is also an explosive issue. Goldkamp
(1976: 183), however, accurately notes that the present paucity
of analysis of this phenomenon leaves criminal justice agencies
free to adopt either of two empirically unsubstantiated "belief
perspectives."
Goldkamp briefly defines those "belief perspectives" as fol­
lows:
Some writers suggest that the disproportionately high
death rate of minorities at the hands of the police can
best be e~lained by the disproportionately h1gh arrest
rate for cnmes of violence, or by assumptwns concern­
ing the suspect's responsibility for his/her own death in
violent pohce-suspect interactions. Others see dis.Pro­
portionate minority deaths as resulting from both Irre­
sponsible use of deadly force by a small minority of
police officers and differential administration of law en­
forcement toward minority citizenry (which in effect
produces disproportionately high arrest and death rates
for minorities in general). Kobler and Knoohuizen,
SOURCE: Reprinted from "Race and Extreme Police·Citizen Violence" by
James J. Fyfe in R. L. McNeely and Carl E. Pope (eds.) Race, Crime, and
Criminal Justice, Beverly Hills, Calif.: Sage Publications, 1981. By permission
of the publisher.

173
 Fahey and Palmer stress the possibility that police mis­
conduct may play a considerable role in generating ci­
vilian deaths. Takagi ascribes disproportionality to the
simple fact that "police have one trigger finger for
whites and another for blacks" [1976: 169].
Harding and Fahey, whose position seems to straddle both
perspectives, suggest that police misconduct is involved in
shooting deaths. Conversely, they also state that the use of
deadly force by police "is not an independent aspect of the race
problem." They write that:
Police conduct is a dependent aspect of general pat­
terns of criminal behavior, patterns that are signifi­
cantly influenced by broader considerations of, for ex­
ample, age, class, and affluence (The) (v)ictimization of
those shot is directly related to contacts of the sort in
which firearms are most frequently used by criminals
[1973: 310].
RACE AND EXTREME POLICE-CITIZEN VIOLENCE:
TWO HYPOTHESES
Harding and Fahey suggest, therefore, that racial dispro­
portionality among police shooting victims may be related to
racial variations among other indices of violence . To some de­
gree, this assertion is supported by the work of Kania and
Mackey (1977), who found that such variations among fatal po­
lice shooting rates across the 50 states were closely related to
variations in reported violent crime and criminal homicide
rates. Because of our access to data on extreme police-citizen
violence in New York City, these prior efforts also suggested
two hypotheses which became the focus of this article.
First, on the theory that police shootings are a corollary of
the frequency of contacts which present the opportunity for
such violence, we postulated that:
H1: Blacks and Hispanics would be overrepresented
among police shooting opponents in relation to their
representation in the New York City population but
there would be less disproportion by race taking into
account the racial representation of arrests for VIolent
crime.

174
 Second, to test the assumption that police shootings are
related to other indices of extreme violence among the races:
H2: Black and Hispanic overrepresentation among po­
lice shooting opponents in New York City would be re­
duced by taking into account the racial representation
in reported murders and nonnegligent manslaughters.

DATA SOURCES
Our major data source for these analyses consisted of New
York City Police Department records of all incidents in which
officers reported discharging weapons and/or being subjects of
"serious assault" (e.g., assault with deadly weapon and/or
which resulted in officer death or serious injury) during the
years 1971-1975. These data included "Firearms Discharge/
Assault Reports" (FDAR's) filed by 4904 officers, of whom
3827 reported discharging firearms in 2926 separate "shooting
incidents." Since not all these involved shooting at other human
beings, we excluded from analysis such events as shootings to
destroy animals, warning shots, and officer suicides. Because of
the relatively low number of female opponents included in our
data and because they were often involved in non-line of duty
shootings (Fyfe, 1978: 145-149), we also excluded them from
analysis. Conversely, because our detailed examination of the
data had convinced us that the frequency of police use of fire­
arms as a means of deadly force is best measured in terms of
officer decisions to point and fire at other human beings, we
included for analysis all such incidents, without regard to their
consequences: with rare exception, missed shots, woundings,
and fatalities are only chance variations of equally grave deci­
sions.
While we considered this shooting data base nearly ideal for
our purposes, 1 we found that available U.S. Census Bureau fig­
ures were less informative. New York City base population fig­
ures provided by the census define only two major racial
groups, "White," which includes "Mexican, Puerto Rican, or a
response suggesting Indo-European stock" (U.S. Department
of Commerce, 1973: B34, App7-8) and "Negro." This inclusion

175
of Puerto Ricans and other Hispanics into the "White" cate­
gory is not a major limitation insofar as measurement of shoot­
ing opponent racial disproportion is concerned, however, since
New York City also develops its own population racial distribu­
tions.
More surprisingly, we also found-with one exception­
that the New York City Police Department does not compile
racial statistics of arrestees. 2 In the absence of this ideal data
source, we decided to employ two surrogate scales of compari­
son. For our analysis of H1, we elected to use Burnham's sam­
ple of the races of 700 persons arrested for murder, nonnegli­
gent manslaughter, robbery, forcible rape, and felonious assault
in New York City (1973: 50). 3 Because we were similarly pre­
cluded from using homicide arrestee data in our examination of
H2, we decided to employ information on the race of homicide
victims. Here, since many studies show that homicides tend to
be intraracial crimes, it might not be unreasonable to consider
the victim index as a crude proxy for the perpetrator index.

H1: ANALYSIS

Given these limitations, our data do indicate that Blacks

and Hispanics are disproportionately represented among those
New York City police shooting opponents whose race is re­
corded. Table 1 demonstrates that Whites, who comprise 64.1 o/o
of New York City's population total , represent only 17.5% of its
shooting opponents. Blacks, conversely, are overrepresented
among shooting opponents in almost exactly the reverse ratio:
60.2% of shooting opponents and 20.5% of total city population
are Black. Hispanics, who number 15.4% of city population,
represent 22.3% of shooting opponents.
Using the racial composition of the general population of
the city as expected frequencies, we derived a chi-square for
these distributions which is significant at the .001 level. Simi­
larly, our obtained Cramer's v (.49)4 suggests a rather high
association between race and the likelihood of being shot at by

176
New York City police. The first part of this hypothesis is there­
fore confirmed.
Turning now to a model (Table 2) which utilized the ethnic
distribution of felony arrests for violent crimes (Burnham's
data) to generate expected numbers of shooting opponents by
ethnicity, it may be seen that there is a fairly close fit. Al­
though the chi-square is still significant, the v value is now only
.09, or much smaller than the v value of .49 reported for Table
1. While Whites remain slightly underrepresented among
shooting opponents, Blacks are also underrepresented. Only
Hispanics are overrepresented. Some caution must be observed
about the finer distinctions in view of the limitations of data on
ethnic classifications.

TABLE 1 RACIAL DISTRIBUTION OF NEW YORK CITY POLICE

SHOOTING OPPONENTS, AND NEW YORK CITY
POPULATION JANUARY 1, 1971-DECEMBER 31, 1975

Shooting New York City

OpponentS' Population

White 17.5% (549) 64.1% (5076022)

Black 60.2 (1889) 20.5 (1621583)
Hispanic 22.3 (701) 15.4 (1216557)
TOTALS 100.0 (3139) 100.0 (7914162)

chi-square = 1488.40
p = .001
v = .49
• Excludes "Other" racial categories because of low statistical signifi­
cance. Ten shooting opponents were identified as members of
"Other" racial groups.
• Calculated from: New York City Police Department, Chief of Field
Services, Summary of Precinct Populations, 1973.

177
 TABLE 2 RACIAL DISTRIBUTION OF NEW YORK CITY POLICE
SHOOTING OPPONENTS, AND PERSONS ARRESTED
FOR FELONIES AGAINST THE PERSON JANUARY 1,
1971-DECEMBER 31, 1975.

Shooting Felony

Opponents Arrest~

White 17.5% 22.2%

Black 60.2 62.4
Hispanic 22.3 15.4
TOTALS 100.0 100.0

chi-square = 12.82
p = .01
v = .09
• Calculated from a sample of 700 persons arrested for murder, non­
negligent manslaughter, robbery, felonious assault and forcible rape
in New York City, 1971 . Source: David Burnham, "3 of 5 Slain by
Police Here are Black, Same as Arrest Rate," The New York Times,
August 26, 1973, 50.

H2: ANALYSIS
Using our murder and nonnegligent manslaughter victim
racial distributions to generate expected frequencies of shoot­
ing opponents by race (Table 3), we find that there is a fairly
close fit. Again Cramer's v is only .09, which suggests that
there is a close parallel between the racial distributions of homi­
cide victims and police shooting opponents.

178
TABLE 3 RACIAL DISTRIBUTION OF NEW YORK CITY POLICE
SHOOTING OPPONENTS, JANUARY 1, 1971-DECEMBER
31, 1975 AND VICTIMS OF MURDER AND
NON-NEGLIGENT MANSLAUGHTER,• JANUARY 1,
1971-DECEMBER 31 , 1975.

Shooting Homicide
Opponents Victims

White 17.5% 22.5%

(549) (1069)
Black 60.2 51.0
(1889) (2419)
Hispanic 22.3 26.5
(701) (1259)
Totals 100.0 100.0
(3139) (4747)

chi-square = 54.68
p = .001
v = .09
• Source: New York City Police Department, Homicide Analysis Unit,
Annual Report, 1976

TABLE 4 AGE CHARACTERISTICS OF NEW YORK CITY MALE

POPULATION BY ETHNIC GROUP

Puerto

Black Rican White

Median Male Age 23.1 19.4 33.3

Percentage of
Male Population 42.0 45.6 27.6
Under 18

SOURCE : U.S. Department of Commerce , Characteristics of the Pop­

ulation, Part 34 NY Section 1, March 1973 U.S. Govern­
ment Printing Office, pp . 34-108, 34-432.

179
FURTHER COMMENTS
Even if both the relationships shown in Tables 2 and 3 were
demonstrated by more comprehensive arrest and victim data,
they would not prove that the disproportionate involvement of
Blacks and Hispanics in shooting incidents is related to their
disproportionate involvement as violent crime arrestees and
homicide victims.
Most specifically, we would still be left with the possibility
that both relationships are spurious and merely reflections of
varying degrees of risk due to differential age distributions
and/or differential enforcement and police deployment prac­
tices. It is still possible that the races are differentially repre­
sented among shooting opponents because more Blacks and
Hispanics fall into the age groups most frequently involved in
these incidents. Alternatively, it is still possible that Blacks and
Hispanics are disproportionately represented among shooting
opponents because police do have "one trigger finger for
Whites and another for minorities." Table 4, which summarized
the age characteristics of New York City's male Whites,
Blacks, and Puerto Ricans (the city's major Hispanic subpopula­
tion), confirms the existence of differential age distributions
among these groups. As Table 4 indicates, New York City's
White males are generally considerably older (median age =
33.3 years) than either its male Blacks (median age = 23.1
years) or its male Puerto Ricans (median age = 19.4 years).
To determine whether similar age discrepancies existed
among the shooting opponents included in our data, we cross­
tabulated opponent race and age. The results of this analysis
are presented in Table 5 and demonstrate that the age distribu­
tions of shooting opponents vary littl_e among races, a finding
that runs counter to those of Jenkins and Faison (1974) and
Kobler (1975). Indeed, our obtained levels of p (.99) and v (.03)
indicate that these three age distributions are so close a s to be
nearly indistinguishable.
Table 5 also suggests that confrontation with armed police
is largely an activity of the young .S More than half (1093) of the
2149 opponents whose race and age are included in our data set

180
 TABLE 5 NEW YORK CITY POLICE SHOOTING INCIDENT OPPONENT RACE BY AGE,
JANUARY 1, 1971-DECEMBER 31, 1975

Opponent Age
Opponent Under Totals
Race 16 16, 17 18, 19 20,21 22,23 24,25 26,27 28,29 30+

White 3 .2% 10.6% 15.0% 14.5% 10.6% 11 .3% 4.9% 7.2% 22.7% 18.9%
(13) (43) (62) (59) (43) (46) (20) (29) (92) (407)
Black 3.6 8 .7 12.8 13.7 12.5 12.7 7 .5 5.8 22.7 57.6
(44) {107) (185) (169) (155) (158) (93) (72) (280) (1236)
Hispanic 2.0 11 .1 10.5 13.1 10.7 11 .1 7 .9 5 .9 27 .7 23.5
(10) (56) (53) (66) (54) (56) (40) (30) (140) (505)
Totals 3.1 9 .6 12.7 13. 7 11.7 12.1 7 .1 6.1 23 .9 100.0
(67) (206) (273) (294) (252) (260) (153) (131) (512) (2148)

chi-square = 3.866
p = .99
v = .03

,.....
00
,.....
are less than 24 years old. Older opponents are not entirely
excluded, however, since almost one-quarter (23.9o/o) of the
group are 30 or more years old. 6

TABLE 6 NEW YORK CITY MALE SHOOTING OPPONENT RATE

PER 10,000 POPULATION, BY AGE AND RACE,•
JANUARY 1, 1971-DECEMBER 31, 1975

Age White/Hispanic Black

15-19 10.04 38.82

20-24 11.40 67.38
25-29 7.62 43.06
30-34 6.02 21.92
35-39 3.88 15.02
40-44 1.77 9.24
45-49 .98 4.54
50+ .16 1.86
Totals 3.95 24.20

• Excludes opponents under fifteen years of age (White/Hispanic n =9;

Black n =17).

Table 6 presents the frequency of shooting opponents per

10,000 population aggregated by Census Bureau age and race
subgroups. Examination of Table 6 reveals that White/Hispanic
and Black age distributions are very similar. For both race
groupings, those 20 to 24 years old are most frequently in­
volved in shootings; subsequently the rates steadily decline for
each age subgroup, reaching the lowest levels among those 50
or more years old. The only major discrepancy in age trends
occurs for the 15-to-19-year age group; the White/Hispanic
male rate is only slightly less than that of the 20-to-24-year
group, while for Blacks the rate of 15-to-19-year-olds is substan­
tially lower than for 20-to-24-year-olds.
Equally as striking as the general similarities in age pat­
terns are the great numeric discrepancies in rates. Specifically,
we find that Black males are six times more likely to have been

182
involved in police shooting incidents (24.20 per 10,000 popula­
tion) than are male White/Hispanics (3.95 per 10,000). Indeed,
the overall Black male rate (24.20) is more than twice as large
as the highest White/Hispanic male rate (11.40). The general
discrepancy ratio holds in each and every age group and is
strongest within the 20-29 year range.

RACIAL DISPROPORTION
To this point, our investigation has shown that the variable
of race is linked to the likelihood of being a police shooting
opponent; in a similar fashion, this risk factor is apparently
linked to arrest rates for violent felonies. The possibility re­
mains that the great numeric disproportion of minorities among
both arrestees and shooting opponents is a function of differen­
tial police enforcement or deployment practices.
Since most prior literature which addresses racial dispro­
portion among opponents (e.g., Goldkamp , 1976; Harding and
Fahey, 1973; Robin, 1963; Jenkins and Faison, 1974) examines
only fatal shootings, we commenced our investigation with an
analysis of incident consequences in terms of opponent injury.
Table 7 provides a crosstabulation of reported shooting oppo­
nent race by injury and demonstrates that these consequences
vary little among the races. Regardless of race, approximately
3 in 5 opponents suffer no injury, 1 in 5 is wounded, 1 in 10 is
killed, and 1 in 10 escapes after the police have shot at him with
unknown effect. Once an officer decides to employ his "trigger
finger," the race of his opponent apparently matters little in
terms of the effect of police shots. Blacks escape with unknown
injuries approximately twice as often (16.6%) as Whites (7.9 %)
or Hispanics (9.8%), but our obtained chi-square (p = .50) indi­
cates that even this variance is likely to be a result of chance.
This lack of variance "within" confrontations, obviously,
does not address numeric disproportionality. Stated most sim­
ply, we can observe little difference among the races once they
become involved in conflict situations. We have not, however,

183
touched on the issue of why so many minority opponents be­
come involved in these incidents in the first instance.

TABLE 7 NEW YORK CITY POLICE SHOOTING OPPONENT RACE

BY INJURY JANUARY 1, 1971 -DECEMBER 31, 1975'

Opponent Injury
Opponent
Race None Wounded Killed Unknown• Totals

White 61.8% 21.0% 9.2% 7.9% 17.6%

(335) (114} (50) (43) (542)
Black 53.4 21.0 9.1 16.6 60.0
(983) (386) (168) (305) (1842)
Hispanic 56.0 22.2 12.1 9 .8 22.3
(384) (152) (83) (67) (686)
Totals 55.4 21.2 9.8 13.5 100.0°
(1702) (652) (301} (415} (3072)

chi-square = 5.45
p =.50
v = .03
• Excludes cases in which opponent race not reported; excludes 4
suicides (1 White, 2 Black, 1 Hispanic) .
• Not apprehended opponents at whom shots were fired with unknown
effect.
c Percentage subcells may not total 100.0 due to rounding.

To examine this question, we attempted to determine

whether there existed significant variation among the types of
shooting incidents in which different races became involved.
We hypothesized that if we found high frequencies of incidents
in which police shot at unarmed Blacks or Hispanics, it would
suggest that police did, indeed, have "different trigger fingers"
for minorities. Conversely, we felt that significant differences
among the races in the precipitating event types and the degree
of danger confronting police might help to explain the dispro­
portion of minority opponents.

184
 To simplify this process and clarify its results, we decided
to employ the race of each incident's "primary opponent." 7 This
resulted in very little loss of accuracy since incidents involving
multiple opponents are overwhelmingly intraracial events. 8
Further, since our operative definition of "primary opponent"
translated into either the only opponent or the one posing the
greatest threat to police (e.g., the most combative, most heavily
armed), we concluded that it was this person's conduct upon
which police reaction (or overreaction) would be principally
based . Thus, a shooting precipitated by a robbery involving one
Black suspect armed with a gun and a White one armed with a
knife becomes a "Black opponent" incident.
Our first measure of variance in shooting incident types
among the races involved an investigation of the events which
precipitated police shooting. As Table 8 and its chi-square sig­
nificance level (.00001) reveal, there is considerable variance
here. Perhaps most striking is the great frequency with which
police confront Blacks at robberies. Indeed, nearly half (45.8%)
of the incidents involving Blacks were reportedly initiated by
robberies. This is a rate nearly twice that of Whites and His­
panics (23.4% and 26.3%, respectively) and represents a raw
frequency (495 incidents) greater than the total of all incidents
involving either Whites (354) or Hispanics (429). We see, in
fact, that robberies involving Black primary opponents com­
prise 26.6% of all the incidents included in Table 8.
Table 8 also demonstrates that Whites are disproportion­
ately more frequently counted among those who confront police
at burglaries (12.4% versus 6.4% and 9.1% for Blacks and His­
panics, respectively). In addition, perhaps because of the gen­
eral relationship between race and social status, the percent­
ages of "Car Stop" incidents involving Blacks (11.3%) and
Hispanics (8.4%) are far smaller than those of Whites (19.5%).
Similarly, the percentage of "Other" incidents, which often in­
clude offduty disputes and the like, is greater for Whites
(10.5%) and Hispanics (9.3%) than for Blacks (4.0%). Con­
versely, Whites are less frequently involved in generally proac­
tive "Investigative Suspicious Person" -or "Stop and

185
 Question"-incidents than are Blacks and Hispanics (7.3% ver­
sus 11.0% and 16.6%, respectively). Finally, Table 8 reveals
considerable difference in the frequencies of " Respond to Dis­
turbance" incidents: 15.6% for Hispanics, 10.2% for Whites,
and 9.0% for Blacks.
A measure of the threat of officer safety at these incidents
is provided by Table 9, which crosstabulates primary-opponent
race with weapon type. Here again, it can be seen that striking
differences exist among the races, with the chi-square proving
significant to .0001. Approximately half of Black and Hispanic
opponents (52.9%) and 48.0%) were armed with handguns.
Nearly half (47.2%) of all incidents included in Table 9 were
police confrontations with Hispanic or Blacks armed with hand­
guns, rifles, machine guns, or shotguns; White handgun, rifle,
machine gun, or shotgun incidents account for 6.9% of all inci­
dents. Conversely, we find that Whites are more frequently
involved in incidents involving no weapon or no assault on po­
lice (15.5%) than are Blacks (7.8o/o) or Hispanics (5.1 %). Whites
are also overrepresented in incidents involving the use of vehi­
cles (16.1% versus 6.1% for Blacks, 6.3% for Hispanics) or
physical force (9.4% versus 4.2% and 5.1o/o) as means of assault­
ing police. Hispanics use knives against police considerably
more often (21.3%) than do Whites (13 .3%) or Blacks (14.1o/o).
Given that most police are killed or seriously injured by guns or
knife wounds, therefore, we would tentatively conclude that
Blacks and Hispanics are more often involved-both propor­
tionately and in terms of sheer numbers-in incidents that
present greater potential danger to police than are Whites.
That potential danger does not nece ssarily translate into
real negative consequences in terms of officer injury is indi­
cated by Table 10, which provides a crosstabulation of FDAR
incident primary-opponent race by degree of officer injury (ex­
cluding non-line of duty injuries, which are not relevant to this
analysis). Here, although the nature and seriousness of nonfatal
injuries are not specified and, in fact, vary considerably, it can
be seen that proportionately more officers are injured in en­
counters with Whites (22.8%) or Hispanics (18.0 %). Proportion­
ately more officers are killed in the line-of-duty by Blacks

186
 TABLE 8 NEW YORK CITY POLICE SHOOTING PRIMARY OPPONENT RACE BY PRECIPITATING EVENT,
JANUARY 1, 1971-DECEMBER 31, 1975

Respond Invest-
Primary to Attempt gating Auto Assault
Opponent Distur­ other Handling Suspicious Mentally Pursuit! on
Race bance Burglary Robbery Arrests Prisoner Persons Ambush Deranged Stop Officer Other Totals

White 10.2% 12.4% 23.4% 4.2% 0.8% 7.3% 0.3% 2.5% 19.5% 8.8% 10.5% 19.0%
(36) (44) (83) (15} (3) (26) (1) (9) (69) (31) (37) (354)
Black 9 .0 6.4 45.8 3.1 1.6 11.0 0.9 1.5 11.3 5.4 4 .0 58.0
(97) (69) (495) (34) (17} (119) (10) (16} (122) (58) (43) (1080)
Hispanic 15.6 9 .1 26.3 4.9 1.2 16.6 0.9 2.1 8.4 5.6 9.3 23.0
(67) (39) (113} (21) (5) (71) (4) (9) (36) (24) (40) (429)
Totals 10.7 8.2 37.1 3.8 1.3 11.6 0 .8 1.8 12.2 6.1 6 .4 100.0
(200) (152) (691) (70) (25) (216) (15) (34) (227) (113) (120) (1863)

Not ascertained = 15
Chi-square = 151 .88078
p = .00001
v = .20

....
00
-::1
~
00
00

TABLE 9 NEW YORK CITY POLICE SHOOTING PRIMARY OPPONENT RACE BY WEAPON,
JANUARY 1, 1971-DECEMBER 31, 1975

Type of Weapon
Primary Rifle/ Knife/
Opponent Machine Cutting Physical
Race None Handgun Gun Shotgun Instrument Vehicle Force Other Total

White 15.5% 32.1% 1.4% 1.9% 13.3% 16.1% 9.4% 10.2% 19.2%
(56) (116) (5) (7) (48) (58) (34) (37) (361)
Black 7.8 52.9 1.3 6.3 14.1 6.1 4.2 7.4 57.8
(85) (574) (14) (68) (153) (66) (46) (80) (1086)
Hispanic 5 .1 48.0 2.6 3.1 21.3 6.3 5.1 8.4 22.9
(22) (207) (11) (14) (92) (27) (22) (36) (431)
Totals 8.7 47.8 1.6 4 .7 15.6 8.0 5.4 8.1 100.0•
(163) (897) (30) (89) (293) (15 1) (102) (153) (1878)
-­
chi-square = 131 .62032
p = .0000 1
v = .19
• Subcell percentages may not total 100.0 due to rounding.
 TABLE 10 NEW YORK CITY POLICE FIREARMS
DISCHARGE/ASSAULT INCIDENT PRIMARY
OPPONENT RACE BY OFFICER INJURY,•
JANUARY 1, 1971-DECEMBER 31, 1975

Officer Injury
Primary
Opponent
Race None Injured Killed Totals

White 76.4% 22.8% .8% 17.5%

(402) (120) (4) (526)
Black 82.2 16.6 1.2 59.6
(1471) (297) (21) (1789)
Hispanic 81.3 18.0 .7 22.8
(557) (123) (5) (685)
Totals 81 .0 18.0 1.0 100.Qb
(2430) .(540) (30) (3000)

chi-square = 1.84

p = .80

v = .02

• Includes only officers wounded or killed in the line of duty.

b Subcell percentages may not total 100.0 due to rounding .

(1.2%) than by Whites (.8%) or by Hispanics (.7%). 9 Although

Table lO's chi-square indicates that differences among these
distributions are not significant (p = .80) it is also important to
note that 55% of the line-of-duty officer injuries and deaths
occur with Black opponents.

SOME CONCLUSIONS
In summarizing this research within the context of prior
literature and the limits of our data, we are led to two major
conclusions. First, Harding and Fahey's assertion that minority
disproportion among police shoot ing opponents is related to
differential age distributions among the races is, in New York

189
 City at least, inaccurate. Our data demonstrate that, while po­
lice shooting opponents are generally young and a greater pro­
portion of the Black population is young, Black males in all age
groups are considerably more liable to become police shooting
opponents than are their White/Hispanics contemporaries.
Our second conclusion deals with whether that greater lia­
bility is associated with greater Black participation in activities
most likely to lead to justifiable extreme police-citizen violence
or with "the simple fact the 'police have one trigger figure for
whites and another for blacks."' Here we are led to choose
Goldkamp's "Belief Perspective II": Our data indicate that
Blacks make up a disproportionate share of shooting opponents
reportedly armed with guns and a disproportionate share of
those reportedly engaged in robberies when police intervened.
If one accepts both the accuracy of these reports and the prem­
ise that opponents armed with guns generally present the
greatest and most immediate danger to police, there is little to
support the contention that Blacks are shot disproportionately
in relatively trivial and nonthreatening situations. A more con­
clusive answer to the question would require the calculation of
shooting rates for specific arrest situations by race. As was
indicated earlier, the lack of race information on arrests pre­
cludes this analysis.
Although our research has not conclusively confirmed
Goldkamp's Belief Perspective II, it has reduced to two the
assumptions upon which one might base acceptance of the "po­
lice misconduct" and "different trigger finger" hypotheses im­
plicit in his alternate theory. First, of course, one might not
accept the accuracy of the reports of Black/gun incidents which
account for most of our data set's Black opponent dispropor­
tion. The sheer number of those shootings (656 are shown on
Table 9), however, is so large as to suggest that the argument
that "irresponsible use of deadly force by a small number of
police officers" accounts for disproportionate minority deaths is
ill-founded.
Second, one might accept the accuracy of these reports, but
properly note that we have not demonstrated that New York

190
 police do not refrain from shooting at Whites in situations com­
parable to those in which they do shoot at Blacks. Since most of
the Black opponents in our data set were reportedly armed
with guns, the assumption based on this observation requires
its proponents to argue that police generally regard Blacks
with guns as more threatening than Whites with guns. Our own
logic and experience, however, suggest that police responses to
such situations are based not upon opponent race, but rather
upon opponent weapon.
Finally, we must qualify our acceptance of Goldkamp's Be­
lief Perspective II . There is nothing in these analyses to sup­
port the contention that the disproportion of Blacks among
New York City police shooting opponents is reflective of police
misconduct or racial discrimination; but the limitations of our
data have prevented us from examining the degree to which
that disproportion is associated with the generally lower socio­
economic position of Blacks. Differences among the shooting
types which characterize the races (e.g., the high incidence of
Black participation in shootings precipitated by robberies,
which are most frequent in blighted inner city areas and the
high incidence of shootings involving Whites and vehicles, often
preceded by car thefts, which are most frequently in middle and
working class areas), however, suggest that this association
may be strong.
Were we to conduct further research based upon data
which included information about opponent socioeconomic stat­
us, we would hypothesize that Harding and Fahey's assessment
of the role of class and affluence in shooting opponent racial
disproportion would be confirmed. Were we successful, our re­
search would strongly indicate that Black shooting opponent
disproportion is neither a consequence of "overreaction" by
individual police officers nor of some racially varying predispo­
sitions toward violent crime. Conversely, it would point up the
continuing existence of An American Dilemma described so
well by Myrdal (1944) a generation ago; Blacks are the mode
among New York's police shooting opponents because they are
also the mode among the lower socioeconomic groups which

191
most frequently participate in the types of activity likely to
precipitate extreme police-citizen violence.

Notes

1. Our reservations here involve the accuracy and degree of detail

provided in the shooting incident reports filed by individual officers.
Because we attempted to limit our analyses to variables reasonably
immune to reporting bias (and often supported by witness statements),
we regarded the question of the veracity of the data we did possess as
one of minor importance. We were more troubled by the impact on our
research of the data we did not possess: 1058 (25.1o/o) of the data set's
shooting opponents were not identified by race. We did not find any
evidence that this information was deliberately withheld from reports
in order to prevent or avoid any sensitive racial issues. This is so for
several reasons. First, the "Firearms Discharge/Assault Report"
forms which served as our primary data source include no caption
requesting "opponent race," so that it appears only on reports filed by
officers who volunteered it. These forms, originally designed to collect
information for training purposes, have more recently been supple­
mented by more complete narrative reports. As a result, the annual
percentage of missing opponent race data declined from nearly 40o/o in
1971 to 5.4o/o in 1975. Second, despite this regular decrease in missing
data, annual known opponent racial distributions have remained rela­
tively constant over the period studied. Third, the percentage of miss­
ing opponent race data is relatively evenly distributed across the city's
police precincts, regardless of the racial characteristics of their total
populations (which one might reasonably expect to impact upon the
characteristics of their shooting opponents). Fourth, our opponent
data suggests that many of these opponents were never seen by the
officers involved: More than 9 of 10 "unknowns" (90.6%) suffered no
injury (72.6o/o) or escaped their confrontation with unknown injuries
(18.0%). Further, our opponent arrest data reveal that 397 (37.5o/o) of
opponents whose race is classified as unknown were not apprehended
by police: In many cases, these individuals were merely shadowy fig­
ures encountered on dimly lit streets or rooftops.
2. Except for homicides, the department began compiling these statis­
tics only in 1976. The department's Homicide Analysis Section began
systematically recording the race of those arrested for homicide in
1973. We elected not to use these data in our analyses because they do
not describe the experience for the full five years of our study and
because they do not identify homicide perpetrators who are not appre­
hended.

192
3. Burnham's sample is not random, but consists of 700 consecutive
arrestees.
4. See Loether and McTavish who describe Cramer's vas follows:
Cramer's v is, so to speak, a properly normed measure of association
for bivariate distributions of nominal variables, it is "margin free" in
that the number or distribution of cases in row or column totals does
not influence its value, nor is it influenced by the number of categories
of either variable . . . Cramer's v . . . can only be thought of as a
magnitude on a scale between zero and 1.0; the bigger the number, the
stronger the association. It can not be interpreted, for example, as the
percentage of variation in one variable explained by the other, nor can
it be interpreted as the proportions of predictive error which may be
reduced by prior knowledge of one of the variables [1974 : 197-198].
5. By aggregating our opponent age data into eight values to conform
with those reported on New York City Police arrest data, we also
found that the age distributions of shooting opponents and violent
felony arrestees closely parralleled each other (v = .09).
6. The decision to create this open-ended age grouping (which includes
individuals up to 79 years old) was made to simplify presentation and
discussion of opponent age. The frequencies of single-year values drop
off dramatically after this point.
7. We defined the "primary opponent" as the only opponent or, in
incidents involving more than one opponent, as the most heavily
armed and/or most aggressive and/or most seriously injured.
8. Intraracial events accounted for 94.4o/o of the multiple-opponent
shootings in which opponent race was reported.
9. These differences would shrink if four alleged politically motivated
"Black Liberation Army" assassinations were considered apart from
other officer deaths perpetrated by Blacks.

References
Breasted, M. (1974) "Police use cars and clubs to quell Brownsville
Riot." Ne:w York Times (September 20): 20.
Burnham, D. (1973) ""3 of 5 slain by police here are Black, same as
the arrest rate." Ne:w York Times (August 26): 20.
Clark, K. (1974) Open Letter to Mayor Abraham D. Beame and Police
Commissioner Michael J. Codd. New York, September 17, 1974.
Fyfe, J. J. (1978) "Shots fired: An examination of New York City
Police firearms discharges." Ph.D. dissertation, State University of
New York at Albany.

193
Goldkamp, J. S. (1976) "Minorities as victims of police shooting: Inter­

pretations of racial disproportionality and police use of deadly force."

Justice System J ournal, 2: 169-183.

Harding, R.W. and R.P. Fahey (1973) "Killings by Chicago police,

1969-70: An empirical study." Southern California Law Review, 4:

284-315.

Jenkins, B. and A. Faison (1974)AnAnalysis of248 Persons Killed by

New York City Policemen. New York: Metropolitan Applied Research

Center.

Kania, R.R.E. and W. C. Mackey (1977) "Police violence as a function

of community characteristics." Criminology, 15: 27-48.

Kobler, A. L. "Police homicide in a democracy" Journal of Social

Issues, 31: 163-184.

Loether, H. J . and D. G. McTavish. Descriptive Statistics For Sociolo­

gists. Boston: Allyn and Bacon.

Milton, C. H., J . W. Halleck; J. Lardner, and G. L. Abrecht (1977)

Police Use of Deadly Force, Washington, DC: Police Foundation.

Myrdal, G. (1944) An American Dilemma: The Negro Problem and

Modern Democracy. New York: Harper & Row.
New York City Police Department (1976) Annual Report. New York:

Homicide Analysis Unit.

--(1973) Summary of Precinct Populations. New York: Chief of

Field Services.

--(1971-1975) Monthly Arrest Report. New York: Crime Analysis

Unit.

New York State Penal Law (1967) Albany: New York State Govern­

ment.

Robin, G. D. (1963) "Justifiable homicide by police officers." Journal

of Criminal Law, Criminology and Police Science, 225-231.

U.S. Department of Commerce (1973) Characteristics of the Popula­

tion. Washington, DC: Government Printing Office.
Wolfgang, M. (1958) Patterns in Criminal Homicide. Philadelphia:
University of Pennsylvania Press.

194
 CHAPTER 8

A GARRISON STATE IN A "DEMOCRATIC"

SOCIETY

PAUL TAKAGI
This paper reports on a study of police officers killed in the
line of duty and civilians killed by the police . The study was
originated in 1971 in reaction to news reporting on the several
mass media outlets at the local and national levels, which fo- .
cused on FBI statistics indicating police officers were being
"assassinated" at an alarming rate. A police reporter for an
edvcational television station alarmed viewers with a report
that 125 law enforcement officers had been killed in 1971, an
increase of almost two and one-half times over 1963 when only
55 police officers were killed in all of that year. Police killings
of citizens, however, were reported as isolated events . Al­
though the death of civilians at the hands of police occurred
from time to time, no news analyst attempted to show this as a
national phenomenon.
Sorel (1950) said people use words in selective ways to cre­
ate alarm. When a police officer kills a citizen, the official lan­
guage is "deadly force ," suggesting to the audience that the
use of force was legitimate. But when a police officer is killed,
it is characterized as "violence," and therefore, illegitimate. In
this way, news, reporting the killing of police officers in 1971,
conjured the idea that the apparent increase in the killing of
police officers was unprecedented. It was seen as an attack
SOURCE: " A Garrison State in a 'Democratic' Society," Crime and Social
Justice: A JourYULl ofRadical Criminology (Spring-Summer, 1974) 5:27-33.

195

caused in part by the rising political militancy among revolu­

tionary groups, and by the increasing race consciousness
among people of color venting their frustrations by attacking a
visible symbol of authority. This interpretation was entertained
by officials at the highest levels. President Nixon, in April of
1971, called upon police officials; and as subsequent events re­
vealed, other representatives from para-military organizations
also met to deal with the "problem."
The approach by officials was to consider the problem one
of defense, and to search for the best technical means and
policies to protect their view of a "democratic" society. It was
viewed as a military problem, and the fortification of the police
under increased LEAA funding and direction became a national
policy (Goulden, 1970).
One hundred and twenty-five police officers died while on
duty during 1971; the actual rate of death, however, did not
increase because of the greater number of police officers who
were on duty during the same year . Even if the number of
police personnel has increased two and one-half times since
1963, the rate of death among police officers should not change.
This is not said in an attempt to minimize the statistics that
concern the officials. One could argue that the rate of police
deaths should decrease. The point is to look at all the statistics,
including previous studies that actually show the killing of po­
lice officers occurs at a relatively stable rate (Bristow, 1963;
Robin, 1963; and Cardarelli, 1968).
The source of data is the FBI's own reports, which show an
increase in the number of police officers killed, from 55 in 1963
to 125 in 1971, along with an increase of over 50 percent in the
numbers of full-time authorized police personnel. The data pre­
sented in Chart 1 show that the rate of such homicides, while
fluctuating from year to year, does not result in a trend either
up or down over the period. The rate did peak nationally in
1967 with 29.9 deaths per 100,000 law enforcement officers.
This includes all ranks from patrolmen to higher officials and
federal agents. Since patrolmen bear the greatest risk of being
killed in the line of duty, they may feel that FBI reports should
be more detailed to accurately reflect the hazards they face.

196
 Reports to the FBI on the numbers of police officers on
duty and the numbers killed may not give a complete picture,
since the agency has been only gradually achieving uniform
reporting. Indeed, the number of reporting agencies increased
since 1963. California, however, has had fairly complete and
uniform reporting throughout the period, and the death rates
among California police are available for the whole decade since
1960. They, too, show a peak in 1967, a year in which 12 offi­
cers were killed. That did not set a trend, however, as the rate
decreased in the next two years.

Homicides of Police on Duty

50

40
0
0
q
0
0 30
(i)
a..
20

*
a:
10

0~-------,--------r-------~-------r----
1960 1962 1964 1966 '. 1968

CHART 1

For the 86 officers who were killed in California from 1960

through 1970, the police apprehended 117 suspects of whom 55
percent were white, 25 percent Black, and 19 percent Mexican­
American. (This is the same percentage distribution of ethnic!
racial groups in California's prison population.) At the time of
this writing, 65 of the 117 suspects were convicted of either
murder or manslaughter, and 7 cases were still pending in
court. W. H. Hutchins, Assistant Chief of t he California Bureau

197
of Criminal Statistics, noted in a paper delivered to the Califor­
nia Homicide Investigators' Conference on March 5, 1971 that
the great majority of homicidal deaths among police officers
occurred in situations where robberies were in progress or
where robbers were fleeing arrest. But, noted Hutchins, "the
ambushing of officers, which has been relatively rare in the
past, accounted for 25 percent of peace officers killed in 1970"
(Hutchins, 1971).
Mr. Hutchins is not entirely correct when he reports that
the majority of police officers killed were in situations involving
armed robberies. An earlier report by his Bureau of Criminal
Statistics indicates: " ... 63 percent of these officers died while
conducting routine investigations, responding to disturbance
calls and taking people into custody ... " (Beattie, 1968:5). A
special study on the deaths of 39 California police officers (1960
through 1966) shows 35 of the 39 died of gunshot wounds , in
some instances by their own guns (ibid.: 11-14).
Klass, Richard J., 25-year-old patrolman, Daly City Po­
lice Department, killed May 6, 1966. Shot with his own
gun by an escapee with whom he was struggling.
LeFebvre, Richard R., 23-year-old patrolman, Long
Beach Police Department, killed August 15, 1965 at
8:00P.M. Died at the scene of a riot when a shotgun in
the hands of a brother officer discharged during a
struggle.
Ludlow, Donald E., a 26-year-old deputy sheriff, Los
Angeles County, killed August 13, 1965 at 9:00 P.M.
Shot to death when brother officers' gun went off dur­
ing struggle at riot scene.
Ross , Charles M., a 31-year-old patrolman, Richmond
Police Department, killed February 9, 1964 at 1:00
A.M. Shot with his own gun while struggling with two
drunks.
The four cases above were classified as homicides. To distin­
guish accidental death from homicide appears to require consid­
erable judgment among those compiling crime statistics, and it

198
is important to understand that these judgment classifications
are included in the annual FBI reports on homicides of police
officers.
It was noted earlier that the killing of police officers peaked
in 1967 with 29.9 deaths per 100,000 law enforcement officers.
Does this mean that law enforcement work is one of extreme
peril? Robin (1963) argues otherwise:
... there is reason to maintain that the popular con­
ception of the dangerous nature of police work has been
exaggerated. Each occupation has 1ts own hazards. The
main difference between police work and other occupa­
tions is that in the former there is a calculated risk ...
while other occupational hazards are accidental and in­
juries usually self-inflicting (ibid.: 230).
Robin adjusts the death rate among police officers to include
the accidental deaths (mostly from vehicular accidents), and
compares the death rate among the major occupational groups.
It is apparent that the occupational risks in law enforcement
are less dangerous than those in the several major industries.
Mining with 93.6 deaths per 100,000 employees is almost three
times riskier than law enforcement, while construction work is
two and one-half times more dangerous, and agriculture and
transportation show considerably higher rates of death than
does law enforcement. Robin correctly concludes that the data
do not support the general belief that law enforcement work is
a highly dangerous enterprise.
TABLE 1 OCCUPATIONAL FATALITIES PER 100,000 EMPLOYEES 1955
Fatality Rate
Occupation per 100,000

Mining 93.58
Construction Industry 75.81
Agriculture 54.97
Transportation 44.08
Law Enforcement 32.76
Public Utilities 14.98
Finance, Gov., Service 14.18
Manufacturing 12.08
Trade 10.25
Table adapted from Robin (ibid. : Table 6).

199
II.
The other side of the coin is police homicides of citizens.
This aspect of police-citizen interaction has received little atten­
tion except in the work of Robin (ibid.) and Knoohuizen, et al.
(1972). For example, the prestigious President's Task Force
Report on the police (1967) devotes not one line to this issue.
What is generally not known by the public, and either un­
known or certainly not publicized by the police and other offi­
cials, is the alarming increase in the rate of deaths of male
citizens caused by, in the official terminology, "legal interven­
tion of police." These are the cases recorded on the death cer­
tificates as "justifiable homicide" by police intervention. After
disappearing onto computer tapes, these reappear as statistics
in the annually published official volumes of "Vital Statistics in
the United States." Here they can be found under "Cause of
Death, Code Number 984," where they have attracted very
little attention.
The deaths of male civilians ages ten years and over caused
by police intervention gradually increased in rate, especially
from 1962 to 1968, the latest year in which nationwide statis­
tics were available at the time of this writing (see Chart 2).
More dramatic is the trend in civilian deaths caused by Califor­
nia police, where the rate increased two and one-half times
between 1962 and 1969. This increase cannot be attributed sim­
ply to an increase in the proportion of young adults in the
population, among whom a larger share of these deaths occur,
because each annual rate is age-adjusted to the age-profile of
the population in 1960. There is an increase in the rate of homi­
cides by police, regardless of the changes in that age profile.
Why should such a trend go unnoticed? The crime rate has,
of course, increased at the same time, and this, it might be
argued , indicates that more males put themselves in situations
where they risk a police bullet. This is the argument that the
victim alone is responsible. But that is too simple an explana­
tion: an increase in such dangerous situations has not led to an
increased jeopardy of police lives, for, as we have seen, their
homicide rate did not increase over the same period.

200
 Homicides Caused by Police

1.0
rn
Q)
15 .8
~
0
0
q .6
0
0
.... .4
Q)
a.
Q)
~
.2
a:
.0
1958 1960 1962 1964 1966 1968

CHART 2

The charts show police to be victims of homicides at an

annual rate of about 25 per hundred thousand police, while
citizens are victims of killings at the hands of police at a rate of
0.5 per 100,000 males ages ten and over, on the national level,
and a rate of about 0.8 in California. This huge difference of 30
to 50-fold cannot be taken literally, because the civilian rate is
based upon all males over age 9 even though most of them
don't have the slightest chance of confronting a policeman in a
desperate situation of anyone's making. There simply is no
other population base to use in computing that rate. The point,
however, is inescapable: the rate of death did not change for
law enforcement officers during a period when it changed criti­
cally for male citizens.

III.
Black men have been killed by police at a rate some nine to
ten times higher than white men. From that same obscure, but
published source in our nation's capital, come the disheartening
statistics. Between 1960 and 1968, police killed 1,188 Black

201
males and 1,253 white males in a population in which about ten
percent are Black. The rates of homicides due to police inter­
vention increased over the years for both whites and Blacks,
but remained consistently at least nine times higher for Blacks
for the past 18 years (see Chart 3).
That proportionately more Blacks are killed by police will
come as no surprise to most people, certainly to no police offi­
cials. The remarkably big difference should be surprising, how­
ever. After all, the crime rate, even if we rely upon measure­
ment by the arrest rate, is higher for Blacks than for whites.
But that does not explain the killing of Black men. In 1964,
arrests of Black males were 28 percent of total arrests, as
reported by 3,940 agencies the FBI, while Black deaths were 51
percent of the total number killed by police. In 1968 the statis­
tics were essentially the same.
Homicides Caused by Police
(United States)

3.0

2.Q
0
0
0
ci
0
...
Q)
a..
Q)
n; 1.0
a:
White Males

1950 1960
CHART 3.

It might be argued that Blacks have a higher arrest rate for

the seven major crimes: homicide, rape, robbery, aggravated
assault, burglary, theft, and auto theft; and that arrests for

202
these crimes will correlate better with deaths by legal interven­
tion of police. In 1968, Black males accounted for 36 percent of
arrests for the major crimes; four years earlier, in 1964, Black
arrests were less than 30 percent during a year when they
suffered 51 percent of the deaths from police guns. Besides, it
is not certain that the major crimes are a more accurate index
of how frequently Blacks and whites commit crimes. Further,
the threshold of suspicion is lower when a policeman encoun­
ters a Black man, thus the arrest rate is biased against Blacks.
No matter how it is viewed, the death rate of Blacks is far out
of proportion to the situations that might justify it.
Black people don't need these statistics to tell them what
has been happening. The news gets around the neighborhood
when someone is killed by the police. It is part of a history. But
white people, especially policy-makers, don' t live in those neigh­
borhoods, and it is important that they explore the statistics
further.
Take the age groups where "desperate" criminals are much
less likely to be found, the very young and the very old. Male
homicides by police during 1964-1968 were:

TABLE 2

Number of Deaths Rate Per Million/Yearly

White Black White Black

Ages 10-14 5 11 0.12 1.75

Ages 65+ 5 14 0.14 4 .76

In proportion to population, Black youngsters and old men

have been killed by police at a rate 15 to 30 times greater than
that for whites of the same age. It is the actual experiences
behind statistics like these that suggest that police have one
trigger finger for whites and another for Blacks. The latest
statistics, those for 1968, give no reason for altering that belief.
Whereas our analysis covered national data on police kill­
ings of private citizens, Robin (1963:229), utilizing the same

203
data for the years 1950 through 1960, examined the rates of
Black and white victims by selected cities. In absolute numbers,
Chicago police accounted for 54.6 percent of the 350 police
slayings of citizens in the eight cities; the mean actual rate,
however, was highest for Miami, with Chicago second. The two
cities with the lowest police "justifiable homicide" rate, Boston
and Milwaukee, killed Blacks in proportion to whites at a ratio
of 25 to 29 times higher.
TABLE 3 RATES OF BLACK AND WH ITE DECEDENTS, BY CITY

Black White

City per 1, 000,000 Black:White Ratio

Akron 16.1 2.7 5.8 to 1

Chicago 16.1 2.1 7.4 to 1
Kansas City, Mo. 17.0 2 .2 7.5 to 1
Miami 24.4 2 .7 8.8 to 1
Buffalo 7.1 .5 12.2to1
Philadelphia 5.4 .2 21.9 to 1
Boston 3.2 .1 25.2 to 1
Milwaukee 13.5 .4 29.5 to 1

A more detailed analysis of police killing of private citizens

was conducted by Robin for the city of Philadelphia. He re­
ports:
Thirty of the 32 cases (28 were Black victims) were
disposed of by the medical examiner, who at the inquest
exonerated the officers involved in the killings on the
grounds that death was due to justifiable homicide. In
the two remaining cases the officers were held for the
grand jury, indicted, tried by a jury, and found not
guilty (ibid.:226).
Black citizens have long argued that the police are committing
genocide on Black people, and there is increasing evidence that
these killings are indeed murder, and that real justice is rarely
if ever carried out in this process. Knoohuizen, et al. (1972)
conducted a study of Chicago police killing of citizens, and pro­
vided further credence to the claim that police are murdering

204
Black citizens. In their report, Knoohuizen and associates ex­
amined the incidents as reported by the police, the reports of
the coroner's office, and testimony or statements by credible
eyewitnesses. In Table 15 they summarized their findings from
which we extracted three cases.
Case 1. The victim was Linda Anderson. Police
action resulting in her death was ruled justifiable homi­
cide because, according to police reports, she was killed
accidentally during attempt to gain entrance to her
apartment by shooting the lock off the door. The part­
ner of the officer, and independent witnesses, corrobo­
rated the police officer's version. An independent inves­
tigation revealed that the officer used a shotgun
standing four feet from the door, did not warn the occu­
pant of 1mpending shot, and missed the lock completely.
Case 2. The victim was Raymond Jones. Police action
was ruled excusable because police officers did not
strike the deceased and were only using the amount of
force necessary to bring the suspect under arrest.
Seven of 9 officers involved in the incident testified and
confirmed each other's story. The report of the Coro­
ner's pathologist, however, revealed that Mr. Jones was
age 31 and in good health. He was also unarmed. The
use of excessive force was implied when 9 police offi­
cers cannot subdue a suspect without causing his death.
Case 3. The victim was Charles Cox. The police report
did not offer a justification or an excuse claiming the
victim died from drug overdose rather than use of po­
lice force. Further reports from the police indicate
blood analysis revealed some drugs in the victim's body.
One of the arresting officers and one of the officers m
charge of the lock-up both testified that the victim ap­
peared all right when in their charge. A pathologist
testified on the basis of his examination of the body
that Cox died of blows to the head .
Knoohuizen and associates conclude from their analysis that in
28 of the 76 cases in which civilians were killed at the hands of
the Chicago police, there was substantial evidence of police mis­
conduct; and in 10 of the 76 cases, there was substantial evi­
dence of criminal liability for manslaughter or murder (ibid.:
61).

205
 Despite grand jury findings in those instances where police
officers are held criminally liable, the courts have been reluc­
tant to proceed with prosecution. All too often, such matters
are thrown out of court or juries return the verdict of not
guilty. For example, Superior Court Judge Ross G. Tharp of
San Diego County dismissed involuntary manslaughter charges
against a California Highway patrolman indicted in the fatal
shooting of an unarmed 16-year-old boy. According to police
reports, Roland R. Thomas was shot by Officer Nelander fol­
lowing a high speed chase in an allegedly stolen car. The car
ran off the road, and Thomas appeared to reach toward his
pocket at which point the officer fired his gun. In dismissing
the case, Judge Tharp observed: "I think the officer deserves a
commendation for doing his duty rather than standing trial.''
The only recent cases in which police officers were held
accountable for killing civilians were shown on a recent TV
program (Owen Marshall , ABC, Saturday, March 2, 1974), in
addition to the highly publicized case in Texas were a 12-year­
old Mexican-American youngster was shot while under custody
in a police car. The circumstances in the latter case were so
gross that a dismissal was out of the question. The court, how­
ever, sentenced the officer to a prison term of 5 years in a state
where sentences of 1,000 years for lesser crimes are not uncom­
mon.

IV.
Authorities have been trying to combat what they view to
be a rash of attacks on police, to the neglect of all the data that
bear on the problem-a problem in which other lives are in­
volved. The problem has existed all along; at least since 1950,
and there is reason to believe for decades before that, Black
people have been killed by the police at a tragically dispropor­
tionate rate, beyond the bounds of anything that would justify
it.
Open warfare between the police and the citizenry might be
one of the outcomes. Two recent attacks upon police station

206
houses, one by a bomb and the other by shotgun wielding assail­
ants resulting in the death of two police officers, are indicative.
In the latter killing, the gunman thrust a shotgun through the
speaking hole of a bullet proof glass shield separating the desk
sergeant from the public. Portions of the police station house
were protected by cyclone fencing. The wall of isolation sur­
rounding the police is not only social and psychological, but
physical, and the breaking down of these walls was considered
by the National Crime Commission to be the single most impor­
tant priority. Yet the federal government in appropriating bil­
lions of dollars for the Law Enforcement Assistance Adminis­
tration program earmarked the funds primarily for the
fortification of the police, thereby contributing to their isola­
tion.
Currently, the concept of citizen participation is being
stressed by the LEAA. The support the police get from some
citizens' groups actually increases the isolation of police from
minority communities. In Oakland, California, such a group,
called Citizens for Law and Order, has a program of needling
judges for their "soft" handling of criminal cases, firing broad­
sides at the press, television and radio, and appearing before
local governmental bodies to promote support for the police and
more "discipline" in schools. Programs like these are based on
the belief that increasing the penalty for crime, increasing the
powers of the police, and invoking police coercion of the citi­
zenry will result in law and order.
Other citizens' groups have encouraged the introduction of
reforms. People have worked on a variety of schemes such as
Civilian Review Boards, psychological testing and screening of
police candidates, human relations training, police community
relations, racially integrated patrol units, and efforts to in­
crease the hiring of Black and other minority officers. To the
extent that they work to improve only the "image" of police,
they fail because the problems go much deeper. And to a major
extent they fail because policemen, most of them willingly and
others unknowingly, are used as the front line to maintain the
social injustices inherent in other institutions and branches of
government.

207
 Perhaps the only immediate solution at this time is to dis­
arm the police. Observers have noted that provinces in
Australia where the police are unarmed have a much lower rate
of attacks upon the police, compared to neighboring provinces
where the police are armed, and the corollary observation, a
lower rate of police misconduct.
Disarming the police in the United States will undoubtedly
lower the rate of police killings of civilians; it does not, how­
ever, get at the causes of police misconduct, particularly to­
ward black people . The findings that Blacks are killed by the
police at a disproportionate ratio in cities like Milwaukee and
Boston, and the attitudes of officials like San Diego County's
Superior Court Judge Tharp, require a more fundamental un­
derstanding of the meaning of policing in contemporary Amer­
ica.

v.
In distinguishing social justice from distributive justice, the
former would not have been obtained, if, for example, Officer
Lelander had been tried and convicted for the killing of a 16­
year-old alleged auto thief; that would have been distributive
justice, because it would have symbolized the fact that the po­
lice would not have received special treatment from the courts.
Instead , the question that must be asked is why the police offi­
cer resorted to deadly force involving an alleged theft. To put it
differently, why was the value of an automobile placed above
the value of a human life? Judge Tharp's comments in dismiss­
ing the case provide a partial answer: "For doing his duty," the
duty being to enforce the laws having to do with the property
rights of an automobile owner. The critical issue here is that
the auto theft laws and for that matter most of the laws in
American society essentially legitimize a productive system
where human labor is systematically expropriated. Examine for
a moment the social significance of an automobile; it involves
an array of corporate systems that expropriate the labor of
people that go into manufacturing its parts, the labor for its

208
 assembly, the labor involved in extricating and processing the
fuel that propels it, the labor of constructing the roads on which
it runs, etc. The fiction of ownership exacts further capital by
banking institutions that mortgage the commodity, and auto­
mobile insurance required by laws that extorts additional capi­
tal. The built-in obsolescence, or more precisely, the deprecia­
tion of the commodity, occurs when the muscle, the sweat, and
human potential have been completely capitalized. These are
the elements embodied in an automobile. It is no longer merely
a commodity value, but represents a social value.
The automobile is a commodity created by varied types of
wage labor. And as noted by men with ideas as far apart as
those of Adam Smith and Karl Marx, the wealth of nations
originate in the efforts of labor. But Marx added that wealth
based on production of these commodities is accrued through
the expropriation of labor power; and thus, the concept of pri­
vate property based on this form of wealth is in essence the
theft of the value-creating power of labor. The criminal laws,
the system of coercion and punishment, exist to promote and to
protect the consequences of a system based on this form of
property.
The rights of liberty, equality, and security are not ele­
ments to be exchanged for the right of property acquired by the
exploitation of wage labor; nor should they be expressed in
relative terms, that is, greater or less than property rights. One
person's life and liberty is the same as the next person's. But in
a society that equates private property with human rights, they
become inevitably reduced to standards and consequences that
value some lives less than others. The system of coercion and
punishment is intimately connected with the inequitable distri­
bution of wealth, and provides the legitimation under the per­
verted notion that "ours is a government of laws" even to kill
in order to maintain social priorities based on private property.
This is the meaning of policing in American society.
Why are Black people killed by the police at a rate nine to
ten times higher than whites? We can describe the manifesta­
tions of racism but cannot adequately explain it. At one level,

209
 we agree with the observation that the existence of racism is
highly profitable. The Black urban ghettos, created by Ameri­
ca's industries, provided the cheap labor power for the accumu­
lation of some of America's greatest industrial wealth at the
turn of the 20th century, and again during World War II. These
urban ghettos still provide a highly exploited source of labor. In
addition, the ghettos themselves have become a place for ex­
ploitation by slum landlords, merchants selling inferior quality
goods at higher prices, a justification for higher premium rates
on insurance, and the victimizing of people under the credit
purchase system. To maintain this situation, the regulatory
agencies, including the police, have ignored the codes govern­
ing housing, food, health, and usury conditions.
In cities across the country, the infamous ghettos are now
deemed to be prime real estate, and the state under the powers
of eminent domain claim for finance capitalism the areas for
high rise buildings, condominiums, trade complexes, and enter­
tainment centers ostensibly for the "people." Under what has
been called urban redevelopment, the police are present to
quiet individual and especially organized protest and dissent,
and the full powers of the state are employed to evict, to dis­
possess, and to humiliate.
At another level, the concentration of capital has produced
on the one hand, a demand for a disciplined labor force and, in
order to rationalize its control, to rely increasingly upon admin­
istrative laws; on the other hand, it has created a surplus labor
force that is increasingly controlled by our criminal laws. The
use of punishment to control surplus labor is not new, having
its roots in early 16th century Europe (Rusche and Kirch­
heimer, 1968).
Historically, people of color came to the United States not
as freepersons, but as slaves, indentured servants, and as con­
tract laborers. They were initially welcomed under these condi­
tions. As these particular systems of exploitation gradually dis­
appeared and the people entered the competitive labor market
a variety of devices were employed to continue oppressing
them, including imprisonment. In the present period described

210
 by some as the post-industrial age, increasing numbers of peo­
ple, and especially Black people, find themselves in the ranks of
the unemployed, which establishment economists, fixing upon
the 5 percent unemployment figure, dismiss as a regular fea­
ture of our political economy. Sweezy (1971) disagrees, arguing
that the "post-industrial" unemployment figures are the same
as that in the Great Depression when one includes defense and
defense related employment data. When arrest and prison com­
mitment data on Black people are viewed from this perspective,
especially the sudden increase in prison commitments from a
stable rate of ten percent up to and during the early period of
World War II to almost double that after the war, there is some
basis to suspect that the police killing of Black citizens is pun­
ishment to control a surplus labor population.
The labor surplus analysis, however, does not explain the
sudden increase in police killing of civilians beginning around
1962. Did the Civil Rights movement in housing, education, and
employment, and more specifically, the militancy of a Malcolm
X, and the liberation movements in Third World nations around
the world, re-define the role of the police? Did finance imperial­
ism in the form of multi-national corporations beginning about
this time create an un-noticed social dislocation? Why do the
police kill civilians at a much higher rate in some cities com­
pared to others, and why do they kill Blacks at a disproportion­
ately higher ratio in cities like Boston and Milwaukee? Why do
California police, presumed to be highly professional, kill civil­
ians at a rate 60 percent higher than the nation as a whole? We
are not able to answer these questions.
We must, however, pause for a moment, and consider what
is happening to us. We know that authorized police personnel in
states like California has been increasing at the rate of 5 to 6
percent compared to an annual population increase of less than
two and one-half percent. In 1960 there were 22,783 police
officers; in 1972 there were 51,909. If the rate of increase con­
tinues, California will have at the turn of the 21st century an
estimated 180,000 police officers, an equivalent of 10 military
divisions. Is it not true that the growth in the instruments of

211
coercion and punishment is the inevitable consequence of the
wealth of a nation that is based upon theft?
America is moving more and more rapidly towards a garri­
son state, and soon we will not find solace by repeating to
ourselves: "Ours is a democratic society."

Notes
The ideas in this section are not original. They come from Fourier,
Godwin, Proudhon, Marx, Kropotkin, and others.

References
Beattie, Ronald H.
1968 "California Peace Officers Killed 1960-66." Bureau of Crimi­
nal Statistics, Department of Justice, State of California (September).
Bristow, Allen P.
1963 "Police Officer Shootings: A Tactical Evaluation," The Jour­
nal of Criminal Law, Criminology, and Police Science 54.
Cardarelli, Albert P.
1968 "An Analysis of Police Killed by Criminal Action: 1961­
1963." The Journal of Criminal Law, Criminology and Police Science
59.
Goulden, Joseph
1970 "The Cops Hit the Jackpot." The Nation (November).
Hutchins, W. H.
1971 "Criminal Homicides of California Peace Officers, 1960­
1970." A report delivered before the California Homicide Investiga­

tors' Conference, Los Angeles, March 5.

Knoohuizen, Ralph, Richard P. Fahey, and Deborah J. Palmer

1972 The Police and Their Use of Fatal Force in Chicago. Chicago
Law Enforcement Study Group.
President's Commission on Law Enforcement and Administration of
Justice
1967 Task Force Report: The Police. Washington, D.C.: U.S. Gov­
ernment Printing Office.

212
Robin, Gerald D.
1963 "Justifiable Homicide by Police Officers." The Journal of
Criminal Law, Criminology, and Police Science 54.
Rusche, Georg and Otto Kirchheimer
1968 Punishment and Social Structure. New York: Russell and
Russe ll.
Sorel, G.
1950 Reflections on Violence, Glencoe, Illinois: Free Press.
Sweezy, Paul M., Harry Magdoff and Leo Huberman
1971 "Economic Stagnation an d Stagnation of Economics."
Monthly Review 22 (April).

213
 CHAPTER 9

USE OF DEADLY FORCE TO ARREST A

FLEEING FELON-A CONSTITUTIONAL

CHALLENGE, PART II

J . PAUL BOUTWELL

RESTRICTIONS UPON USE OF DEADLY FORCE

THROUGH DEPARTMENT POLICY
A most significant effort toward reform of the common law
rule has come through law enforcement administrators.
Whether in response to persuasive police commentary, 1 na­
tional study commissions2 or because of tragic incidents in the
community, 3 many executives of law enforcement agencies have
prepared written policy detailing restrictions on the use of
deadly force for purpose of making an arrest. In many in­
stances, the policy is more restrictive than the state statutory
standard. This is understandable. The fact that deadly force is
legally justified does not mean that it is always wisely utilized .
Riots, for example, have been attributed to an officer's legal,
but unwise, use of deadly force. 4 The legislature determines the
legal use of deadly force; the administrator promotes its wise
use.

SOURCE: "Use of Deadly Force to Arrest a Fleeing Felon-A Consti·

tutional Challenge, P art II" FBI Law E nforcement Bulletin (October
1977) 46:10:27-31.

214
 Many law enforcement administrators are concerned that if
an officer is sued, the department's firearms regulation will be
admitted into evidence, and where more restrictive than state
law, will create liability where none might otherwise exist. This
is not necessarily the case. To begin with, states differ on ad­
missibility of departmental policy. Decisions in California and
Florida illustrate the different responses. For example, in a
California case, a police officer shot at and killed a fleeing felon.
The shooting was a justifiable use of deadly force under state
law. The police tactical manual pertaining to the use of fire­
arms, however, justified the use of deadly force only if neces­
sary to save the officer, a citizen, a brother officer, or a pris­
oner from death or grave bodily harm. The Supreme Court of
California held the manual was admissible on the ground that
an employee's failure to follow a safety rule promulgated by his
employer, regardless of its substance, served as evidence of
negligence. 5
On the other hand, in the State of Florida, at least two
district courts of appeal have reached an opposite result. In one
case, officers covering a rock concert observed from a rooftop
two teenagers trying the doors of a number of vehicles in the
parking lot and finally entered a van. The rooftop officers di­
rected officers on the ground to arrest them. As an officer
attempted to arrest one of the boys, a struggle ensued and the
officer fell to the ground after receiving a blow to the face. The
youth ran, and the officer shot the plaintiff in the leg. Florida
has codified the common law rule. Over the officer's objection
in a civil suit, the court admitted into evidence a departmental
o.'der on the use of firearms, which was in effect at the time of
the shooting. The order authorized the officers to use firearms
to apprehend a fleeing felon, but only when the officer reasona­
bly believes the fleeing person has committed either (1) a vio­
lent crime to the person of another, or (2) a crime against
property that clearly demonstrates a wanton and reckless disre­
gard for human life. On appeal, the officer contended that the
trial court erred in admitting this order. The appeals court
agreed. While the departmental regulation may be applicable

215
for departmental discipline of its own members, the regulation
would not affect the standard by which the officer's criminal or
civil liability was measured. To admit the public safety order
constituted reversible error. 6
Whether departmental regulations will create liability
where none might otherwise exist is more difficult. Americans
for Effective Law Enforcement (AELE)1 makes the following
points: (1) Police chiefs and other administrators should not be
dissuaded from promulgating safety rules and policy directives
due to the threat of civil liability; (2) it is inconsistent with
modern management to leave unfettered discretion (as to when
an officer may use his firearm) to the lowest ranks-this is not
to suggest that any particular restrictive policy is meritorious,
only that planning and policymaking should be centralized at
the highest administrative levels; and (3) written directives
which restrict a police officer's action beyond the requirements
of state law should contain an explanation of their intended
purpose. Suggested wording is as follows:
''This directive is for internal use only and does not
enlarge an officer's civil or criminal liability in any way.
It should not be construed as the creation of a higher
standard of safety or care in an evidentiary sense, with
respect to third party claims. Violations of this direc­
tive, if proven, can only form the basis of a complaint
by this department, and then only in a nonjudicial ad­
ministrative setting. " 8
The wise administrator, concerned about potential liability
problems with regard to the use of deadly force , will discuss
this topic with a legal adviser. He certainly wants to know what
effect his policy might have on his officers' potential liability.
He needs to be clear as to who will pay the civil judgment, if
one is awarded, arising out of a deadly force case. 9

THE INTERPLAY BETWEEN A STATE'S JUSTIFIABLE

HOMICIDE STATUTE AND CIVIL LIABILITY
A state legislature defines what constitutes justification for
an act otherwise criminal. 10 A state civil court defines what

216
 constitutes privilege for conduct otherwise tortious. 11 Query:
Can a state civil court adopt a definition of an officer's privilege
in the use of deadly force, that is more restrictive than the
state's legislative standard, expressed through its justifiable
homicide statute?
The question underscores the distinction between the two
areas of the law-criminal and civil. The legislature of the state
has the legitimate authority to define crimes and defenses, and
generally the civil courts retain the common law authority to
define torts and their defenses. So the simply answer to the
question is yes; civil courts may adopt a definition of privileged
conduct that is more restrictive than the state's justifiable
homicide statute. It should be emphasized, however, that most
courts have refused to do so.
A recent Minnesota case illustrates the point. Early one
morning, an off-duty officer, dressed in civilian clothes but who
carried his .38-caliber snub-nose revolver, drove a marked po­
lice department "take-home" squad car, which he was autho­
rized to use, to pick up the morning newspaper. On his return,
he observed a station wagon traveling at an excessive rate of
speed collide with a parked car. Two boys got out, yelled some­
thing into the station wagon, and then ran. As the officer
stopped his squad car, another person alighted from the driv­
er's side of the wagon and ran. The officer jumped out of the
squad car and shouted "Stop, police." As he chased one boy, he
repeatedly shouted similar warnings, finally calling out, "Stop,
or I'll shoot." The plaintiff ignored the warnings and continued
to run. The officer fired a warning shot into the ground, but the
plaintiff only ran faster. The officer again yelled, "Stop, or I'll
shoot." When this warning failed to produce results, the officer
aimed and fired a shot, intending to hit the plaintiff in the
lower part of his body. Instead of striking the plaintiff in the
legs, the bullet struck the plaintiff in the nape of the neck,
permanently crippling him.
In his complaint, the plaintiff alleged defendant's liability
on two theories-battery and negligence. The trial court sub­
mitted the case to the jury on the theory of negligence alone.

217
The jury found for the officer. They found also that the plain­
tiff's negligence was the proximate cause of his own injury. The
plaintiff appealed. He argued that it was error for the trial
court to leave out the issue of battery. In addition, the plaintiff
sought to have the Supreme Court of Minnesota adopt a civil
liability standard for privileged conduct, a standard that would
be more restrictive than the state's justifiable homicide statute.
Minnesota's justifiable homicide statute follows the common
law rule.
The Supreme Court of Minnesota held that the trial court
had improperly framed the issue in the case in terms of negli­
gence rather than battery and remanded the case for a new
trial. The court wrote that while they were not technically
bound to follow the statutory formulation of the justifiable
homicide statutes, they would nevertheless do so and defer to
the legislative policy in defining tort liability. The police officer
contemplating the use of force under emergency conditions
should not be held to conflicting standards of conduct by the
civil and criminal law. The confusion which would be engen­
dered by such a situation can only produce unfair and inequita­
ble results. The Court wrote:
"It is in the legislative forum that the deterrent effect
of the traditional rule may be evaluated and the law­
enforcement policies of this state may be fully debated
and determined.... The legislature, and not this court,
is the proper decision maker." 12
In order for a police officer to raise an affirmative defense
of privileged use of his firearm in a suit alleging battery, the
officer must bear the burden of proving: (1) That he had proba­
ble cause to believe that the person sought to be arrested either
committed or was committing a felony, and (2) that he reasona­
bly believed the arrest could not be effected without the use of
a firearm .
CONSTITUTIONAL ANALYSIS OF THE USE OF DEADLY
FORCE TO ARREST A FLEEING FELON
The most significant development in litigation regarding
the common law fleeing felon rule is the federal constitutional

218
challenge made upon the use of deadly force to arrest a nonvio­
lent, fleeing felon. Such a challenge may be made by a plaintiff
seeking either declaratory or injunctive reliefY Most fre­
quently, however, the plaintiff merely files a claim under title
42, United States Code, section 1983, 14 alleging the violation of
a constitutional right. This legislation was enacted April 20,
1871, with the purpose of providing a remedy for the wrongs
allegedly being perpetrated under color of state law. Thus,
1983, as it is often called, creates a right to sue law enforce­
ment officers personally for depriving another of " . . . any
right, privileges, or immunities secured by the Constitution and
laws ...."(of the United States). Such suits may be filed in the
U.S. district courts under the provisions of title 28, United
States Code, section 1343.
Prior to 1961, it was thought the plaintiff had to exhaust
possibilities that local or state remedies would give relief before
coming to the federal court. In a 1961 landmark decision, the
U.S. Supreme Court established the principle that the right to
sue police officers under 1983 was completely independent of
any state remedies that might be available. The Court stated,
"It is no answer that the State has a law which if enforced
would give relief. The federal remedy is supplementary to the
state remedy, and the latter need not be first sought and
refused before the federal one is invoked." An officer could no
longer regard abstention or exhaustion of local remedies as
useful in defending an action under 1983. 15
Thus, a plaintiff may commence a section 1983 action
against an officer in federal court, or he may file a civil suit in
state court. It is sometimes asked how a state civil lawsuit
brought in a state court and arising out of the same set of facts
differs from a 1983 suit. Some general observations on the
nature of a state law suit are useful before discussing some of
the recent 1983 cases.

State Tort Action Distinguished

State civil lawsuits arising out of an officer's use of his
firearm are not unusual. A suit may develop from its negligent

219
use as well as from its intentional use. In the latter case, the
distinction between justifiable force and excessive force is im­
portant.

Negligence
Probably the most widely recognized duty of a law enforce­
ment officer is that of requiring him to avoid negligence in his
work. Our society imposes a duty upon each individual to con­
duct his affairs in a manner which will avoid subjecting others
to an unreasonable risk of harm. This, of course, also applies to
law enforcement officers. If his conduct creates a danger recog­
nizable as such by a reasonable officer in like circumstances, he
will be held accountable to others injured as a proximate result
of his conduct and who have not contributed to their own harm.
These general principles are well-known concepts in the law of
negligence.
They mean that actions taken by officers in apprehending
criminals must not create an unreasonable risk of injury or
death to innocent persons. The creation of risk is not in and of
itself negligence; however, the law does require a reasonable
assessment of harm's likelihood and regards as negligent any
act which creates a risk of such magnitude as to outweigh the
utility of the act itself.
Under the civil court system, if the police officer owed no
duty to the complainant, he will not be penalized even if the
plaintiff in fact suffered some injury. An officer will be liable
only where it is shown that (1) he was obliged to do or refrain
from doing something, and (2) the plaintiff was injured because
of the officer's failure to comply with this obligation or duty.
Assume that Officer A shoots at B, a felon fleeing in a
congested downtown area, but misses B and hits C, an innocent
bystander. C, in a civil suit against Officer A in State court, will
allege that Officer A was negligent in the discharge of his fire­
arm. The gist of C's suit is that Officer A has breached his duty
to C.

220
 Intentional Torts
Another category of torts is termed intentional torts. In a
negligence suit, the officer will not be liable unless he foresaw,
or should have anticipated, that his acts or omissions would
result in injury to another. An intentional tort is the voluntary
doing of an act which to a substantial certainty will injure an­
other. It does not have to be performed negligently to be ac­
tionable. Examples of such torts are false arrest and assault
and battery. Assume Officer A intentionally shoots and seri­
ously injures B, a fleeing felon. B may bring a civil suit in state
court alleging that he has been battered, an intentional tort.
The gist of B's action is that Officer A used excessive force in
his effort to apprehend him and the use of his firearm was not
justified under the circumstances. It is not alleged that Officer
A was negligent-he did what he intended to do-namely,
shoot B. The essential elements of the tort of battery are intent
and contact. Privilege, however, is an affirmative defense to
the tort of battery. Usually the officer must bear the burden of
proving the essential elements of the defense. A few jurisdic­
tions reach a contrary result, adopting the rule that a police
officer's act is presumed lawful. 16 In final analysis, the
reasonableness of the force used in making an arrest under all
the circumstances is a question of fact for the jury or other
trier of fact (such as a judge in a bench trial), and the standard
usually expressed is the conduct of ordinary prudent men under
existing circumstances. Not a very precise standard to be sure.

Notes
1. Chapman, "Police Policy on the Use of Firearms." Police Chief,
July 1967, at 16, 26-27. McCreedy & Hague, "Administrative and
Legal Aspects of a Policy to Limit the Use of Firearms by Police
Officers," 42 Police- Chief, January 1975, at 48.
2. President's Comm'n. on Law Enforcement and Administration of
Justice . Report: The Challenge of Crime in a Free Society 119 {1967);
Task Force Report: The Police 189-90 {1967); 1 Nat'!. Comm'n. on
Reform of Fed. Crim. Laws, Working Papers 269 (1970).

221
3. Bart, "Inquest Lightens Tension in Watts," New York Times, May
21, 1966, p. 13, col. 1.
4. The San Francisco riot of 1966 was said to have started after a
juvenile was shot and killed while fleeing from a stolen car. Davis,
"Calm is Restored in San Francisco," New York Times, Sept. 30,
1966, p. 1, col. 5.
5. Grudt v. City of Los Angeles, 468 P. 2d 825 (Cal. 1970).
6. City of St. Petersburg v. Reed, 330 So. 2d 25 (Fla. App. 1976). See
also, Chastin v. Civil Service Board of Orlando, 327 So. 2d 230 (Fla.
App. 1976).
7. Americans for Effective Law Enforcement, Inc. (AELE) is a na­
tional, not for profit organization whose purpose is to provide a voice
for the law-abiding citizens through responsible support for profes­
sional law enforcement. As a citizen-supported research and action
organization employing three attorneys and three legal assistants, all
of whom have law enforcement backgrounds, AELE also publishes the
Legal Liability Reporter, and the staff has sponsored workshops
across the country on civil liability.
8. AELE Legal Defense Manual, "Admissibility of Police Written Di­
rectives in Litigation," Brief No. 76-5, p. 14 (October 1976).
9. A.B.A. Standards for Criminal Justice, the Urban Police Function
(approved draft, 1973) §5.5, provides: "In order to strengthen the
effectiveness of the tort remedy for improper police activities, munici­
pal tort immunity, where it still exists, should be repealed and munici­
palities should be fully liable for the actions of police who are acting
within the scope of their employment as municipal employees."
10. Justification is based on a determination that an act is legal be­
cause circumstances negate the validity of the normal rules of criminal
liability. Such defenses recognize that under such circumstances the
value protected by law is eclipsed by a superseding value. Note, Statu­
tory Reform, 75 Colum. L. Rev. 914 (1975).
11. Privilege in the law of torts is a defense to what might have been
an actionable wrong. It excuses such conduct, hence no liability occurs.
Comment. 11 Harv. Civ. Rights-Civ. Lib. L. Rev. 361.
12. Schumann v. McGinn, supra note 1, at 537. The dissenting opin­
ion of Justice Rogosheske is instructive. He pointed out that the crimi­
nal statute distinguishes between the killing of felony and misde­
meanor suspects, whereas sound policy dictates that tort law should
distinguish between the killing of dangerous and nondangerous crimi­
nal suspects: "Surely a police officer should not be imprisoned if he
mistakes a nondangerous for a dangerous felony suspect and uses his

222
firearm against the former. However, unless he is in violation of spe­
cific instructi ons (emphasis added) his employer ought to bear finan­
cial responsibility for mistakes committed in the line of duty. Viewed
in this way, it does not follow, as the majority declares, that under the
rule urged a police officer contemplating the use of force under emer­
gency conditions would be held t o conflicting standards of conduct by
the civil and criminal law. A police officer who makes a mistake and
uses deadly force against a nondangerous felon would know unequivo­
cally that he is committing a civil wrong. The legislature and the
courts of this state, out of awareness of his difficult job in these emer­
gency circumstances, will not jail him for his mistake, but in no way
can that justify granting immunity for a civil wrong .... Rather, and
hopefully, it would lead all police officers in Minnesota to do what
some, if not most, well-trained and experienced police officers already
practice, which is to follow the rule that the use of deadly force is not a
proper arrest procedure for nondangerous, nonthreatening felons."
13. Generally, the way to challenge the constitutionality of a state
statute is to seek injunctive relief under 28 U.S.C. §2281. Upon proper
application, a three-judge court will be convened to hear and deter­
mine the constitutionality of the challenged statue. See, Cunningham
v. Ellington, 323 F . Supp. 1072 (W.D. Tenn . 1971).
14. 42 U.S.C. §1963 reads as follows: "Every person who, under color
of any statute, ordinance, regulation, custom or usage, of any State or
Territory, subjects or causes to be subjected, any citizen of the United
States or other person within the jurisdiction thereof to the depriva­
tion of any rights, privileges, or immunities secured by the Constitu­
tion and laws, shall be liable to the party injured in an action at law,
suit in equity, or other proper proceeding for redress."
15. Monroe v. Pape, 365 U.S. 167 (1961).
16. West v. Nantz, 101 S.W. 2d 673 (Ky. 1937); Wall v. Zeeb, 153 N.W.
2d 779 (N.D. 1967); Modesett v. Emmons, 292 S.W. 855 (Tex. Com.
App. 1927).

223
 CHAPTER 10
POLICE POLICY ON THE USE OF FIREARMS

SAMUEL G. CHAPMAN
The traditional decentralization of governmental organiza­
tion and authority in the United States has led to the establish­
ment of a vast array of police agencies. As a result, there exists
a proliferation of police departments across the nation which
has developed autonomously with little conscious design for any
uniformity in policies or rules of procedure. Such lack of stan­
dard procedures is especially apparent when one surveys regu­
lations purporting to govern the use of firearms in various po­
lice agencies throughout the United States. 1
Whether firearms use policies are written or "oral," the
broad variation in comprehensiveness and utility reflects more
than the traditional American spirit of individualism and non­
conformity. It reflects, in far too many instances, a failure on
the part of police administrators to provide adequate guidance
for officers faced with situations where they must decide in­
stantaneously whether or not to use their firearms in discharg­
ing official responsibilities. This state of affairs is particularly
disturbing because it exists at a time when police agencies seek
to gain greater public cooperation and respect while organized
interest groups of varying political persuasions seem poised to
exploit police misjudgment on a scale previously unknown in
American law enforcement annals.

SOURCE: Reproduced from The Police Chief (July 1967 issue) with the per·
mission of the International Association of Chiefs of Police.

224
 In a democratic society, the manner in which the police
officer uses his weapon may be critical to the effectivensss of
his organization, for accomplishment of the total police mission
is dependent, at least in part, upon the cooperation of a major­
ity of the citizenry. Imprudent or indiscreet use of firearms will
arouse public indigation and alienate public support of the po­
lice agency.

DISARM THE POLICE?

If, in fact, police use or misuse of firearms in our society
constitutes a potential threat to the public support necessary to
the accomplishment of the total law enforcement mission, an
obvious solution to the problem would be to disarm the Ameri­
can police. Like all simple solutions, however, such a course of
action warrants closer examination. But from time to time pro­
posals are offered which call for disarming American police
officers. An example of one such proposal, made by a university
professor, says in part:
The use of firearms, restricted in England to only
those who absolutely need them in the course of their
duties, should be equally restricted in American juris­
dictions. By not carrying a gun, the traffic officer, for
example, who has more extensive contact with the pub­
lic than any other policeman, would instill greater confi­
dence in the public and remove a psychological barrier
between himself and the public which firearms inevita­
bly produce. Most officers never handle dangerous per­
sons, and therefore, do not need a gun. In fact, It is
likely that the carrying of weapons imparts a false
sense of pride in pohce officers and gives them a poor
excuse for continuing a lack of education and trainmg. 2
Three aspects of the professor's statement require re­
sponse. First, the British police routinely carry no sidearms for
reasons other than tradition. They go unarmed because it is
very difficult for a British citizen to purchase personal arms, as
the professor notes. The police also go unarmed on the premise
that being armed would offer inducement for the criminal to
arm himself. And they go also unarmed because the need t o

225
 bear arms for self-defense seems extremely limited in Britain.
That the British police have little need to bear firearms
routinely for self-defense becomes evident when one learns that
from January 1, 1946 through September 30, 1966 (almost 21
years) only 19 sworn members have been fatally injured "as a
result of having been attacked in a manner intrinsically likely to
cause death."3 Of these victims, 13 succumbed to gunfire, three
to stabbing, and three were fatally injured in other fashions. 4
These 19 victims include the three unarmed Metropolitan Police
detectives murdered in London on August 12, 1966 in what is
called "the worse crime against the [British] police in 56
years."5
By way of contrast to the British experience the FBI re­
ports that from January 1, 1946 through September 30, 1966,
1014 American police officers have been murdered in the line of
duty. 6
The British populace in a densely populated, realtively
small area is considered more homogeneous and stable in terms
of mobility than persons in the mobile , vast American nation.
With regard to owning firearms, if a British resident has one,
he must carefully control his weapon. Also he does not carry his
firearm except for good reason and he must renew his firearm
certificate each year. There is no comparable control in the
. United States. A Home Office official notes:
. . . possession is restricted to persons who have ob­
tained a certificate from a Chief Constable of Police.
Applications for certificates are subject to the most
careful consideration and experience leads me to sug­
gest that the number of new certificates issued is rela­
tively low. Certificates have to be renewed annually and
here, again, the case for continuance is carefully scruti­
nized. We do not keep details of the number of new
certificates issued each year, but .. . in February 1965,
the number of certificates in force in England and
Wales totalled 220,496 . . . Most of the certificates is­
sued refer to .22 rifles used for target shooting or ver­
min destruction . Very few certificates are granted for
pistols or revolvers. Where the problem lies is in the
number of illegal sales and possession of firearms, . . .
we have no rehable statistics about those. 7

226

 Second, the notion that certain classes of American police

personnel need no arms is misleading. For example, during the
six-year period between 1960 and 1965 , 84 of the 278 American
police killed in line of duty were fatally injured while making
arrests, including those for traffic violations, or while trans­
porting prisoners. Twenty were killed while off duty but inter­
ceding to prevent the commission of crime, and 17 men were
killed without warning, the result of a sudden unprovoked at­
tack. Furthermore, almost all uniformed officers assigned to
patrol the streets or highways are charged with enforcing traf­
fic regulations consistent with their other multi-form activities. 8
It is almost impossible to separate those whose primary mission
is to enforce traffic regulations from those whose duties encom­
pass the full range of uniform patrol activities. Finally, even
those personnel who serve in traffic- or highway patrol-oriented
assignments face dangerous persons regularly and effect nu­
merous arrests. For example, in 1965, California Highway Pa­
trol field officers captured three of the suspects listed as among
the "Ten Most Wanted" by the FBI. This indicates the great
danger to which these men are regularly exposed in the course
of issuing traffic citations or warnings. 9
An additional consideration is that on many occasions lives
and property are saved because of the judicious use of firearms
by law enforcement officers. In other instances the knowledge
that officers are armed and trained in the use of lethal weapons
has proven sufficient to induce dangerous criminals to surren­
der or refrain from violence.
Finally, regardless of how often a policeman must resort t o
his firearm and the degree of hazard which one associates with
police service, it is evading the question to conclude that fire­
arms are unimportant or unnecessary solely on the basis of the
infrequency of combat situations. Since frequency of combat
circumstances cannot be anticipated, it is essential that the offi­
cer who has occasion to use his weapon no more than once in 20
years or, happily, never encounters police combat, should be
fully informed and aware of the legal and ethical aspects of the
police use of firearms.

227
 Perhaps the best summary of reasons why the American
police should be armed is, ironically, found in Britain's "Police
Journal":
... U.S. authorities are [not] wrong to arm their police .
They have, indeed, no alternative: the public is armed,
and has a constitutional right to be so, and the inci­
dence of violent crime is such that law-abiding citizens
have only too good a case for possessing arms to defend
themselves and their families. What it does mean,
surely, is that every .Possible step should be taken in
this country to minimize the amount of firearms in pri­
vate hands and to confine their possession to those who
have legitimate cause to own them . It also means that
the police and the courts of justice should make unmis­
takably clear that the criminal use of firearms will
result in swift arrest and draconian punishment. 10
British police experience offers no solid precedent for dis­
arming the American police. It seems clear that this should
happen (if at all) only after the American public is disarmed.
The American custom of arming police officers is as strongly
based on evident need as the British practice to the contrary.
In the meantime, there remains the problem of developing
effective policy to govern the use of firearms by American po­
lice officers. Unlike many matters of police routine, the use of
firearms in a combat situation is an area of concern not only to
the police organization but to the community as a whole.

A CONVERGENCE OF INTERESTS
The protection of life, one of several primary police goals, is
steadfastly pursued by police personnel throughout the United
States. Commendable as the goal may be and laudable as the
total police effort designed to achieve it, many police adminis­
trators have failed to formulate firearms policies or have pub­
lished policies which are, upon scrutinizing examination, ab­
surd, weak, or meaningless and in essence do not serve to
protect life. The vacuum so created must be filled. Policy, fol­
lowing a gunshot death, is easily rewritten; human life, once

228
taken, lingers only in photographs, mementos, epitaphs, and
newspaper accounts.
The image and reputation of a public organization as well as
the life of citizens are not the only factors at stake when the
police resort to the use of firearms. The municipality may well
find itself a co-defendant in civil actions as a consequence of
police firearms use. This fact was amply illustrated by a 1959
New Jersey Appellate Division ruling which asserted that a
municipality is liable under civil law when a police officer negli­
gently shoots an innocent citizen, and it is shown that the offi­
cer has not received adequate instruction from his department
on how and when a firearm may be used. u
In addition to serving the best interests of the public, the
government, and the police department, an effective firearms
policy is of vital interest to the police officer as an individual.
The officer who must ultimately reach a decision whether to
shoot faces the possibility of lawsuit and lifetime remorse if his
decision to shoot proves, in light of subsequent examination,
somewhat more faulty than his aim. On the other hand, the
decision not to shoot (or faulty marksmanship) may be per­
ceived as holding its own unhappy consequences if he himself is
the target of the person he chose not to shoot.
Assaults on police officers are, it seems, becoming more
numerous, and constitute a continuing threat to social order.
The most convincing evidence is that during the six-year period
of 1960 through 1965, 278 American police officers were slain
while performing their duty. 12 During this period, 96 percent of
all officers murdered were killed by rifles, shotguns, or hand­
guns, with handguns predominating. There were 362 suspects
involved in the 279 killings. Of these, 304 were arrested, 43
were killed justifiably by police at the time of or shortly after
the incident, 13 committed suicide, one died a natural death,
and one drowned before being taken into custody. A review of
FBI files reveals that of the 362 persons involved in the fatal
incidents, 76 percent had prior records of arrest, and about 181
of these persons had previously been arrested for assaultive
types of crimes. More than one of every four suspects was on
parole or probation when he murdered the officer.

229
 Non-fatal assaults on police officers continue at a rate
which alarms the nation's police. The trend is apparent when
considering that in 1960, 9621 were physically assaulted; in
1964, 18,001 were victims of physical assault; and in 1965,
20,523 officers were assaulted, with almost four of every ten
attacked suffering personal injury. 13
The apparent increase in the risk of physical violence suf­
fered by the nation's law enforcement officers argues most ef­
fectively for the development of sound policy for police use of
firearms. Such statistics suggest that the individual officer is
being called upon more frequently to make critical decisions,
often under conditions of extreme physical exertion and emo­
tional stress, regarding the use of force in the performance of
his duty. In an emergency situation, the officer equipped with a
thorough understanding of a realistic and effective policy for
the combat use of firearms will be in an advantageous position
to react instantly and prudently.
Adequate guidelines for police use of deadly force should be
the concern of everyone with an interest in the complex process
of law enforcement in modern society. The citizen is interested
in his personal safety and that of his family. The governmental
unit cannot afford to ignore the possibility of civil suit or the
political ramifications of improper police conduct. The law en­
forcement agency must consider its fundamental goal of the
protection of life and the vital nature of its relationships with
the public it serves. And finally, the individual police officer is
entitled to effective guidance in an area fraught with potential
physical danger and civil liability.

THE FIREARMS POLICY MILIEU

In an area of such immediate professional, political, and
public concern, it is both proper and timely to consider how
comprehensively today's American police officer is trained in
the use of his personal weapon, and to what extent his parent
police organizat ion has concerned itself with promulgating
department-wide policy which concisely, yet clearly, outlines
conditions under which he may wisely use his firearm.

230
 Most police departments have recognized and acted upon
the need to train their personnel to attain and maintain a cer­
tain degree of proficiency in the accurate discharge of the di­
verse weapons in the police arsenal. The most widely used of
these is the revolver. Police training also normall y includes at
least some instruction on the care and maintenance of the per­
sonal sidearm and on basic safety rules both on and off range.
Since police officers are taught how to shoot, it might be
assumed that police departments have concomitantly prepared
adequate regulations governing the use of firearms, and that
police officers are instructed when they may shoot. Actually,
research discloses that such is not the case; many departments
have never reduced firearms rules or policy to written form.
Instead, when to fire is frequently trusted to the judgment or
discretion of officers as individuals. Some departments have
ignored the issue completely and have never considered articu­
lating such a policy. Finally, some departments function with
policies so outdated or unrealistic that they actually have no
practical application, and are worthless and often dangerous as
guidance to police personnel. 14
The consequence is that while officers know the mechanics
of care and use of their firearms, many have little or no under­
standing of when the weapon may be employed. This paradox is
similar to teaching someone to drive an automobile while ne­
glecting to instruct him on motor vehicle regulations. It might
be as logically argued, as it often is in the case of firearms
regulations, that the driver's "common sense" coupled with a
warning not to crash into anybody unless absolutely necessary
would suffice to enable the driver to operate his vehicle at large
on the highways.
Some police administrators disclaim the need for written
rules and regulations governing the use of firearms on the
premise that an oral policy is sufficient to guide the actions of
officers under any combination of circumstances. There is, of
course, nothing inherently wrong with a comprehensive oral
policy that is fully understood and uniformly interpreted by all
members of the department. Unfortunately, an oral policy may

231
 prove to be an excuse for the department's failure to examine
thoroughly the issues at stake and to provide concrete guidance
for its officers. Oral policy too often means no policy at all.
Some police departments which at one time relied exclu­
sively on oral policy have prepared written firearms rules and
regulations. However, many written policies prove to be incom­
plete or confusing, and consequently neither fit the needs of the
department, the individual officer, nor the community. Essen­
tially, personnel are left to operate within the vague limits of
what constitutes their individual "understanding" of the mean­
ing of departmental policy. For example, for years the only
directive regarding firearms of one police agency of more than
100 sworn members in southwestern United States was:
"Never take me out in anger; never put me back in disgrace ."
Until this situation was changed, this department, like many
others, simply paid lip service to the issue of police firearms
policy formulation; the police administrator, when asked
whether his department had a written firearms policy, could
respond affirmatively. However, a review of his policy shows it
to be as conspicuous by its insufficiency as a garden hose at a
three-alarm fire.
Another department of over 60 sworn members in the
Rocky Mountain area adopted a policy which limited use of
firearms to self-defense and against misdemeanants. All other
issues remain undiscussed:
Except in self-defense, an officer shall not use a deadly
weapon or take life to effect arrest for a misdemeanor,
whether his purpose is to kill or merely to stop the
other's flight. This is true even though the offender
cannot be taken otherwise.
One police department of about 100 men in the southern
United States devoted almost eight pages of its rules and regu­
lations to a careful description of policies surrounding uniform
allowances and specifications. Paradoxically, in the less than
one page it devoted to the use of firearms , the department
completely failed to present any ground rules which officers
should observe in how and when to use their weapons . Officers

232
of this department were left to operate according to the follow­
ing ambiguous warning:
Unnecessary and careless handling of firearms may
cause accidents, and the drawing, aiming, or snapping
of firearms within Police Headquarters, or in other
places, is forbidden.
Other police firearms policies, assembled from several
American municipal law enforcement agencies, are contained in
their entirety in such statements as:
Officers shall not immediately fire their gun except as
authorized by law.
Leave the gun in the holster until you intend to use it.
Shoot only when absolutely necessary to apprehend a
criminal who has committed a major felony.
Never pull a sidearm as a threat, and if it is drawn, be
prepared to use same.
It is left to the discretion of each individual officer
when and how to shoot.
Such written statements do not represent the entire fire­
arms policy for any police agency. What happens, of course, is
that the department has as mimy firearms policies as it has
members, with each "policy" sharing only the meager core fur­
nished by such vague and inadequate official policy as noted
above. Some forces (including one very large one) simply in­
struct their officers to "read the law book" to discover firearms
policy. Even if all police officers were required to have a degree
in law, a "read the law book" approach invites embarrassment,
and consequent lack of compliance.
The net effect of an inadequate written firearms policy, or
one which is oral, is to shift the burden of full responsibility for
using firearms to the shoulders of the police officer at the level
of execution. The field officer carries heavy responsibility with­
out commensurate guidance. The reluctance of a police agency
to formulate and publish a comprehensive firearms policy

233
seems to indicate that police administrators have either failed
to understand the importance of such regulations, neglected to
consider the multitude of socio-legal factors inherent in police
use of deadly force, or have avoided committing themselves in
what is perceived as a delicate policy area. Whether through
ignorant neglect or conscious avoidance of responsibility, the
police administrator who fails to provide his men with an ade­
quate firearms policy is failing to meet his responsibility to his
chief executive, his community, and his men. To the extent that
the members of a department are more knowledgeable about
how to shoot than when to shoot, the police administrator may
be charged with a critical management failure that ultimately
may lead to an untimely shooting, human tragedy, censure, and
suit.
That such charges are made is confirmed in a recent news­
paper account:
The American Civil Liberties Union of Washington
has requested an inquiry into Seattle Police Depart­
ment regulations governing police use of firearms .
David Guren, ACLU executive director, told Mayor
J. D. Braman that police directives on the use of weap­
ons are va~e and put 'tremendous burdens of interpre­
tation' on mdividual officers. 15

THE FORMULATION OF FIREARMS POLICY

The police administrator who undertakes to formulate a
firearms policy for his agency quickly confronts several factors
which must be acknowledged and faced if his policy is to be
both realistic and effective. Such limiting aspects may be classi­
fied, respectively as social, legal, operational, departmental,
and supervisory.

Social Factors
In theory, legislation constitutes the will of the people. It is
a formalized expression of public policy. In a theoretical sense,

234
 then, the police firearms policy could be based exclusively on
existing legal structure and legislation. Unfortunately, existing
law, because of its complexity and lack of susceptibility to
change, often fails to be in exact accord with the desires and
opinions of the majority of the public. Thus any attempt to
formulate policy in the sensitive area of the police use of fire­
arms must take into account the current state of public opinion
in addition to the position of existing law.
Public definitions of "proper" police conduct reflect wide
variations geographically. For example, provincialism, in a
sense, can influence firearms policy. Reputedly "tough" com­
munities or cities in frontier-like settings may tolerate more
aggressive police behavior than the more sedate, sophisticated
college town. Variations in police firearms policy may be noted
between what New Englanders will accept as compared to per­
sons living in the southwestern United States. "Gunslinger"
police officers carrying sidearms with inlaid ivory pistol grips
would attract only nominal attention in a few American locali­
ties, but in most an officer so armed would bring horror, con­
sternation, and a switchboard deluge at city hall. Communities
which are dominated by particular religious groups or whose
culture is closely allied with a particular ethnic background may
be broadly influenced in a social sense by tradition, heritage, or
belief which predisposes one toward the police agency. Such
communities would surely reject certain kinds of police conduct.
The police administrator must not be swayed by the de­
mands of vociferous, radical, or extreme elements of the popu­
lation, but should take into consideration the political and social
tenor of his jurisdiction as a whole when preparing a firearms
policy. He must take into account the type of regulation the
public will approve and support as "right" or "fair" within the
framework of the objectives of the police operation. While the
law authorizing the police to use deadly force to apprehend or
prevent the escape of a felon makes no distinction as to the
suspect's age, sex, or specific felony, the administrator can rest
assured that the public will make such distinctions and react
accordingly if, say, a youth be shot. This sort of consideration is

235
embodied in the introduction to a firearms policy currently in
effect in one midwestern city:
Our policy provides more stringent requirements
than does the law, but our policy is based upon estima­
tion of the degree of protection needed by the peoyle of
the City of - - and the kind of action they wil sup­
port.
Regardless of what may be legally permissible in a commu­
nity, public opinion must be reflected in the policies of the po­
lice agencies which represent them.

Legal Factors
Obviously, a police administrator has no license to contra­
dict the provisions of existing law as he drafts a firearms pol­
icy. Yet laws and interpretations of laws vary from jurisdiction
to jurisdiction, and, although there are some areas of general
agreement on the meaning of the law, there are many areas in
which legal practice varies considerably and often rests in a
twilight zone of ambiguity. In the formulation of policy, the
administrator must not only take advantage of firmly estab­
lished law, but must recognize the shades of gray as well. State
laws governing the use of firearms by police officers will gener­
ally be found either under the justifiable homicide provisions of
the criminal code 16 or in the form of statutes which deal specifi­
cally with the use of deadly force by police officers. 17 Finally,
some states regulate the use of deadly force by the police
through case law resulting from court interpretations of exist­
ing statutes and from judgments stemming from civil actions. 18
Self Defense. A police officer is clearly justified in using
deadly force when it is necessary to save himself, a citizen, or a
prisoner from death or grave bodily harm, regardless of the
offense for which an arrest is being made. This is also true in
the case of overcoming the resistance of a person already in
custody who resorts to an assault likely to produce death or
grave bodily harm. Even if the arrest is illegal, it has generally
been held that an officer may kill in self defense if the arres­

236
tee's resistance constitutes aggression sufficient to justify the
application of the self-defense rule. 19
An officer, of course, is permitted to use only such force as
is necessary to overcome resistance encountered or to protect
his person from great bodily harm. He is not permitted to use
deadly force to protect himself from assaults which are not
likely to have serious results. The self-defense concept applies
equally to overcoming attacks by juveniles since an assailant is
not necessarily less desperate or dangerous simply because of
his youth.
Misdemeanants. An officer may not use deadly force to
effect the arrest of a person suspected of being a misdemean­
ant. This restriction, however, does not infringe upon an offi­
cer's right to self-defense should he be attacked by a misde­
meanant suspect or prisoner.
The use of firearms in halting a fleeing misdemeanant is
unwarranted, regardless of whether the attempted arrest is
with or without a warrant. In such cases, the value of human
life is considered to supersede the importance of immediate
apprehension. In a moral sense, this restriction is justified,
even though in some jurisdictions the very act of fleeing consti­
tutes a felony, and an officer may be permitted by law to use all
possible force, including firearms, to prevent the escape of any
person arrested on a warrant after notice of arrest has been
furnished. The Pennsylvania Supreme Court stated the princi­
ple clearly:
... where a misdemeanor has been committed and is
charged in the warrant, flight from an officer, even
after actual carture and custody, there having been no
conviction;~ wil not justify the use of a deadly weapon
... An omcer has no right to shoot a person who is
merely running away from him without committing any
violence, when under arrest or to avoid arrest for mis­
deamor. He must at least stop short of force which will
result in the sacrifice of human life; and a killing in such
a case is manslaughter at least. 20
Felons. Far less susceptible to clear statement are rules
governing the use of deadly force against the felon. The rule

237
 that an officer may, if necessary, take the life of a felon in
effecting an arrest seems linked to ancient law under which
almost all felonies were punishable by death. This was true as
recently as 1800.21 Since 1800, however, various state legisla­
tive bodies have given felony status to such non-violent offenses
as statutory rape, larceny, sodomy, extortion, perjury, forgery,
adultery, bigamy, incest, and pandering, to name a few. In
consequence, the traditional concept that an officer may use
whatever force is necessary to effect the arrest of a known
felon has become complicated.
In most states it is still lawful for the police to risk killing a
suspect when necessary to effect his arrest, or to prevent the
escape of a person guilty of a felony if no other means of pre­
venting flight is reasonably available. However, the application
of this principle has from time to time aroused public resent­
ment to the point where opinion ha s virtually suspended the
policeman's option to use firearms under conditions where the
suspect indeed offered no resistance but merely fled. 22
In response to the statutory broadening of the felony cate­
gory, many police agencies have arbitrarily classified some felo­
nies as major, atrocious, extreme, dangerous , serious , or hei­
nous to differentiate them from felonies which are less
aggressive in character and presumably less offensive to the
collective conscience of contemporary American society. Al­
though undoubtedly humanitarian, such arbitrary subclassifica­
tion has indeed complicated , not simplified, the formulation of
police firearms policies. The natural result has been the appear­
ance in some existing firearms policies of such terms as serious,
dangerous , and major felonies. These terms usually are not
further described or defined, with the result that an officer
must apply his own split-second interpretation of the classifica­
tion in the heat of tense police combat.
The rationale which supports the subclassification of felo­
nies is that society will not tolerate the killing of a person
escaping from a crime that would carry a three or four-year
sentence or committing the common law offense of fleeing from
a police officer punishable by an even lighter sentence. The

238
appeal of this argument is even further augmented in those
states where capital punishment has been abolished entirely. 23
It seems to follow that when many states, even after conviction
of murder, do not invoke the death penalty, then police use of
firearms should occur only under circumstances which strongly
justify force which may have fatal consequences. This principle
seems valid whether or not a state has capital punishment.
There is really only one avenue open to the police adminis­
trator who is designing a modern, utilitarian, concise firearms
use policy. It is to flatly prohibit the use of any fatal force
except in self-defense or the defense of a citizen after all other
means have failed. Such a policy clears the issue of shooting at
fleeing felons, obviates the difficulties involved in arbitrarily
subclassifying offenses, and eliminates public clamor over lives
being taken for seemingly trivial offenses. Such a position
would coincide with the policy which the FBI has adopted, the
essence of which is explained by Director J. Edgar Hoover:
The FBI's policy with respect to the use of firearms
in connection with official duties is relatively simple and
certainly comes within the moral, ethical and legal
framework governing the use of force by a law enforce­
ment officer in performance of duty. We teach our per­
sonnel, and the same policy is advocated at firearms
schools by our firearm experts, that a law enforcement
officer should use only that degree of force necessary to
overcome resistance to a legal duty he must perform. In
other words, an officer should use his firearm only for
self-defense of another. I have long believed that a
highly developed skill in the use of firearms by Jaw
enforcement LPersonnel] actually saves lives, and cer­
tainly equips the officer to better protect himself and
the citizenry of his community against the criminal who
follows no legal, moral or ethical code. Many criminals
have candidly admitted they offered no resistance to
FBI arrest, although armed, because of the FBI's repu­
tation for training its Agents in the proficient use of
firearms. 24
The modern policy permits the use of fatal force only after
all other means have failed in those instances when the officer

239
believes the felon suspect has used or threatened fatal force
during his criminal act and if he is not apprehended at once he
may seriously injure or kill someone. The American Law Insti­
tute proposed the following as guidelines in the General Princi­
ples of Justification section of its proposed Model Penal Code:
(2) Limitations on the Use ofForce
(b) The use of deadly force is not justifiable un­
der this Section unless:
(iv) the actor believes that:
(1) the crime for which the arrest is
made involved conduct including the
use or threatened use of deadly
force; or
(2) there is a substantial risk that the
person to be arrested will cause
death or serious bodily harm if his
apprehension is delayed. 25
If all departments formulated firearms use policies which
included the above principles and these policies were judiciously
enforced, many of the tragic incidents of the past could have
been averted.
A classic example of an officer's dilemma whether or not to
use fatal force is the circumstance where an officer comes upon
two men who are struggling in a public place. One man, upon
seeing the officer, flees, while the other shouts to the officer
that the fleeing man has just committed an armed robbery.
Does the officer fire? Has he basis for doing so? Does the officer
have sufficient facts to justify the use of deadly force? Is it
better that the suspect escape? These are some of the issues
which flash across the officer's mind in that agonizing moment
of decision-making.
Suspicion or flight alone or in combination are an insuffi­
cient basis for shooting at a suspect. There must be firmly
based knowledge that the suspect has indeed committed a fel­
ony, that the offense was serious and aggressive in nature, that
every other means to stop flight has been exhausted, and that
the suspect, if not apprehended, may seriously injure or kill

240
someone. Even after the split-second evaluation above, the offi­
cer has one more question to answer: "Am I justified in kill­
ing?"

Operational Factors
No firearms policy can be formulated to include specific
instructions covering every possible operational contingency,
but the police officer is certainly entitled to clear guidance in
such relatively common operational situations as firing at re­
sisting offenders, shooting at or from moving vehicles, using
firearms against juveniles, and firing warning shots.
Resisting Offender. A sound principle to follow whe n deal­
ing with the resisting offender is that the same amount of
deadly force may be used to overcome resistance as is autho­
rized to prevent the flight of the same individual. This principle
does not, however, abridge an officer's right to self-defe nse.
Indeed, where significant resistance is encountered, the self­
defense rule will take precedence.
At Moving Vehicles. While it is rare that the police are
specifically prohibited by statute from firing at moving vehicles,
justification for such action may be difficult to find in many
instances. This is especially true when officers have no concrete
evidence that a felony has been committed . In some instances
an officer may know only that the fleeing vehicle has failed to
heed a siren, red light, or roadblock, but may have no knowl­
edge of the driver's reason for evading arrest. Granting the
sure knowledge of a felony, an officer must consider the safety
of the general public. This is especially true in urban a reas,
regardless of the hour. Also, on occasions hostages have been
forced to accompany suspects fleeing in a motor vehicle. An­
other consideration is that gunfire aimed at a moving vehicle is
notoriously inaccurate except in television epics, in class D
movies, and in popular detective comic strips. Finally, standard
police cartridge loads from a .38 calibre revolver often do not
penetrate the metal portions of a motor vehicle but instead
ricochet. Consequently, shots fired at vehicles are far more
likely to constitute a hazard to public safety than enhance it.

241
 From Moving Vehicles. Even though the crime and circum­
stances may justify the use of firearms, municipal police admin­
istrators have substantial basis for requiring their officers to
refrain from firing at fleeing vehicles from pursuing cars or
motorcycles. On the other hand, state or county administrators
may not flatly prohibit such a practice. They may permit the
passenger in a two-man patrol unit to fire from a moving vehi­
cle, if circumstances described earlier warrant such action, pro­
vided there is no apparent danger to persons in the area. Such
gunfire should occur only when almost abreast of or extremely
close to the fleeing and armed criminals. Firing from moving
solo motorcycles (unless a second man is in a sidecar) and from
one-man motor patrol units should be flatly prohibited.
Juvenile Offenders. There is no legal distinction in the use
of deadly force against juveniles as compared to adults. But if a
police department adopts the model firearms use policy sug­
gested below as its guidelines, there need to be no distinction
drawn between firing at adults or juveniles. This is so because
the provisions of the model policy are so demanding that fatal
force is justified only when circumstances have deteriorated to
such a point that officers or citizens are in immediate serious
danger. Under such circumstances, it is not really relevant
whether the suspect is adult or juvenile; age (and indeed sex) is
a secondary factor.
On the other hand, if a police department does not adopt a
firearms use policy as restrictive in nature as that proposed in
the model, guidelines must be drawn which specifically address
the juvenile issue. The police administrator must be certain that
his men fully comprehend the social impact of firing at, wound­
ing, or killing a juvenile.
Firearms may be used against juveniles only under two
circumstances: (1) In the necessary defense of a citizen's life, or
(2) to preserve the life of an officer or a prisoner. In essence,
they are the provisions of the model policy. Under no circum­
stances can the use of fatal force be justified against persons
recognized as or suspected of being below age 18. This places a
grave discretionary decision upon the officer because almost 50

242
 percent of all arrests for serious offenses in 1965 were persons
under 18 years of age. 26
Should a department's policy include provisions not as re­
strictive as those contained in the model, it is essential that an
officer is certain he is directing his fire at an adult felony sus­
pect when not shooting in defense of life. Murky weather condi­
tions, darkness, or the positioning of the sun or artificial lights
render it all the more difficult for an officer who has had only a
fleeting look at a suspect to decide whether the fleeing person
is an adult or juvenile. Added clothing in rainy or cold weather
may cause a slender youth to assume the general proportions of
a middleweight adult. Since police combat often occurs under
the most adverse weather conditions and in an environment
which itself is "spooky" the officer must feel even more certain
that he is justified in using his firearm and that the suspect is
an adult. He must under no circumstances fire upon mere suspi­
cion and at persons who run away merely to avoid facing the
police.
Warning Shots. Interestingly enough, the 1961 Michigan
Firearms Use Study revealed that in some quarters the warn­
ing shot has come to be regarded as a sort of "civil right," with
both public and criminal element expressing indignation when
the custom is bypassed in favor of more immediate action.
Whether the warning shot tradition was an outgrowth of poor
police marksmanship or "professional" courtesy, the practice
has nothing to commend it and should be terminated. This is
consistent with the best interests of public safety and the pre­
vention of incidents likely to result in civil or criminal action
against an officer or his department.
The officer who fires his weapon when deadly force is not
authorized risks wounding or killing an offender or some inno­
cent bystander. Thus, a gun should never be fired over some­
one's head or into the ground merely to frighten a suspect into
submitting to arrest. For every suspect who surrenders upon
hearing warning shots there are others who flee that much
faster.
A final and compelling basis for prohibiting warning shots
is that officers other than the one who fired a warning shot

243
may easily be decoyed into killing a suspect by believing that
the officer's shot was indeed offered to kill, not to warn. When
policy is such that brother officers have but one basis to inter­
pret shooting-that there exists strong cause to take life-it is
clear that there is no room for warning shots.

Departmental Factors
No matter how comprehensive a firearms policy may be and
how carefully it is tailored to be compatible with legal and
social requirements, it will be effective only to the extent that it
is understood by the police officers who must conform to its
provisions and by supervisors who must assure that policy is
followed. Without oversimplification, the policy should be re­
duced to compact terms, free from legal terminology and com­
plicated sentence structure.
The most powerful tool at the disposal of the police admin­
istrator in securing department-wide understanding and accept­
ance of his policies is a comprehensive recruit and regular in­
service training program. And the chief who will not tolerate
irresponsible display or use of firearms is the key to overseeing
such a policy, as are supervisors who are alert to poor police
practice.
In the recruit training program the departmental firearms
regulations and the reasons behind them should be fully ex­
plained. And a few minutes at roll call from time to time can be
effectively employed to refresh veteran officers on the proper
use of weapons and to air practical firearms problems arising in
day-to-day service. Visual training aids depicting combat situa­
tions which cause officers to make independent decisions based
on policy should be stressed.
There are a host of moral or ethical considerations involved
in the use of deadly force. Such issues should be discussed at
length with officers undergoing training. The San Diego, Cali­
fornia, Police Department's recruit training program is perhaps
the nation's best in helping young men and women confront
these problems. However, any effort to treat firearms morality
per se is unlikely to meet with any degree of success. It is easy

244
 enough, for example, to point out that the Bible says, "Thou
shalt not kill," but if each officer has not yet accepted this as a
guiding principle, it is unlikely that any police training program
can firmly implant a sense of right and wrong in the adult
officer. In short, a moral sense of right and wrong as a guide to
behavior becomes an inherent part of personality long before an
individual becomes an adult and a police officer. For this rea­
son, the training program may in part be designed to refresh
the individual's sense of right and wrong, but it cannot be ex­
pected to seriously modify pre-existing attitudes. The police
firearms training program, then, should discuss broad societal
and police-oriented implications of existing patterns of moral­
ity, and emphasize what the officer is or is not permitted to do
by law of the nation and states, ordinances of the community,
and the rules of the department.
In any event, morality is a personal matter and the use of
firearms is a policy matter. If an officer' s personal morality is
in harmony with the department's firearms policy, all the bet­
ter. If not, the officer must be provided with a policy that will
act as the conscience of the community to fill any void existing
personal conscience. This highlights the importance of control
in ensuring that performance at the level of execution is con­
sistent with policy.

Supervisory Factors
The conscientious police administrator will recognize the
fact that effective control of firearms policy will require a con­
cert of high calibre, first-line supervision and the development
of internal machinery to assure that firearms regulations are
obeyed and violations investigated.
The role of positive supervision, basic to implementing any
policy, is perhaps even more critical in the control of firearms
use regulations. While most police officers will readily accept
rules and regulations which are realistic and easily understood,
they invariably look to their supervisors for subtle, individual

245

attitudes toward the formal policies of the organization. The

attitude of the supervisor will be reflected either consciously or
unconsciously in the behavior of the subordinate. For this rea­
son it is essential that the supervisor understand and fully ac­
cept the organization's firearms policy.
Experience has time and again revealed that people re­
spond favorably to positive leadership. It follows that efficient
supervision, positive in nature, will inspire compliance with
firearms regulations on the part of most officers and will iden­
tify those few subordinates who seemingly are unable to comply
with firearms directives. Officers who continue to be unwilling
or unable to adhere to established policy should be retrained or
reassigned to positions where they will be less likely to endan­
ger themselves, the safety of the public, and the reputation of
the police agency.
The critical and sensitive nature of the use of firearms by
the police demands not only the implementation of a modern
firearms policy, accepted by departmental personnel and super­
vised at all levels of command, but the creation of machinery
within the department to insure that such policy is obeyed and
that violations are investigated. This control is achieved by a
process of review of every incident in which police officers re­
sort to the use of firearms. The Oakland, California, Police
Department review process is outstanding.
Departmental regulations should require that an officer re ­
port, in writing, every occasion upon which a weapon is dis­
charged outside of authorized range firing. The officer's imme­
diate supervisor should be obliged to investigate the
circumstances and report on the facts to the divisional com­
mander. A board of review or inquiry, comprised of departmen­
tal command personnel, should be created to investigate both
accidental and intentional discharge of firearms. The police le­
gal advisor or a representative from the district attorney's or
corporation counsel's office may be invited to assume an advi­
sory role if a case seems to so warrant.
Such a review board should be charged with the dual func­
tion of enforcing compliance with firearms regulations and of

246
identifying areas in which policy revision is required or addi­
tional training needed. Administrators should stress the posi­
tive aspects of the inquiry board system in order to maximize
departmental acceptance and, more important, to insure a con­
tinuing flow or field experience which will serve as a basis for
future policy formulation.
Having considered the many complex factors involved in
the formulation of an effective firearms use policy, the police
administrator can translate his conclusions into a written policy
that will adequately reflect the policy position of the depart­
ment and yet remain within the understanding of operational
personnel. Drafting this document will be no easy accomplish­
ment, but the preparation of a clear, concise , and comprehensi­
ble regulation is critical to the success or failure of t he depart­
ment's policy.

A MODEL FIREARMS USE POLICY

The model firearms policy which fits every possible occur­
rence is impossible to draft. But one may be written which sets
forth fundamental philosophy and purpose regarding the poten­
tially fatal use of force by police. Such a policy may also stand
as a carefully devised, practical, understandable , and acceptable
guiding force with a suitable control mechanism to assure that
the policy is indeed adhered to. The Berkeley, California, Police
Department adopted a revised firearms use policy in 1966
which is one of the nation's best. It is similar to the model
policy which follows.
The model firearms use policy proposed below includes four
major sections: (I) A summary statement of policy, (II) the reg­
ulation proper, (III) procedure following a firearms discharge,
and (IV) the review mechanism.
The principles contained in Sections I and II are sufficiently
universal in nature to be applied to a police agency of any size
or location throughout the United States. Sections III and IV,
however , may have to be revised so that its language conforms
with the size, organization, and rank and title terminology of
individual departments.

247
 I. POLICY
The policy of this department is that members
shall exhaust every other reasonable means of
apprehension before resorting to the use of fire­
arms.

II. REGULATIONS27
A. An officer shall not discharge firearms in the
performance of his police duties except under
the following circumstances and after all
other means fail:
1. In the necessary defense from death or se­
rious injury of another person attacked .
2. In the necessary defense of himself from
death or serious injury when attacked.
3. To effect an arrest, to prevent an escape,
or to recapture an escapee when other
means have failed, of a felony suspect
when:
a . The crime for which the arrest is sought
involved conduct including the use or
threatened use of deadly force; and
b. There is a substantial risk that the per­
son whose arrest is sought will cause
death or serious bodily harm if his appre­
hension is delayed.
4. To kill a dangerous animal or one that hu­
manity requires its removal from further
suffering, and other disposition is imprac­
tical.
5. To give alarm or to call assistance for an
important purpose when no other means
can be used.
6. For target practice at an approved range.
B. Firearms shall not be discharged under the
following circumstances:
1. As a warning.

248
 2. At moving or fleeing vehicles unless the
circumstances come within the provisions
of Section IIA, 1, and 3 of this policy.
C. An officer shall file a written report through
established channels to the police chief imme­
diately following the purchase, replacement,
loss, or other disposition of a police firearm,
and shall list a complete description including
the serial number. A report concerning the
loss (including theft) of a police firearm shall
include all facts surrounding the loss.
III. PROCEDURE TO BE FOLLOWED WHEN
FIREARM IS DISCHARGED (Except at an ap­
proved range)

A. Notification and report by member involved.

1. Whenever a member discharges his fire­
arm either (a) accidentally, or (b) in the
performance of police duty, he shall ver­
bally notify his on-duty supervisor as soon
as time and circumstances permit, but in
no event later than the conclusion of his
current tour of duty. If a command officer
is not on duty in his division at the time of
the discharge, the member shall verbally
notify the ranking officer on duty at the
time.
2. The member who discharged his firearm
shall file a written report of the incident
through established channels with the po­
lice chief and a carbon copy with the mem­
ber's superior within 16 hours of the inci­
dent.
3. If the member who discharged his firearm
is hospitalized or fatally injured during the
tour of duty and incapable of filing the re­
port required in paragraph 2 of this sub­

249
 section, his supervisor is responsible for fil­
ing as complete a report as possible
pending further departmental investiga­
tion.
B. Investigation by a Command Officer.
1. Each discharge of firearms shall be inves­
tigated personally by the on-duty com­
mand officer of the member involved. If
the discharge occurs when there is no com­
mand officer on duty in the division to
which the member is assigned, the ranking
command officer then on duty shall per­
sonally conduct the investigation when no­
tified that the discharge of firearms has
taken place.
2. After conducting a thorough investigation
of the circumstances attending the dis­
charge of firearms, the command officer
shall submit a detailed written report of
the results of the investigation to the po­
lice chief through channels. The report
shall also contain the observation and con­
clusions of the command officer as to
whether the discharge was justified and in
accordance with this order.
IV. BOARD OF REVIEW
A. Membership of the Board
1. There is established a board of review con­
sisting of the following members, along
with others who may be designated by the
police chief:
a. The commanding officer of the patrol divi­
sion who is designated chairman of the
board.
b. The commanding officer of the training
section.
c. One supervisory officer of the member
who discharged the weapon.

250

 B. Meetings of the Board.

A meeting of the board shall be called by the

chairman within a reasonable time after the

report of a firearms discharge comes to his

attention.

C. Authority of the Board.

1. The board is authorized to review the cir­
cumstances attending each discharge of
firearms by a member of the department,
and to recommend to the police chief disci­
plinary action. The police chief makes the
final decision whether disciplinary action is
to be taken against the officer, and the
nature and extent of the action.
2. The board shall make/or receive recom­
mendations for the modification of the de­
partment's firearms use policy and shall
make recommendations concerning train­
ing necessary for the effective implemen­
tation of such policy.

Regardless of the exact format adopted, a department's

firearms policy should be written and each member should be
furnished a copy. In addition, the administrator may wish to
have produced a "brief" of the policy which would extract the
essential elements in a condensed form easily handled and car­
ried in the pocket or notebook of officers for periodic reference
purposes.

IMPLEMENTATION
Policy-making is an internal matter, a process which
springs from within each of America's 40,000 police agencies.
The law may dictate, the public may react, and the police de­
partment may suggest, but the police administrator must be
held ultimately responsible for the formulation and implementa­
tion of the firearms policy best suited to the needs of all con­

251
cerned. The chief also bears responsibility for insuring that pol­
icy, once made, is adhered to.
While recognizing that policy formulation is an internal
matter, it is equally apparent that the notable range in scope
and design found among police firearms use policies throughout
the United States is neither efficacious nor desirable. This be­
comes especially evident when one acknowledges that in princi­
ple the 40,000 American police agencies enforce essentially the
same laws and may resort to using deadly force under approxi­
mately similar conditions and circumstances. One may conclude
that it is reasonable as well as timely to encourage the nation's
police to adopt a firearms use policy which is essentially uni­
form in nature.
In its 1967 report to the President, the National Crime
Commission recommended:
A comprehensive regulation should be formulated
by every chief administrator to reflect the basic policy
that firearms may be used only when the officer be­
lieves his life or the life of another is in imminent dan­
ger, or when other reasonable means of apprehension
have failed to prevent the escape of a felon suspect
whom the officer believes presents a serious danger to
others. 28
The model firearms use policy described earlier includes the
policy principles recommended in the National Crime Commis­
sion Report. The model policy provisions are humane , yet oper­
ationally feasible.
If the Commission recommendation is to be implemented,
national and state-level police associations must commit them­
selves to adopt a model policy statement as contained herein .
These associations include many influential groups such as the
International Association of Chiefs of Police, the National Sher­
iffs' Association, the Fraternal Order of Police, and the Inter­
national Conference of Police Associations, as well as branches
of such associations at the state and local level throughout the
United States. And several non-police governmental profes­
sional associations such as the International City Managers'
Association, the National League of Cities, the National Associ­

252
ation of County Officials, the National Conference of Mayors,
and the National District Attorneys Association, to mention
only a few, are concerned with the formulation and the imple­
mentation of effective firearms policies.
Police professional and governmental organizations, as
partners dedicated to fostering broad improvements in law en­
forcement, may advise, assist, and encourage the police to im­
plement sound principles and policy. The model firearms use
policy presented above may be recommended and supported on
the basis of its utilitarian and humane nature, concise direction,
and internal accountability. The fact that it costs nothing to
install, other than the time involved in some careful staff study
backed by recruit, and in-service training time, warrants notice .
It is timely for police administrators to act.

CONCLUSION
Undeniably, the ultimate decision to shoot or not to shoot
rests with each officer as an individual. Consequently, the po­
lice administrator must insure that every officer has a clear
understanding of what is required, permitted, and forbidden by
departmental policy and by the law. This responsibility can only
be met by providing each officer with clear, concise, and com­
prehensive policy guidance and the training and supervision
necessary to insure that the department's policy becomes a
constituent element of the behavior pattern of every individual
concerned.
The very nature of the question of using deadly force in our
free society dictates that the issue will continue to be the focus
of a wide variety of groups external to the police organization.
Effective law enforcement in the United States demands the
implementation of uniform, humane, and workable regulations
governing the police use of firearms. The complexities of a
rapidly changing society will no longer permit the risk of pro­
crastination.
Our large southern police department concludes its fire­
arms training course with the following terse but meaningful
laconism which summarizes the intent of this entire exercise:

253
Quick action is often necessary with no time for deliberations
and you must be right. Ask yourself one question before you
fire-AM I JUSTIFIED IN KILLING? If you are in doubt, DO
NOT FIRE.

Notes
1. There have been two such surveys. The f'lrst, completed in 1961,
surveyed the 71 police departments serving cities of 10,000 and more
population in Michigan. See Samuel G. Chapman and Thompson S.
Crockett, '"'Gunsight Dilemma: Police Firearms Policy," Police 6:40­
50, 60-70, March-April and May-June, 1963. Portions of the Chapman­
Crockett research are included in this report, originally prepared for
the President's Commission on Law Enforcement and Administration
of Justice, with permission from the Charles C Thomas Publishing
Company, Springfield, Illinois. The other report, "Police Regulations
Governing Use of Firearms Survey," was published by the Cincinnati,
Ohio Police Division on April 22, 1964, Stanley R. Schrotel, Chief of
Police. This survey reported police firearms policies in 45 of the 51
cities in the United States over 250,000 population.
2. Gerhard Falk, "The Police Dilemma in England and America," in
Interdisciplinary Problems in Criminology: Papers of the American
Society of Criminology, 1964 (Columbus, Ohio: Ohio State University),
p. 119. Also see: Gerhard Falk, "The Public's Prejudice Against the
Police," American Bar Association Journal, 50:756 August 1964.
3. Letter to Samuel G. Chapman, Assistant Director of the Presi­
dent's Crime Commission, from B. N. Babbington, Director of the
Police Research and Planning Branch, Home Off'lce, London, England
(November 17, 1966).
4. Ibid. Of the three men fatally injured by means other than being
shot or stabbed, one was struck on the head in a chase; another was
killed in a melee; and the third was thrown from a moving vehicle and
run over by a car travelling in the opposite direction .
5. David Anable, "British Police Peril Stirs Arms Debate," The Chris­
tian Science Monitor, August 17, 1966, eastern edition, 5:415.
6. Letter to Samuel G. Chapman, Assistant Director of the Presi­
dent's Crime Commission from Jerome Daunt, Chief, Uniform Crime
Reporting Section, Federal Bureau of Investigation, (December 20,
1966).
7. Babbington, loc. cit.

254
 8. Crime in the United States, Uniform Crime Reports, 1965 (Wash­
ington, D.C. : U.S. Department of Justice, Federal Bureau of Investi­
gation), pp. 33-38. Also see: Allen P. Bristow, "Police Officer Shoot­
ing: A Tactical Evaluation," Journal of Criminal Law, Criminology
and Police Science, 54:93-95, March-April 1963. Professor Bristow's
research discloses that in approximately 18 percent of the incidents his
evaluation considered, the officers were approaching stopped vehicles,
many of which had been halted for routine traffic arrests.
9. Operational Analyses Section, California Highway Patrol, Sacra­
mento, Calif., Bradford M. Crittenden, Commissioner. In addition to
non-fatal assaults on its personnel, the CHP has had 15 of its person­
nel killed by criminal gunfire during the past 30 years. The "Top Ten"
program was started on March 14, 1950, by FBI Director J. Edgar
Hoover. Relying heavily on cooperation of the nation's press, the suc­
cess box score shows 215 captures to April 1966. Twenty-seven of
these were apprehended in California. See the FBI Law Enforcement
Bulletin, 35:2-5 and 19, July 1965.
10. "Firearms," The Police Journal, 38:108, March 1965. Also see:
Quinn Tamm, "Shall Society Disarm Its Police?" The Police Chief,
31:7, August 1964, and Tamm, "Dangerous Dreams of Wishful Men,"
Police, 10:31, March-April 1966.
11. McAndrew v. Mularchuk and Keansburg, 33 N.J. 172, 162 A. (2d.)
820, 88 A.L.R. (2d) 1313 {1960). In this decision the victim, a five year
old boy, was shot in the back at chest level by the defendant who was a
uniformed police officer attempting to apprehend a disputant in a
street argument. The judgment was based on the theory that the
municipality was negligent in permitting an officer to carry a revolver
without having received any training in its use and that the municipal­
ity was liable for acts of its employee while he was acting within the
scope of his employment. Also see: Judge Bartlett Rummel, "Police
Firearms Training: An Inquiry Into the Governmental Duty to Provide
Adequate Training," The National Rifleman, 3:17-22, August 1963.
12. Source: Crime in the United States, Uniform Crime Reports, 1965
(Washington, D.C.: U.S. Dept. of Justice, Federal Bureau of Investiga­
tion), pp.33-38.
13. Ibid., p. 153.
14. The 1961 Michigan firearms policy survey found that 27 of the 50
police agencies which furnished information had no written policies to
govern the police use of firearms. These agencies relied on oral policy
or expected officers to use discretion. In the 1964 Cincinnati survey,
op cit., three of the 45 largest American cities reported they had no
written firearms use policy, and a review of the 42 written policies

255

discloses some which were carefully prepared and comprehensive in

nature and one whose only guidance is to prohibit "warning" shots.
Another instructed officers to "exercise the greatest possible caution"
when using a firearm, and about 10 urged officers to use "good judg­
ment."
15. The Oregonian, December 7, 1966, 18:4.
16. As in Arizona, California, Colorado, Louisiana, Minnesota, New
York, Oregon, Texas, and Washington.
17. As in Georgia, Illinois, Indiana, and Massachusetts.
18. As in Florida, Hawaii, Michigan, New Jersey, Ohio, and Pennsyl­
vania.
19. Roy Moreland, Modern Criminal Procedure (Indianapolis: Hobbs­
Merrill Co., 1959), p. 40; but also see Oscar Leroy Warren and Basil
Michael Bilas, Warren on Homicide, Vol. 1 (Buffalo, New York: Den­
nis & Co., 1938), pp. 662-663. Finally, the court in New Jersey v.
Koonce, 89 N.J. Super. 169, (1965), said: " ... Self-help is antisocial in
an urbanized society .. ..[It seems to guarantee] escalation into blood­
shed. We declare it to be the law of this state that a private citizen
may not use force to resist arrest by one he knows or has good reason
to believe is an authorized police officer, whether or no t the arrest is
illegal."
20. Commonwealth v. Loughead, 218 Pa. 429 (1907). The principle
was set forth even earlier in Reneau v. Tennessee, 70 Tenn. 720, 721
(1879).
21. Moreland, op cit. , p. 29. Also see Lewis E. Lawes, " Capital Pun­
ishment," Encyclopedia of Criminology (New York: The Philosophical
Library, 1949), p. 43.
22. Louisiana is one state which by statute restricts the use of deadly
force to self-defense and to the defense of endangered persons. This
restriction is imposed by the "justifiable homicide" and "defense of
others" sections of the Louisiana Criminal Code, sections 20 and 22.
23. The general trend in American capital punishment seems charac­
terized by abolishment through law or disuse and by legislation to
make a sentence of death permissive rather than mandatory. The
trend to limit the use of the death penalty began early, in 1794, when
Pennsylvania first legislated a category of murder which was not a
capital offense but rather a question of degree. Today nine states
(Alaska, Hawaii, Iowa, Maine, Michigan, Minnesota, Oregon, West
Virginia, and Wisconsin) and Puerto Rico have abolished the death
penalty. Four other states (New York, North Dakota, Rhode Island,
and Vermont) retain it only for a second-time murderer, the slaying of

256
a police officer, or murder by a "lifer." Only one person was executed
in the entire United States in 1966 and only one Federal prisoner has
been executed since 1957. In the military services, the Army's last
execution was in 1961 and the Navy has not executed anyone since
1842. See: Curtis J. Sitomer, "California Groups Attack Capital Pen­
alty," The Christian Science Monitor, April 8, 1967, central edition,
1:3&4.
24. Memorandum from FBI Director J. Edgar Hoover to Attorney
General Nicholas DeB. Katzenbach submitted on April14, 1966.
25. American Law Institute: Model Penal Code, Article 3, Section 3.07
(2) (b) (iv) (1 and 2), May 4, 1962.
26. Crime in the United States. Uniform Crime Reports, 1965 (Wash­
ington, D.C.: U.S. Department of Justice, Federal Bureau of Investi­
gation), p. 114.
27. If a department elects to adopt a policy which permits officers
broader license than proposed in the model, than it should include a
statement in Section II which says, in essence, that: "An officer shall
not fire at persons known to be or suspected of being juveniles (per­
sons less than 18 years of age) except under circumstances which come
within the provisions of Section II A, 1, 2, and 3 of this policy."
28. The President's Commission on Law Enforcement and Adminis­
tration of Justice, The Challenge of Crime in a Free Society, (Washing­
ton, D.C.: U.S. Government Printing Office, 1967), p. 119.

257
 CHAPTER 11

ADMINISTRATIVE INTERVENTIONS ON

POLICE SHOOTING DISCRETION: AN

EMPIRICAL EXAMINATION

JAMES J. FYFE

ABSTRACT
In August, 1972, the New York City Police Depart­
ment promulgated administrative shooting guidelines
and shooting incident review procedures far more re­
strictive than former statutory " defense of life" and
"fleeing felon" justifications for police shooting. Using
a data base that includes all reported New York City
police firearms discharges and serious assaults on police
between 1971 and 1975, this article examines the ef­
fects of the new guidelines and procedures on shooting
frequencies, patterns, and consequences.
Great decreases in "fleeing felon" shootings,
"warning shots," and shooting-opponent injuries and
deaths were found to be associated with the new rules.
This change also appeared to have a favorable effect on
line-of-duty officer deaths and serious injuries. The im­
plications of these findings are discussed.

SOURCE: Reprinted from Journal of Criminal Justice 7, James J. Fyfe, "Ad­

ministrative Interventions on Police Shooting Discretion: An Empirical Exami·
nation," Copyright (1978), Pergamon Press, Ltd.

258
 Perhaps the major paradox of the American system of jus­
tice is the discretionary latitude it allows many of its police
officers in the use of their firearms as a means of "deadly
force." While the system-and the society-argue and agonize
over the death penalty, police shootings generally draw little
attention. It is likely, in fact, that the 1977 execution of Gary
Gilmore generated more publicity and debate than did the more
than 2,500 police shooting deaths that occurred during the ten­
year period between that event and the last previous court­
ordered exercise of deadly force (Milton eta!., 1977:33).
It is true that police use of deadly force differs dramatically
from court-ordered death sentences. Police often use their fire­
arms as the last resort against real and imminent peril; they
often have no choice but to shoot. Judges who elect to impose
the death penalty, in contrast, usually select it from among a
range of alternatives after lengthy deliberation in the safety of
their chambers. Despite these differences, it is ironic that the
system has devised rigid devices to control and review court­
ordered death sentences, but has generally maintained a hands­
off policy where police decisions to shoot are concerned. In­
deed, the system zealously controls the court's power to take
lives on evidence of guilt beyond a reasonable doubt, but fre­
quently gives its police a blank check in deciding whether or not
to shoot at those they have probable cause to arrest.

POLICE SHOOTINGS
Legal Controls and Review Procedures
While Milton et a!., (1977:41-43) detect both a "recent ea­
gerness of the judiciary to impose restraints upon police con­
duct" and an increase in the number of civil actions alleging
excessive police force, police shooting discretion in many juris­
dictions is limited only by the broad common law fleeing-felon
and defense-of-life rules (National Advisory Commission.
1973:18). In addition, police shooting is often subjected to re­
view procedures of questionable effectiveness (Harding and
Fahey, 1973).

259
 The fleeing-felon rule is very briefly, but accurately, de­
scribed by Wilson as police "authority to use deadly force to
prevent escape from any felony charge" (1972). Milton et al.,
(1977:39) note that this principle was defensible when " virtually
all felonies were punishable by death," but question its wisdom
in an era in which the death penalty is all but extinct.
An only slightly lengthier definition of the defense-of-life
rule is offered by Rhine (1968:834):
rrt is the police officer's] general right to use deadlJ'
force for defense of self and others against threats of
death and serious bodily action. In addition, law en­
forcement authorities never have a duty to retreat be­
fore using deadly force, and may always use this for
defense of others solely upon reasonable belief that
they are being threatened with death or serious bodily
harm.
The most striking features of these guidelines are brevity
and breadth. Further, while many states have supplemented
them by legislating more narrow and clearly delineated statu­
tory limits on police shooting discretion, the record suggests
that police officers are rarely penalized for violating them. 1 The
adjudication of violations of either codified common law princi­
ples or more restrictive legislative guidelines requires that they
be subjected to the criminal process. Here, one finds that, even
among those cases that do come before the courts,2 often the
only civilian eyewitness to a police shooting is its subject-if
surviving. The only version of a police shooting that comes to
court attention, therefore, is likely to be that of the police offi­
cer involved. Further, even where alternative versions are of­
fered, the prosecutor must decide to take action if a case is to
go to trial. Several reasons have been proposed to explain the
reluctance of district attorneys to take these cases to trial.
Rhine (1968:856) suggests that it is very difficult to prove crimi­
nal intent in police killings. Harding and Fahey (1973: 298,299)
point out that elected prosecutors may find that both constitu­
ent concern with law and order and the need to maintain a
cooperative relationship with police militate against such prose­
cutions.

260
 Administrative Controls and Review Procedures
The broad nature of legal restrictions on police shooting
discretion and the difficulties of enforcing them have led many
to argue that police agencies should formulate narrower admin­
istrative guidelines and internal procedures for the review of
shootings (President's Commission, 1967:189, 190; American
Bar Association 1973:125-31). Milton et al., (1977:45-57) report
a "clear trend" toward the adoption of such policies. They note
also, however, that where administrative standards are opera­
tive, they vary widely and are frequently contradictory or (per­
haps intentionally) vague. As a result, "their impact on the
conduct of police officers is questionable." Further, the adop­
tion of administrative guidelines and review procedures is often
resisted by police who perceive such rules as arbitrary restric­
tions on their ability to defend themselves (Berkley, 1969;
McKiernan, 1973). This argument is based on the premise that
such regulations would promote police reluctance to shoot when
necessary for self-defense out of fear that their split-second,
life-or-death decisions will be subject to leisurely second­
guessing (Rubinstein, 1973:333).

NEW YORK CITY: A TEST CASE

One agency that has adopted clearly delineated administra­

tive shooting guidelines and review procedures is the New York
City Police Department. In August, 1972, that department
promulgated Temporary Operating Procedure 237 (T.O.P. 237),
a directive that narrowed officer shooting discretion considera­
bly more than did New York's statutory provisions. In addition ,
T.O.P. 237 established a high-level Firearms Discharge Review
Board (FDRB) to investigate and adjudicate all officer firearms
discharges.
T.O.P. 237 refined New York's penal law restrictions on
police shooting (which are based on the American Law Insti­
tute's 1962 Model Penal Code and permit officer use of deadly

261
force to "defend life" or to arrest for several specified "violent
felonies") by providing that:
a. In all cases, only the minimum amount of force will be
used which is consistent with the accomplishment of a
mission. Every other reasonable means will be utilized
for arresting, preventing or terminating a felony or for
the defense of oneself or another before a police officer
resorts to the use of his firearm.
b. A firearm shall not be discharged under circumstances
where lives of innocent persons may be endangered.
c. The firing of a warning shot is prohibited.
d. The discharging of a firearm to summon assistance is
prohibited, except where the police officer's safety is
endangered.
e. Discharging a firearm at or from a moving vehicle is
prohibited unless the occupants of the other vehicle are
using deadly physical force against the officer or an­
other, by means other than the vehicle. (NYPD, 1972:1)
Except for some minor changes in 1973 (NYPD, 1973:2),
these provisions have been in effect since 1972. So too, has the
FDRB, which is chaired by the Chief of Operations (the depart­
ment's highest ranking sworn officer), and which also includes
as members two deputy police commissioners and the supervi­
sor of the Police Academy's Firearms Unit. FDRB is empow­
ered to conduct hearings at which it may question civilian wit­
nesses, the officer involved, the officer's commander, or any
other officers. Its findings are submitted in the form of recom­
mendations to the commander of the officer involved. These, a
review of case dispositions reveals, fall into one or more of the
following categories:
1. The discharge was in accordance with law and depart­
ment policy.
2. The discharge was justifiable, but the officer should be
given additional training in the use of firearms or in the
law and department policy.
3. The shooting was justifiable under law, but violated de­
partment policy and warrants department disciplinary
action.

262
 4. The shooting was in apparent violation of law and should
be referred to the appropriate prosecutor if criminal
charges had not already been filed.
5. The officer involved should be transferred (or offered
the opportunity to transfer) to a less sensitive assign­
ment.
6. The officer involved should be the subject of psychologi­
cal testing or alcoholism counselling.3
T.OP. 237 became the core of this study, which attempts to
examine the impact of that directive on the frequency, nature,
and consequences of police shooting in New York City.

DATA SOURCES
The primary data for this study consist of 4,904 Firearms
Discharge/Assault Reports, (FDAR) or all of those filed by offi­
cers who had reported discharging their firearms (N = 3,827)
or being the subjects of serious assaults (assaults with deadly
weapons that may have resulted in officer death or serious
injury) between January 1, 1971 and December 31, 1975. These
FDAR reports, supplemented by various personnel records,
were converted to computer data and analyzed using the Statis­
tical Package for the Social Sciences (Nie et al., 1975).
We would have preferred to have included in this data set
reports on shootings and assaults on officers that occurred dur­
ing the two or three years immediately preceding 1971. While
this endeavor might have strengthened our analysis by provid­
ing a clearer description of shooting frequencies and patterns
before the promulgation ofT.O.P. 237, we were precluded from
undertaking it by the unsystematic and incomplete shooting
and assault data available. Of necessity, therefore, we con­
tented ourselves with the inclusion of data on shootings and
assaults on officers reported during the nineteen months pre­
ceding T.O.P . 237.
A potential weakness of this data base (or any other that
consists of incident reports) involves the degree to which events
have not been reported or have been inaccurately reported.
Because officers who fired justifiably would not be disciplined

263
for shooting but would be charged for failing to report, we
concluded that missing data would most often include shootings
of questionable justifiability. Even in violative cases, however,
officers would be unlikely to omit reports of shootings in which
they had (or thought they might have) hit someone or some­
thing. They would also be unlikely to omit reports of shootings
perceived as likely to be brought to official attention by third
parties, including their colleagues. Because of New York City's
population density and because its police rarely work alone, it is
unlikely that more than a very few police could fail to file inci­
dent reports in confidence that their shootings would not other­
wise come to light. We concluded that the problem of missing
data was of minimal import.
The problem of inaccurate reports was viewed as more sub­
stantive. Following the T.O.P. 237 ban on warning shots, for
example, we had found a great and unexpected increase in
reported accidental "shots in the air" fired by police who
"tripped on curbs" while pursuing fleeing suspects (Fyfe,
1978:316-28). FDRB generally recommends disciplinary action
against officers who fire warning shots, but usually refers acci­
dental shooters to nonpunitive tactical retraining classes. It is
probably, therefore, that this specific pattern change reflects
altered reporting behavior rathe r than changes in actual field
behavior. To minimize the effects of this and any other distor­
tions, we attempted to limit our analyses to variables reasona­
bly immune to reporting bias. 4

ANALYSIS

Shootings and Intra-community Violence

In related research (Fyfe, 1978:32-106), we had found
strong correlations among the geographic distributions of
shooting incidents, arrests for felonies against the person, and
reported murders and non-negligent manslaughters (Pearson's
r = +0.62 and +0.78, respectively). Therefore, we commenced
the present investigation by examining the relationships be­
tween these variables over the five years studied. If we found

264
that these associations also existed over time, it would be rea­
sonable to conclude that changes in shooting frequencies were
at least in part attributable to these nonorganizational varia­
bles.
Table 1, which contrasts annual shooting incidents (which
may include one or more officer shooters who fired at the same
time and place) with reported criminal homicides and arrests
for felonies against the person, presents strong evidence to the
contrary. After peaking in 1972, homicides remain fairly con­
stant over the period studied (the relatively large 1973-1974
decrease is only a 7.5 percent decline), while arrests register a
regular annual increase and firearms discharge incidents de­
cline annually after a large increase between 1971 and 1972.
More specifically, columns four and five reveal considerable
variation in annual ratios of homicides/shootings and arrests/
shootings before and after the promulgation of T.O.P. 237. In
1971, there were 2.33 reported criminal homicides for every
police shooting in New York City; this ratio declines slightly (to
2.11) in 1972, then increases considerably over 1973 and 1974
to a high of 3.67 in 1975. Perhaps most significant for the
purposes of police administrators, the table indicates that the
annual ratio of arrests/shootings has been nearly doubled over
the five years studied (from 47.62 in 1971 to 86.88 in 1975).

T.O.P. 237
The decline in reported shooting incidents in the face of a
continuing increase in arrests since 1972 suggests the interven­
tion of another variable. A logical first subject of investigation
in looking for such an event is T.O.P. 237, which became effec­
tive in late 1972, after which the relationship seems to have
changed. Our examination of the association of T.O.P. 237 with
decreased shooting frequencies commenced by dividing the five
years under study into two-month periods and displaying the
number of reported officer shooters and shooting incidents for
each, as shown in Figure 1. The two-month observation periods
were chosen to refine the trend as far as possible without losing
information: they reduce the total of observations from sixty to

265
t--:)
O"l
TABLE 1 VIOLENT CRIMES AND POLICE SHOOTINGS BY YEAR, 1971-1975
O"l

Ratios
Reported Felony Police Homicides/ Arrests/
Year Homicides• Arrests•· b Shootings Shootings Shootings

1971 18.2% 17.1% 21.5% 2 .33 47.62

(1466) (30002) (630}

1972 21.0 18.9 27.5 2.11 41 .18

(1691) (33070) (803}

1973 20.9 20.1 19.6 2.93 61.26

(1680} (35163) (574}

1974 19.3 21.7 16.1 3.30 80.62

(1554} (37971) (471}

1975 20 .5 22 .2 15.3 3 .67 86.88

(1645} (38922) (448)

Totals 100.0 100.0 100.0 2.75 59 .85

(8036} (175128) (2926)

• Calculated from : New York City Pol ice Department (December, 1971-1975). Monthly arrest report.
• Includes murder, non -negligent manslaughter, forcible rape, robbery, felonious assault.
NOTE: Subcell percentages may not total 100.0 due to rounding.
thirty and allow for the data to be cut September 1, close in
time to T.O.P. 237 (August 18, 1972) and its slightly altered
successor, I.O. 118 (August 27, 1973).
The chart, and the overall r's of -0.62 and -0.64 for offi­
cer shooters and shooting incidents, indicate that both phenom­
ena have fluctuated somewhat, but have been declining steadily
since the period during which T.O.P. 237 became effective; "of­
ficer shooters" show a peak of 210 during May and June, 1972,
decline to 175 during July and August, 1972, and never again
reach either level. Similarly, "shooting incidents" peak at 149
during the May-June, 1972, period, decline to 141 during July­
August, 1972, and remain below those levels for the duration of
the period studied.
One cannot argue, of course, that incident declines are en­
tirely attributable to T.O.P. 237. Indeed, the "shooter" and
"incident" r values of 0.38 and 0.41 serve notice that fewer
than half these variations are explained by the passage of time.
Many other variables over which the department has little or no
control (e.g., economic and social conditions, the numbers of
officers available for street duty) are certain to have affected
these frequencies. On the other hand, splitting the chart at
September 1, 1972, produces Pearson's r values of +0.88 (offi­
cer shooters) and +0.94 (shooting incidents) for the earlier per­
iod and respective r's of -0.70 and -0.66 for the latter. It is
evident, therefore, that T.O.P. 237 was accompanied by rather
dramatic changes in the frequencies with which New York City
police officers reported discharging their firearms.
1.0. 118
The second department firearms policy statement was I.O.
118, which was issued on August 27, 1973. The major purposes
of this directive were to clarify T.O.P. 237's ambiguities and to
establish decentralized "area level" review boards, 5 but the doc­
ument appears at the beginning of a six-month decline in shoot­
ing incident frequencies and an eight-month decline in officer
shooter frequencies. Some percentage of these decreases may
be attributable to a cold weather slump. But the declines con­
tinued in the absence of other possible explanations (e.g., the

267
 FIGURE 1. POLICE FIREARMS DISCHARGES, JANUARY 1, 1971-DECEMBER 31, 1975.
!:':)
0)
00 1.0 . 118 r ,. n·
overall = -0.62 0.38 30

OHicer

; :·jP. 237
1171-8131172 • +0.88 0 .77 10

Shooters

9172-12175 = -0.70 0 .49 20

Incidents overall

1171-8/31172
.=
-0.64
+0.94
0.41
0 .88
30

10

I ' 9172-12175 = - 0.66 0.44 20

r.. ,. ,. . ,. . ,./
//

\
\ / A\

I
I \.,.-,.,
l'J I
Jl\
\ I
"\
I\

/ \I
.., \
'V ,........'
\
reponed olf•cer
' shoote rs

reponed shool•ng
1oc1dents

o L---------------------------~L-------------L----------------------------------------
1971 1972 1973 1974 t975

•N • number of blmOnlhly penods.

department was, at that time, adding to its ranks; reported
homicides and arrests for violent crime did not show similar
declines), and are at least as severe and as long as those that
followed the issuance of T.O.P. 237.
Because so few observations are included in this second
period, it was decided not to attempt to test their possible
significance statistically. It would be interesting, however, to
continue following the frequencies of shooting incidents and
reported shooters consequent to other clarifications, procedural
alterations, and minor discretionary changes. If further de­
clines occured, one might postulate that substantial influence is
exerted on shooting frequencies by continuing emphasis on lim­
its to individual officer discretion or the institution of decentral­
ized reporting and review procedures.

Weekly Means
The two-month periods studied thus far are, of course,
rather inexact and may be criticized if used as a basis for com­
parison because they include varying numbers of days. July/
August periods, for example, encompass sixty-two days, while
January/February includes only fifty-nine (in all years except
1972, which was a leap year). In addition, bimonthly figures do
not allow the data to be split precisely at the effective date of
T.O.P. 237 (August 18, 1972).
To provide more comparable figures, the five years under
study were therefore split at midnight, August 18, 1972, and
means of reported officer reason for shooting were computed
weekly for each period (period 1 weeks = 85.1; period 2 =
175.7). The results are presented in Table 2. Its column totals
reveal that a weekly mean of 18.4 officers reported discharging
their firearms prior to T.O.P. 237 and that this figure dropped
to 12.9 officers after the directive was put into force (this rep­
resents a decrease of 29.9 percent). Further, the table shows
important changes in the reasons given by officers in reported
discharging of firearms {p = 0.001; v = 0.28). Indeed, while the
weekly mean reported "defense of life" shootings has de­
creased (from 11.9 to 9.0), the pre-T .O.P . 237 percentage for

269
these shootings (65.8 percent) has increased since the order (to
70.6 percent). Concomitantly, both the weekly mean and per­
centage of "prevent/terminate crime" shootings (usually
fleeing-felon situations) have decreased substantially (from 3.9
and 21.4 percent to 0.6 and 4.6 percent, respectively). Con­
versely, Table 2 demonstrates that both the weekly means and
percentage of suicide attempts have remained relatively con­
stant. As one would expect, T.O.P. 237 has little effect in deter­
ring suicides.
Warning Shots
While Table 2 demonstrates reduced shooting frequencies
and varying reported reasons for shooting, it leaves unan­
swered many questions about other consequences of T.O.P.
237. Its frequencies, for example, are confounded because they
include warning shots, which were not treated separately but
were coded on the basis of the officer shooter's reported intent.
If, for example, a police officer fired a shot into the air in an
attempt to stop a fleeing burglary suspect, the shooting would
be classified as "prevent/terminate crime" and would be in­
cluded in Table 2. What Table 2 does not provide, therefore, is a
measure of T.O.P. 237's effect in reducing shots fired at fleeing
burglars or other opponents. Instead, it indicates that before
T.O.P. 237, 3.9 officers per week reported discharging their
firearms to prevent or terminate crimes, regardless of whether
or not they fired at or over the heads of suspects, and t hat this
mean has subsequently declined to 0.6. Since, in addition to
mandating the use of " every other reasonable means ... for
arresting, preventing or terminating a felony . .. before a po­
lice officer resorts to the use of a firearm," T.O.P. 237 flatly
prohibits warning shots, one might expect that much of the
decrease discussed above is attributable to lower warning shot
frequencies rather than to reduced numbers of shootings at
human targets.
To control for this possibility and to better measure relative
frequencies of officers who shot at targets (except, obviously,
accidental shooters), Table 3, which excludes all officers who

270
 TABLE 2 OFFICER-REPORTED REASON FOR SHOOTING
Pre- Weekly Post- Weekly Weekly
T .O.P. 237 Mean T .O.P. 237 Mean Totals Mean

Defense 65.8% 11.9 70 .6% 9.0 67 .9% 10.0

of Life (1 016) (1582) (2598)

Prevent or 2 1.4 3.9 4.6 0 .6 11 .3 1.7

Terminate (330) (103) (433)
Crime

Destroy 4.4 0.8 11 .4 1.5 8.4 1.2

Animal (68) (255) (323)

Suicide 0 .7 0 .1 0 .8 0 .1 0.8 0.1

Attempt (11) (18) (29)

Accidental 3.6 0 .7 9.0 1.2 6.7 1.0

(56) (201) (257)

Other 4.1 0.7 3.7 0.5 3 .8 0.6

(63) (83) (146)

Totals 40 .8 18.4 59.2 12.9 100.0 14.7

(1562) (2265) (1827)•
NOTE: Chi-square= 318.82; p = 0.001 ; v = 0.28. Subcell percentages may not tota l 100.0 due to rounding.
r-:> • All column totals include 18 pre-T .O.P. 237 and 23 post-T.O.P. 237 cases in which reason for shooting was not
,....
-.::) reported.
~ TABLE 3 OFFICER-REPORTED REASON FOR SHOOTING, EXCLUDING WARNING SHOTS
-::)
~
Pre- Weekly Post- Weekly Weekly
Reason T.O.P. 237 Mean T.O .P. 237 Mean Totals Mean

Defense 72.7% 10.6 70.7% 8.7 71 .5% 9.0

of Life (902) (1536) (2438)

Prevent or 13.9 2.0 4.3 0.5 7.8 1.0

Terminate (172) (93) (265)
Crime

Destroy 5.5 0.8 11 .7 1.4 9.4 1.2

Animal (68) (254) (322)

Suicide 0.9 0.1 0.8 0.1 0.8 0.1

Attempt (11) (18) (29)

Accidental 4.5 0.7 9.2 1.2 7.5 1.0

(56) (201) (257)

Other 2.5 0.4 3 .3 0.4 3.0 0.4

(31) (71) (102)

Totals 36.3 14.6 63.7 12.3 100.0 13.1"

(1240) (2173) (3413)
NOTE: Chi-square = 102.62; p = 0.001; v = 0.17.
• Total weekly mean includes 36 cases in which a reason for shooting was not reported.
reported firing only warning shots, was calculated. This table
presents some important differences from the data in Table 2.
Its total (3,413 shooters) indicates that 414 of the shooters re­
ported in the previous table had fired only warning shots. Fur­
ther, most of the reported warning shots (304 of the 414) took
place during the period preceding T.O.P. 237. Consequently,
the weekly mean number of officers who reported firing "in
defense of life" during the early period has declined to 10.6 and
the weekly mean "prevent/terminate crime" shooters is almost
halved from 3.9 to 2.0). While this table's exclusion of "warning
shots" still leaves significant pre- post-T.O.P. 237 differences in
the mean weekly frequencies of reported reasons for shootings,
the differences shrink considerably; indeed, the relative per­
centage of reported "defense of life" shootings decreases. 6 In
terms of weekly shots fired at people, however, this table also
demonstrates significant decreases following T.O.P. 237.
T.O.P. 237 and Firearms Discharge/Assault Generated
Injury
In measuring the impact of T.O.P. 237 on officer injury, we
chose to include in our analysis only injuries and deaths sus­
tained in the line of duty. In this manner, we eliminated from
consideration injuries on which one might reasonably expect
that T.O.P. 237 would have no impact (e.g., officer suicides,
injuries accidentally suffered while handling or cleaning weap­
ons. This does not, however, limit the analysis to injuries sus­
tained by on-duty officers. Off-duty officers hurt while "taking
police action" (e.g., while attempting arrests, stopping and
questioning suspicious persons) are defined by the department
to have been injured in the line of duty. Thus, for example, an
off-duty officer who is shot while attempting to apprehend the
perpetrators of a robbery that the officer witnesses is deemed
to have been injured only because of his or her status and
actions as a police officer. The injury was sustained in response
to "duty's call." If, conversely, whether the officer is off or on
duty when injured in circumstances that do not involve the
performance of the police function or that involve negligence on

273
~
-:)
~

TABLE 4 REPORTED LINE-OF-DUTY INJURIES SUSTAINED BY OFFICERS IN FIREARMS

DISCHARGE/ASSAULT INCIDENTS•

Officer Pre- Weekly Post- Weekly Weekly

Injury T.O .P. 237 Mean T.O.P. 237 Mean Totals Mean

Injured 95.9% 4.4 96.9% 2.5 96.4% 3.1

(376) (438) (814)

Killed 4.1 0.2 3 .1 0 .1 3.6 0.1

(16) (14) (30)

Totals 46.4 4.6 53.6 2.6 100.0 3.2

(392) (452) (844)

NOTE: Chi-square= 0.14;p = 0.70;0 = 0.14.

• Excludes incidents not involving confrontations with human opponents (e.g., destroying injured animal) .
the officer's part, the injury is defined as non-line of duty. As a
case in point, an officer stabbed in a street robbery by individ­
uals who do not know that he or she is a police officer is re­
garded as having been injured as would any citizen in similar
circumstances. The injury, therefore, is classified as non-line of
duty. Similarly, an on-duty officer who is injured by, for exam­
ple, the accidental discharge of a carelessly handled weapon
while not engaged in a specific police action, is also recorded as
having suffered a non-line of duty injury.
Table 4, which presents comparative frequencies of FDAR­
generated, officer-line-of-duty injury and death before and after
T.O.P. 237, demonstrates that the injury vs. death percentages
remained fairly constant across the two periods. Once again,
however, one finds that weekly means for injured and killed
officers vary considerably. Before T.O.P. 237, 4.4 officers a
week suffered nonfatal FDAR-generated, line-of-duty injuries:
after T.O.P. 237, the figure declines to 2.5. Similarly, the fre­
quency with which officers are killed in the line of duty drops
from 0.2 (one every five weeks) to 0.1 (one every ten weeks).
Table 5, which presents frequencies of known shooting inci­
dents opponent degree of injury for the pre- and post-T.O.P.
237 periods, shows reductions in these frequencies as well.
Again, one finds that the relative chances of opponent injury
have remained fairly constant across the five years studied. Of
opponents whose degree of injury was known to the police,
approximately seventy percent suffered no injury, slightly more
than twenty percent were wounded and just over nine percent
were killed during both periods.
But in examining the weekly means, we again find consid­
erable reductions in all our categories following T.O.P. 237.
Most specifically, Table 5 shows us that New York City police
wounded a weekly average of 3.9 opponents and that they
killed 1.6 prior to T.O.P. 237; during the period between that
intervention and December 31, 1975, these means fell to 2.3
and 1.0 respectively.
While both tables 4 and 5 present rather striking differ­
ences between the pre- and post-T.O.P. 237 periods, several

275
~
-::)
~

TABLE 5 KNOWN INJURIES SUSTAINED BY OPPONENTS IN POLICE SHOOTING INCIDENTS

Opponent Pre- Weekly Post- Weekly Weekly

Injury T.O.P. 237 Mean T.O .P. 237 Mean Totals Mean

None 69 .0% 12.3 70.9% 8 .2 70 .1% 9 .5

(1043) (1435) (2478)

Injured 21 .8 3.9 20.0 2.3 20 .8 2.8

(329) (405) (734}

Killed 9.2 1.6 9.1 1.0 9 .1 1.2

(139) (184) (323)

Totals 42.7 17.8 57.3 11.5 100.0 13.5

{1511) (2024) (3535}

NOTE: Chi-square = 0.03; p = 0.99; v = 0.01 .

considerations prevent our concluding that these differences
are entirely attributable to that directive and they limit the
degree to which equivalent reductions might be predicted for
other agencies. Simply stated, the reduced frequencies that are
presented in both tables are based on two observations made
over a five-year period, during which other variables might
have been expected to influence police and citizen injuries.
Those other variables (e.g., injuries resulting from the unpro­
voked police assassination attempts by the radical "Black Lib­
eration Army," which were most frequent during the pre­
T.O.P. 237 period; police personnel deployment), however, are
beyond the scope of the present inquiry. 7 We elected, therefore,
to more clearly identify the effect of T.O.P. 237 on police and
citizen injury by increasing the number of observations in our
examination of those phenomena during the five years studied.
Figure 2 contrasts the bimonthly frequency of civilian
FDAR-generated deaths and injuries (classified as one entity on
the premise that degree of injury is subject to chance varia­
tions) with the frequency with which officers were shot or
stabbed in the line of duty (including only what are generally
the most serious injuries). This figure demonstrates a general
decrease in both these phenomena between 1971 and 1975. In
addition, the figure reveals that the relationship between citi­
zen injuries and officer shot or stabbed frequencies is subject to
considerable variation. In some periods, citizen injuries far out­
weigh officer shot or stabbed frequencies, while in others, the
gap is closed considerably: during May and June, 1973, officer
injuries even exceeded citizen injuries. Again, except for these
periods, and despite the gross incident and injury reductions
cited earlier, the relative risk of FDAR-generated officer of
citizen injury or death looks relatively stable.

CONCLUSIONS
Our examination has demonstrated that a considerable re­
duction in the frequency of police shooting accompanied New
York City's direct intervention on the firearms discretion of its
police officers. Further, our data indicate that this reduction

277
!:'-:)
"'1
00

FIGURE 2. POLICE OPPONENT INJURIES AND DEATHS AND POLICE OFFICERS SHOT
OR STABBED IN THE LINE OF DUTY (N = 30).
T.O .P. 237 1.0. 118

/\
I \ 1\
/
/'" \
\
\.--, I
I
/ \
\ ,,,\
\I ,., ~
" I \

'\ I \\ t'\~l' \
' , ....,1I \'
"'
'\I \
\ \.--..}I
I
I
\
\
\ ' '/ \" , ; ok~~~~nents injured/

officers shot/stabbed
in line of duty
1971 1972 1973 1974 1975
was greatest among the most controversial shooting incidents:
shootings to prevent or terminate crimes, which frequently in­
volve police shots at fleeing felons. To the extent that this New
York experience may be generalized to other agencies, there­
fore, an obvious consequence of the implementation of clear
shooting guidelines and their stringent enforcement is a reduc­
tion of injuries and deaths sustained by suspects who would
face far Jess severe penalties even if convicted after trial.
Of equal significance to the police administrator is the fact
that these shooting decreases were not accompanied by in­
creased officer injury or death . Conversely, since both these
phenomena appear to be associated with the frequency of shoot­
ing incidents and related citizen injury, both declined pursuant
to T.O.P. 237.
In the most simple terms, therefore, the New York City
experience indicates that considerable reductions in police
shooting and both officer and citizen injury and death are asso­
ciated with the establishment of clearly delineated guidelines
and procedures for the review of officer shooting discretion.

Notes
This paper is a revision of a paper presented a t the Annual Meeting of
the American Society of Criminology, Atlanta, Georgia, November,
1977.
1. Kobler (1975:164) reports that only three of the 1,500 police shoot­
ers he studied were criminally punished for their action.
2. It is probable that officers who do violate statutory provisions (e.g.,
b:' shooting at fleeing misdmeanants) but who both miss and fail to
apprehend their targets do not often come to court attention.
3. Firearms Discharge Review Board cases disposed of with recom­
mend ations that include more than one of these categories most typi­
cally involve shootings deemed to have violated departmental guide­
lin es and to have indicated that transfers from sensitive (and often
desirable) assignments are appropriate. Narcotics officers and other
plainclothes personnel who shoot in violation of departmental" policy,
for example, are often recomme nded for transfers to administrative
assignments or to patrol duty in outlying areas where the chances of
encountering circumstances provocative of weapons use are slight. In

279
such cases, we considered the board's recommendation for disciplinary
action to be its major and most severe finding and coded dispositions
under that single heading. Similarly, because we regarded criminal
charges to be the major and most severe finding in all cases in which
they resulted, we included such cases in that single category, although
they also invariably involved some departmental disciplinary action.
Cases involving criminal charges are held in abeyance by FDRB pend­
ing court disposition: officers convicted of criminal use of firearms are
then summarily dismissed from the department: all officers acquitted
of such criminal charges during the period we studied were subse­
quently subjected to departmental trials for their actions.
By the time data collection for this report ceased (August, 1976),
the Firearms Discharge Review Board had adjudicated 2,155 shoot­
ings. Their major recommendations in these cases were as follows:
Within law and department guidelines 70.8% (1525)
Retrained in law or tactics 18.3 (395)
Disciplinary action 7.7 (167)
Criminal charges 1.2 (26)
Transfer 0.6 (13)
Psychological or alcoholic counseling 1.3 (29)
Total 100.0 (2155)
4. This attempt was not always successful. In the absence of informa­
tion to the contrary (e.g., FDRB or judicial findings), we accepted
"officer reported reason for shooting" at face value. This exception,
however, affects only our examination of changes in the nature of
shootings. It has little or no significance as far as total shooting fre­
quencies are concerned.
5. 1.0. 118's only new discretionary parameter was the statement that
"(t)he discharge of a firearm at dogs or other animals should be an
action employed ONLY when no other means to bring the animal
under control exists."
6. A good part of this percentage decrease, however, is due to the
great increase in frequency in reported "destroy animal" shootings.
Had these remained constant at their pre-T.O.P. 237 level (0.8 per
week), the post-T.O.P. 237 "defense of life shots at people" percent­
age would have been 74.6 percent.
7. We conducted detailed examinations of the effects of several other
variables on shooting frequency, type, and consequences in Fyfe (1978)
and found that they did not significantly alter the findings reported in
this study.

280
 References

American Bar Association Project on Standards for Criminal Justice

{1973). Standards relating to the urban police function. New York:
American Bar Association.
Berkley, G.E. {1969). The democratic policeman. Boston: Beacon
Press.
Fyfe, J.J. (1978). Shots fired: An examination ofNew York City police
firearms discharges. Ph.D. dissertation. State University of New York
at Albany.
Harding, R.W., and Fahey, R.P. (1973). Killings by Chicago police,
1969-70: An empirical study. Southern California law review, 6:284­
315.
Kobler, R.M. (1975). Police homicide in a democracy. Journal ofsocial
issues, 31:163-181.
McKiernan. R.M. (1973). Police shotguns: Devastating to the animals.
The New York Times, February 7:35.

Milton, C.H.; Halleck. J.W.; Lardner, J.; and Abrecht, G.L. (1977).

Police use of deadly force. Washington, D.C.; Police Foundation.

National Advisory Commission on Criminal Justice Standards and

Goals (1973). Police. Washington, DC: U.S. Government Printing Of­
fice.
New York City Police Department (1973), Interim order 118.
_ _, (1972). Temporary Operating Procedure 237.
_ _, Crime Analysis Unit (1971-1975). Monthly arrest report. De­
cember.
Nie, N.H.; Hull, C.H.; Jenkins, J.G.; Steinbrenner, K.; and Bent. D.H.
(1975). Statistical package for the social sciences. New York: McGraw­
Hill.
President's Commission on Law Enforcement and Administration of
Justice (1967). Task force report: The police. Washington, DC: U.S.
Government Printing Office.
Rhine, B. (1968). Kill or be killed: Use of deadly force in the riot
situation. 56 California law review: 829.
Rubinstein, J. (1973). City police. New York: Farrar, Strauss and
Giroux.
Wilson, J.V. (1972). Deadly force. Police chief December.

281
 CHAPTER 12

ALWAYS PREPARED: POLICE OFF-DUTY

GUNS

JAMES J. FYFE

ABSTRACT

Even while off duty, American police are expected to

be armed and to actively intervene in situations threat­
ening to life, property, or order. This article considers
the assumptions upon which that expectation is based
and suggests that they may be ill-founded. Research to
determme whether armed off-duty police actually in­
crease community violence levels is recommended as a
prerequisite to an informed reconsideration of the ap­
propriate role for off-duty police officers.

POLICE OFF-DUTY GUNS

Whether or not they are technically on duty, American po­
lice almost everywhere are expected to be armed and ready for
action. As a 1978 survey of 49 major police departments found,
25 departments-51 percent-reported permitting officers to
carry off-duty guns, and 24-49 percent-reported requiring
that officers be armed off duty. 1 None required officers to leave
their guns at police stations at the completion of the working
day. If it is safe to generalize from these departments to the
rest of America's 450,000 police, it is likely that at any time
there are about in the country approximately 300,000 armed
off-duty police.
Because the "twenty-four hour cop" is an American tradi­
tion, the presence of these 300,000 off-duty guns is not surpris­
ing. Nor is it surprising that there is so little debate over the
SOURCE: "Always Prepared: Police Off.Duty Guns," The Annals of the A meri­
can Academy ofPolitical and Social Science (November, 1980), 452-72-81. Re­
printed with the permission of the American Academy of Political and Social
Science.

282
merits of adding these weapons to the great number in circula­
tion among citizens. Police, unlike most handgun owners, are
psychologically screened and tested, trained, and sworn to pro­
tect life and property. Police possess weapons to protect public
interests, whereas citizens possess weapons to protect their
own private interests. Thus is likely that even handgun control
advocates, if they have thought about the issue at all, regard
police off-duty guns as a category of weapons separate and
distinct from those in the hands of private citizens.
Despite this apparently reasonable distinction, and despite
the fact that 300,000 off-duty police guns are a drop in the
bucket of the estimated 40 million handguns in America,2 the
issue of off-duty weapons raises a number of important ques­
tions. While the benefits are unclear, the costs are potentially
substantial. Over one in ten of the officers killed by felons na­
tionwide from 1972 to 1978 were off duty at the time, 3 and
almost one in four of the 239 officers killed by felons in the New
York City Police Department since 1844 were off duty.4 Off­
duty guns are also important in police killings of citizens; sev­
eral studies have found from 12 to 17 percent of homicides by
police to occur while the officers are off duty. 5

WHY ARM OFF-DUTY POLICE?

As suggested previously, the major rationale for arming

off-duty officers is the notion that police professional responsi­
bilities should not be constrained by their scheduled working
hours. In many jurisdictions, off-duty police may be disciplined
or prosecuted for failing to respond in an "appropriate" man­
ner to situations threatening to life, property, or order. 6
Thus the justification for arming off-duty police varies with
the interpretation of this responsibility. If it is " appropriate"
for off-duty officers to passively make observations of crimes in
progress and to relay information to on-duty colleagues, one
might argue that there exists little justification for off-duty
guns. If, on the other hand, off-duty officers are expected to

283
intervene actively in circumstances which threaten life, prop­
erty, or order, ther exists greater justification for equipping
them with firearms.
A second rationale for arming off-duty police is that of
deterrence. Here one would argue that justification for arming
off-duty officers increases with the degree to which potential
offenders are deterred from criminal behavior by the actual or
feared presence of armed off-duty police.
A related consideration is that of officer safety. A conse­
quence of police work is that some of those questioned, ordered
to move on, ticketed , investigated, or arrested may wish to get
even with officers or to impede cases against themselves. Dis­
arming off-duty police, it may be argued, might increase both
the temptation and the opportunity for their disgruntled clien­
tele to do so. Thus there are unanswered questions about the
extent to which off-duty guns ensure the safety of officers from
those seeking revenge or forcible prevention of damaging testi­
mony. Further, disarming off-duty officers may also affect their
willingness to engage forcefully, but properly, in on-duty en­
forcement or order-maintenance activities. Why make enemies
at work when one may be vunerable to them after work?

CONSEQUENCES OF ARMING OFF-DUTY POLICE

The best way to test the validity of these justifications for
arming off-duty police is to examine them and to weigh them
against negative consequences of providing police with off-duty
guns. This is difficult for several reasons. It is probable, for
example, that the bars widely known to be frequented by off.
duty officers are held up less often than bars remote from po­
lice stations. Beyond that observation , however, it is nearly
impossible to say much about the deterrent value of off-duty
guns because no attempt to measure it systematically has been
undertaken.
In addition, the data that are available on more concrete
events are frequently rather one-sided. Police often point with
pride to incidents in which armed off-duty officers have bravely
and honorably protected the public interest. Less frequently do

284
they attempt to publicize cases in which off-duty officers have
used their guns unwisely or dishonorably. As a result, little is
known about the negative consequences of arming off-duty po­
lice. There is little knowledge of the degree to which off-duty
guns, which ostensibly serve to decrease public violence, also
serve to increase or escalate violence. It would be useful to
know how often police off-duty guns are deliberately misused or
are the instruments of tragic accidents. It would also be useful
to know how often armed off-duty police make bad situations
worse by well-intended but ill-advised interventions in threats
to life, property, or order.
The question of consequences is extremely important. The
measurement of consequences should, in fact, serve as the basis
for any definition of the appropriate role of off-duty officers. As
Safer notes, "proof of frequent misuse of weapons" might lead
to the conclusion that "disarmament of off-duty officers would
diminish the community's level of violence." 7 Disarmament
would also necessarily result in limiting the "appropriate" off­
duty police role to that of passive observer and information
transmitter. Such a definition would be a major police reform
because, as the survey of off-duty weapons policy suggests, the
prevailing definition of the appropriate off-duty police role is
that of active intervenor.

REDEFINING THE ROLE OF OFF-DUTY POLICE

Police agencies, like all bureaucracies, are conservative in­
stitutions. Even given evidence of frequent misuse of off-duty
guns, it is likely that attempts to disarm off-duty police and to
redefine the off-duty role would be strongly resisted. Police
chiefs- among whom are some advocates of strict public hand­
gun controls-might argue that, despite a few freak accidents
and some regrettable incidents, off-duty guns are a major con­
tributor to public safety and to society's ability to protect itself
against criminality. They might also point out that a precise
assessment of the positive consequences of off-duty guns was
precluded by the absence of data. Further, it could be argued,
attempting to complete an off-duty gun cost-benefit equation by

285
 acquiring those data would require experimental conditions­
disarming off-duty police-that might prove too costly to soci­
ety and to officers themselves. Finally, it might be pointed out
that since police are paid only for their formal working hours,
the financial costs of the extra protection of off-duty guns are
negligible.
In these justifications for off-duty guns, a cynic might find
the suggestion of a hidden agenda. The economic argument
may be a rationalization offered by police administrators unwill­
ing to face up to the question of off-duty guns for reasons more
closely related to organizational labor tranquility than to public
safety. Police unions and labor groups exert considerable influ­
ence over police managerial decisions. Police unions often base
their salary demands in part on the public expectation that
their members will keep and bear arms 24 hours a day. Their
logic generally runs as follows:
Police deserve more money than other civil servants for
two reasons. First, their duties are uniquely demandin_g
and dangerous: they alone are expected to put their
lives on the line agamst society' s enemies.
Second, they alone are expected to fulfill their job du­
ties, whether or not they are technically working: they
alone are expected to be equipped With the tools of
their trade at all times. SanitatiOn workers are not ex­
pected to clean dirty streets once they sign out of work ,
but off-duty police are expected to help keep the streets
free of crooks. Fire fighters do not carry off-duty axes,
but police carry off-duty guns.
Further, no other workers have to worry constantly
about safeguarding lethal weapons from their curious
children. To know how much a police officer is worth,
you have g-ot to look at the number of officers killed
taking pohce action off duty. While you are doing that
take a look at the little kids who have found daddy's
gun in the closet and killed themselves while playing
with it.
These are potent arguments which make it clear that police
unions do not regard off-duty guns as cost-free protection. Con­
versely, unions regard their responsibility and willingness to be

286
armed off duty as a blue chip at contract negotiations. It is also
likely, therefore, that they would strongly resist efforts to ei­
ther disarm off-duty police or to limit the off-duty role to that
of observer and information transmitter. Further, such efforts
can easily be interpreted as ''handcuffing the police and reduc­
ing police protection." Thus it is probable that chiefs would
choose not to be publicly identified with these efforts unless
there existed an extremely unambiguous and convincing case
that they are in the public interest.
The absence of much of the information upon which such a
case might be made reinforces the suggestion that administra­
tors are not anxious to deal with the issue . Data related to the
number of lives saved and the number of important arrests
effected by off-duty police, for example, would be relatively
easy to collect, but little or no comprehensive information on
these events is at hand. As noted previously, we also know little
about the deterrent effects of off-duty guns, a subject about
which data are more difficult, but not impossible, to collect and
analyze.
Despite these omissions, it is possible to make a preliminary
assessment of the consequences of arming off-duty police. Such
an assessment suggests that it might be appropriate to fill the
information voids cited previously and to conduct more com­
plete and detailed evaluations of the merits of police off-duty
guns.

WHAT DO POLICE DO WITH OFF-DUTY GUNS?

Police in New York City are required to carry off-duty guns8
while within the limits of the city and may do so at their option
outside the city. Between 1971 and 1975, 681 New York City
police officers reported discharging their firearms while off
duty. 9 Most of these events occurred for apparently meritorious
reasons, but others did not. Table 1 shows that roughly seven in
ten of these shootings reportedly involved the defense of life or
prevention of crime. If we add to those the number of incidents
in which officers destroyed dangerous or hopelessly injured ani­
mals (3.2 percent), we find that about three quarters of the

287
events ostensibly involve some law enforcement or order main­
tenance function .
TABLE 1 OFF-DUTY NEW YORK CITY POLICE REASONS FOR
SHOOTING, 1971 -75

Reasons for Shooting Percent Number

Self-defense 55.8 380

Defense of others 2.5 17
Prevention or termination of crime 13.1 89
Destruction of animal 3.2 22
Suicide attempt 3.8 26
Accident 12.6 86
Other/not ascertained 9.0 61
Total 100.0 681

What about the other quarter? Officers used their guns

against themselves often enough to make "suicide by gun" the
largest single cause of violent officer death and an event nearly
as frequent as line of duty deaths during the years included in
the table. During 1971-75, 30 New York City officers died at
the hands of others in the line of duty; 25 off-duty officers shot
and killed themselves, and one other attempted to do so but
sustained a nonfatal wound. 10 Almost half of the "nonenforce­
ment" shooters-12.6 percent of the total-reportedly fired
their guns accidentally. In doing so, they shot themselves, fam­
ily members, friends, other officers, and total strangers, for
example:
Patrolman Hines and his wife were in the kitchen ...
Officer removed his off-duty gun from the top of a six
foot china closet, noticed that the holster strap was
loose and pushed the gun firmly in the holster discharg­
ing the gun. Mrs. Hines was across the room and the
bullet struck her above the right ear. She was pro­
nounced D.O.A . by Dr. Thompson.
In another case which appears to have been accidental , the
officer was found lying dead on the floor in front of a
bureau and his service revolver with one shot fired was
288
 found in an open top drawer. Apparently Ptl. Adams
was standing at a right angle in front of the bureau and
discharged one shot from service revolver. Interroga­
tion of parents, in-laws, and wife of officer indicated no
medical history or other known reason for Ptl. Adams
to have willfully committed this act.
A substantial percentage of off-duty shooters-nine percent
of the total-fired their guns for "other" reasons or for reasons
that are not readily articulable. One officer, for example, shot
and killed his wife and his mother-in-law; in other cases:
While apparently intoxicated, Ptl. Jones was
ejected from a St. Patrick's Day party in a bar and grill
and fired six shots from his revolver into the premises,
wounding two persons .
. . . [Officer who shot his wife once in the head and
three times in the backl stated that he had been mar­
ried for seven years and had various problems with his
wife; their problems focused on mother-in-law trouble,
his wife's spending- of money, and his studying for the
sergeant's promotiOnal examination.
After a dispute with my wife, I went outside and
fired 12 shots into a tree behind my house in order to
vent my frustrations.
Various explanations for such bizarre violence may be
found in the growing body of literature on police occupational
stress which generally concludes that the police career places
extreme psychological pressures on its incumbents and their
families. 11 Sadly, in the previous cases, it also placed in the
officers' hands the means to deal with those dilemmas in a
destructive manner. Thus it is apparent that some officers use
their off-duty guns for purposes other than those for which they
were intended. 12
Less apparent is the number of unwise or imprudent shoot­
ings concealed within Table 1's more seemingly admirable r ea­
sons for shooting. A suggestion of their numbers appears in
Table 2, which describes the departmental adjudications of off­
duty shootings that occurred after 18 August 1972, when a
formal shooting review and adjudication process was estab­
lished. 13

289
TABLE 2 NEW YORK CITY POLICE DEPARTMENT
ADJUDICATIONS OF OFF-DUTY SHOOTINGS, 18
AUGUST 1972-31 DECEMBER 1975

Adjudications Percent Number

Shooting justifiable 58.5 224

Officer retrained in law and/or
tactics 15.1 58
Officer disciplined 16.7 64
Officer arrested 5.5 21
Officer transferred 0.3 1
Officer referred to psychological
and/or alcoholic counseling 3.9 15
Total 100.0 383

The table shows that the department itself found that more
than four in ten shootings warranted some administrative or
criminal sanction: 15.1 percent of the officers were retrained in
law and/or tactics; 14 16.7 percent were disciplined by the depart­
ment; 5.5 percent were arrested; and the remainder were
transferred to other assignments or referred to the depart­
ment's psychologists or alcohol abuse counselors. Thus while
three quarters of the shootings between 1971 and 1975 involved
some law enforcement or order maintenance activity, only
slightly more than half of the shootings adjudicated during
most of those years were found to have been beyond official
reproach.
Even among those found formally justifiable, however,
questions of prudence and impact on violence may be raised .
More than one third of the shootings were immediately precipi­
tated by robberies. A review of those incident reports causes
one to speculate on how appropriate it is for lone off-duty offi­
cers to attempt to intervene in such events. One example in­
volved an officer dining in a restaurant when two males en­
tered and announced a stick-up.
I was seated at a table in the dining area. I was
ordered to get over to the bar. I leaped from my chair
saying "police" and fired my revolver two times at (sus­
pects].

290
 After returning the officer's fire, both suspects escaped.
Since the robbery was averted and nobody was injured, this
unquestionably courageous intervention was not a total failure.
On the other hand, it did escalate violence considerably and did
create a great risk to the lives of the officer and other patrons,
as well as to the robbers. It takes little imagination to conceive
of possible far less happy endings to this event. The officer
could not have been certain that on-duty police had not been
summoned to the robbery . Thus he ran the risk of being mis­
taken for a robber by other officers, a scenario which has also
occurred and ended tragically. 15 Indeed, once this officer inter­
vened , literally anything could have happened. Eight months
later, in fact, this same officer was again dining in a restaurant
and was confronted by two armed robbers. On that occasion,
the exchange of gunfire left the officer and both robbers dead.
The most immediate question raised by such events is what
would have happened had the officer refrained from interven­
ing and me rely observed and transmitted information. Both
sets of robbers would almost certainly have escaped with the
proceeds of their endeavors. Although it is not likely , both sets
of robbers might have been apprehended later as a result of
information and descriptions provided by the officer. 16 Most im­
portant, the officer would almost certainly have survived both
robberies.
This last observation will raise police eyebrows. A firmly
held belief among police is that it is advisable for officers to
take forcible action when confronted by situations like those
described previously. This is so because it is believed that rob­
bers who discover a police officer among their victims are likely
to kill him simply because he is a police officer. Thus, it is
better to resist than to passively submit and risk the possibility
of discovery and execution. 17 This belief has powerful implica­
tions. The officer operating under it perceives the gun duel
with robbers as not merely appropriate, but as imperative.
Logic suggests, however, that this last belief is ill-founded .
Holdup men commit their crimes for pe rsonal gain. Their goal ,
apparently, is to survive their acts and to quietl y enjoy their

291
earnings. Why, then, would they wantonly execute police offi­
cers and gratuitously become the subject of manhunts for
"coldblooded cop killers?"
This logic is supported by experience. This New York City
data include literally dozens of cases in which armed off-duty
police responded to the sudden appearance of robbers' over­
whelming firepower by submitting passively. Many were dis­
covered to be police and were relieved of their guns and
badges. None was wounded or killed merely "because he was a
cop," 18 but many were wounded or killed while resisting.
The apparent reluctance of criminals to kill police officers
also suggests that police fear of assaults motivated by desire
for revenge or prevention of testimony may be unfounded. One
who contemplates such an act does so knowing that records of
his prior encounters with his potential victim will make him a
prime suspect. He also knows that his offense will be investi­
gated, prosecuted, and penalized more intensely than any other
he may have committed. Such attacks to impede justice, then,
appear quite unlikely regardless of whether or not police are
armed off duty. Further, the relative infrequency of such at­
tacks on citizen witnesses, jurors, prosecutors, and judges rein­
forces the notion that off-duty guns make little difference
where violence of this nature is concerned .

CONCLUSION
The classic and often derided conclusion of social scientists
is to recommend further research. In the present case, how­
ever, it is genuinely appropriate to conclude by recommending
a test of the assumption that armed off-duty police contribute
to the public good.
American police are citizens and police officers. Considera­
ble effort has been expended to eliminate distinctions between
them and the communities they serve. Some distinctions, how­
ever, are both desirable and necessary and thus are not subject
to these efforts. It is desirable and necessary that on-duty po­
lice fulfill the role of active intervenor in threateni:1g situations.

292
 It is also necessary, therefore, that they be distinguishable from
most citizens by being armed during that time.
It is less clear that it is desirable and necessary for police to
attempt to fulfill the active intervenor role while off duty. To do
so, they must continue to be armed and therefore distinguish­
able from other citizens. Thus it may be difficult for police to
relate their own life experiences to those of the unarmed citi­
zens for whom they work. Further, this distinction may be Jess
than desirable because police may be even more vulnerable than
most citizens to the forces that lead to gun abuse.
Before police practice their craft, they are usually screened,
tested, and trained. While they practice their craft, they are
subject to stresses far more psychologically demanding than are
most citizens. The combination of these job stresses and ready
access to off-duty guns sometimes ends tragically. Many laud­
able efforts to avert such tragedies have focused on eliminating
or neutralizing the stresses which precede them, but little or no
attention has been given to the desirability or necessity of ac­
cess to the off-duty guns which complete the tragedies.
The question of the desirability or necessity of off-duty
guns does not involve only intentional and accidental misuse.
When off-duty police do use their guns in well-intended inter­
ventions, it is not at all clear that they reduce violence. Con­
versely their actions in threatening situations may even create
actual violence where only potential violence exists.
This negative effect is often due to important qualitative
distinctions between the situations in which on-duty and off­
duty police typically intervene. On-duty police are usually ad­
vised by their radio dispatchers of potential violence. Since this
usually occurs while they are at a distance from the scene, they
have the opportunity to plan their approach to it and to coordi­
nate their efforts with colleagues. On-duty police are also usu­
ally in uniform and are clearly identifiable to other officers.
Off-duty police who intervene in potential violence rarely
enjoy such luxuries. They are typically not given any warning
of impending events but rather, are suddenly confronted by
suspects whose guns are already drawn. Off-duty police are not

293
typically in the company of colleagues, but are alone or with
friends or family. They do not usually have instant access to
police communications systems. They are usually in civilian
clothes and are thus easily mistaken for armed suspects by
arriving police. Finally, they are far more likely than on-duty
officers to have judgment and reflexes dulled by liquor.
In such circumstances, it is rarely desirable or necessary
that off-duty police distinguish themselves from other citizens
by attempting to actively intervene nor is it fair to them or
other citizens. Indeed, it may be most fair to require off-duty
police to leave their guns in their lockers with the rest of their
uniforms.

Notes
1. John F. Heaphy, ed., Police Practices: The General Administrative
Survey (Washington, DC: Police Foundation, 1978), item 20.
2. Matthew G. Yeager, Joseph Alviani, and Nancy Loving, How Well
Does the Handgun Protect You and Your F amily? Technical Report:
Two (Washington, DC: United States Conference of Mayors, 1976),
p.l.
3. Mona Margarita, Criminal Violence Against Police in the United
States (Washington, D.C.: U.S. Bureau of Justice Statistics, forthcom ­
ing).
4. Mona Margarita, "Criminal Violence Against Police" (unpublished
Ph.D. dissertation, School of Criminal Justice, State Unive rsity of
New York at Albany, 1980), p.83.
5. Lawrence W. Sherman, "Homicide by Police Officers: Social
Forces and Public Policy" (grant proposal submitted to the Center for
Studies in Crime and Delinquency, National Institute of Mental
Health, (1977), p.3.
6. See, for example, J. Gerald Safer, "Deadly Weapons in the Hands
of Police Officers, On Duty and Off Duty, " J. Urban Law, 49(3):577.
7. Ibid., p.579.
8. New York City police may use as "off-duty guns" either the regula­
tion four-inch barrel .38 caliber service revolver or a lighter and more

294
compact two-inch barrel .38 caliber revolver. The former must be pur­
chased by officers, and the latter is an option chosen by most for
reasons of convenience. Some officers, however, carry their service
revolvers both on and off duty.
9. For access to these data, I am indebted to former Chief of Person­
nel Neil J. Behan, former Assistant Chief Patrick Fitzsimons, and
Lieutenant Frank McGee of the New York City Police Department.
10. Three other officers committed suicide by gun during working
hours.
11. A Law Enforcement Assistance Administration bibliography, J.T.
Skip Duncan, Robert N. Brenner, and Marjorie Kravitz, Police Stress:
A Selected Bibliography (Washington, D.C.: U.S. Government Printing
Office, 1979), lists 112 articles, books, and monographs and 33 films.
These cite as factor s causing stress such police job characteristics as
danger, unpredictability, boredom, the need to repress emotions, ro­
tating shift work, quasi-military organizational structure and disci­
pline, lack of lateral transfer opportunities, inadequate career develop­
ment , frustration about court appearances, concern over criticism, and
a police image distorted by the media. It cites as effects of stress
disproportionate police rates of suicide, alcoholism, heart disease , gas­
trointestinal disorders, and marital, family, and emotional problems.
12. Neither is such abuse of off-duty guns confined to the officers
themselves. Included in the data examined in this study, for example,
are nine incidents in which officers' off-duty guns were used against
them by wives and female friends. Five officers were wounded and
four were killed in these shootings.
13. New York City Police Department, "Temporary Ope rating Proce­
dure #237," (18 Aug. 1972).
14. A finding that retraining is appropriate most often reflects the
reviewers' opinion that a shooti ng itself was justifiable, but might have
been avoided had the officer handled the events which precipitated it
differently , for example, by calling for help rather than by attempting
to confront a dangerous person alone.
15. Nine of the shootings studied involved "mistaken identity" ex­
changes of shots between uniformed police and plainclothes or off-duty
officers at crime scenes. Three resulted in death.
16. The successful escape of the first set of robbers-who may even
have been the same individual s involved in the second robber y­
argues against this outcome. Had the officer concentrated his efforts
on obtaining good descriptions at the first robbery, however, he might
have improved the possibility of apprehension.

295
17. Joseph Wambaugh's The Onion Field (New York: Delacorte
Press, 1973) is an excellent treatment of a variant of this belief.
18. One robber threatened to do so, but was dissuaded by his accom­
plice who made precisely the arguments cited above. In another case,
an officer was one of several restaurant patrons searched and locked
in a restroom by robbers. His victimizers kept their promise to leave
his badge on the restaurant's bar and to deposit his gun in a nearby
mailbox, so as not "to get him in trouble."

296

 CHAPTER 13

OBSERVATIONS ON POLICE DEADLY FORCE

JAMES J. FYFE

The subject of police use of deadly force 1 has generated

considerable controversy in recent years. Questions related to
deadly force have been raised and analyzed by presidential com­
missions, police practitioners, researchers contracted by police
practitioners, radical criminologists, more traditional academ­
ics, social activists, law reviews, and popular writers.2 Two sali·
ent questions addressed from these widely varying perspectives
are
1. Should the police be permitted to shoot at unarmed and
nonviolent fleeing suspects?
2. To what degree have police agencies insulated them­
selves and their members from criticism, as well as from crimi­
nal or civil liability, by "covering up" incidents in which police
officers have shot others in violation of law or administrative
regulations?

SOURCE: " Observations on Police Deadly Force," Crime and Deli nquency
(July , 1981), 27:3:376-389. Reprinted with the kind permission of the National
Council on Crime and Delinquency.

297
FLEEING SUSPECTS
The power of the police to use deadly force is derived from
the English common law, which authorized the use of deadly
force in defense of life and in order to apprehend persons com­
mitting or fleeing to avoid arrest for a felony. While few would
quarrel with the defense of life rule, many do find fault with
fleeing felon justifications for police use of deadly force. More
specifically, they argue that the rule enabling police to use
deadly force to apprehend any fleeing felon is overly broad,
obsolete, and inconsistent with the philosophies that guide most
of the operations of American criminal justice. 3
Their arguments are based upon changes that have oc­
curred since the fleeing felon rule became part of the common
law . Centuries ago, they note, the fleeing felon rule may have
been defensible because all felonies were capital crimes. We
have long since ceased executing all felons, however. In recent
years, very few persons convicted of even the most heinous
crimes have been put to death. This precipitous drop in execu­
tions has not been accidental: The state's power to execute has
been challenged and carefully reviewed at every level of gov­
ernment. Recognition of the irreversibility of the death penalty,
notwithstanding its questionable deterrent value, requires that
all decisions be carefully made and reviewed, and that defend­
ants be executed only after they have presumably received
every benefit of due process. On the other hand, police are
often permitted to make hasty decisions "at three o'clock in the
morning, in the rain" to take the lives of first-time offenders
suspected of such felonies as burglaries of unoccupied commer­
cial buildings, use of stolen credit cards, and auto larcenies. If
apprehended and convicted, many of these unarmed nonviolent
suspects would receive very minor sentences, especially in
crowded big city courts; certainly, none would be executed.
In the distant past the fleeing felon rule may have
functioned (and been necessary) to protect the lives of unarmed
citizens required to participate in apprehending fugitive s who
knew they faced execution if brought to court and convicted.
Thus, there was considerable overlap between the defense of

298
life and fleeing felon rules. Just as we long ago ceased execut­
ing all felons, however, we long ago ceased requiring unarmed
citizens to apprehend felons by engaging in hand-to-hand com­
bat, if necessary. Police are now solely responsible for the ap­
prehension of felons, and the risk to the armed and trained
police officer pursuing an unarmed thief today is generally far
lower than was the risk facing the unarmed and untrained citi­
zen attempting to bring to trial a suspect to whom arrest and
conviction were literally matters of life and death.
Many legislatures have modified the fleeing felon rule.
Eight states have limited justifiable use of police deadly force
to defense of life situations and to apprehension of persons who
have at least threatened the use of deadly force, or whose con­
duct indicates that delay in apprehension would create a sub­
stantial risk to others of death or serious bodily harm; and ten
states have limited the use of deadly force in arrest to situa­
tions involving persons suspected of crimes of violence . Thirty­
two states operate under the "any" fleeing felon rule, which
authorizes use of deadly force to apprehend even unarmed sus­
pects fleeing from nonviolent felonies. 4
Those who favor allowing police to shoot at such nondan­
gerous fleeing felons base their argument upon grounds of de­
terrence and general law enforcement effectiveness,5 on the
assumption that the power of the police to shoot unarmed flee­
ing nonviolent felony suspects serves as a deterrent to many
who would otherwise commit felony crimes. Although relevant
data on police shootings are scant, there is little empirical sup­
port for this assumption.
If this hypothesis were correct, one might anticipate find­
ing that relatively high police shooting rates would be accompa­
nied by relatively low crime rates, because of the great number
of deterred potential offenders. Gerald Robin, however, reports
that annual rates of justifiable police homicides during the
years 1950-60 in ten American cities varied nearly twenty-fold
(from .40 to 7.06 per 1,000,000 population).6 FBI crime data for
these cities for the same years demonstrate no similar variance
in either direction.

299
 More recent studies of police shooting rates among Ameri­
can cities also show that these rates vary greatly, with little
apparent association with either crime or arrest rates.7 Stated
simply, nothing in the research to date suggests that a high
frequency of police shooting reduces crime rates in any way.
Conversely, given findings that many of the urban civil disor­
ders of the 1960s were precipitated by police shootings,8 it has
been suggested that such shootings have even caused an in­
crease in crime and disruptions of social order.9
Some indication of the reasons for the absence of a clear
association between police shooting and reduced crime rates
may be found in the criminological literature on deterrence.
Donald Newman notes,
General deterrence rests on a pleasure-pain princi­
ple, namely that potential offenders will be prevented
from seeking the pleasures of crime if there is sufficient
assurance that sWift and certain punishment will follow
any violation....
The idea of deterrence is evidently attractive, for it
has existed as a form of social control from the very
earliest times. It persists as operationally viable, even
in the face of evident failure. It rests on a strange mix­
ture of conceptions about the nature of man and human
behavior, in part viewing man as rational, able to calcu­
late gains and losses and in/art treating him as an
animal, able to be conditione to obey. As its rational
basis, the pleasure-pain principle adopts the legal fic­
tion of the reasonable and prudent man . What prudent
human is motivated more by pain than pleasure? Deter­
rence is also seen as a conditioning process resting on
the same principle involved in Pavlov's experiments
with dogs, not only imprinting avoidance on those sub­
ject to painful stimuli, but also conditioning all others
who know about and identify with the punished. It is a
sort of conditioning-once-removed approach.
Some persons can be deterred or conditioned from
at least some deviant acts under ideal laboratory condi­
tions. However, even under controlled conditions the
most serious crimes, the ones most desired to be de­
terred (e.g., murder, assault and sex crimes) are proba­
bly the least amenable to deterrence. In addition, the

300
 real world is not a laboratory, and there are significant
gaps in the speed and certainty of deterrent responses.
Most crimes are not solved, and it is generally known
that most verpetrators get away at least for a long
time. A pohceman cannot be put into every household,
warehouse, or on every street corner. Yet despite its
conceptual weakness and operational difficulties, deter­
rence persists as a major response to the challenge of
crime prevention.
. . . [Deterrence] is probably the most popular of all
crime prevention approaches because it is the cheapest
and most simplistic. Witness the resurgence of death
penalty provisions which almost always rest on deter­
rence claims. Given the horrors of kidnaping and mur­
der, and the usually vague, infinitely complex, and con­
troversial nature of other long-range prevention
proposals, it is easy to see the political attractiveness of
reestablishing the electric chair. No matter that it is
likely to fail in its objective. It represents precise and
immediate action in the face of pressures to do some­
thing about these crimes. 10
In short, we know little about the real effects of deterrence.
What we do know suggests that its effects depend in large
measure on the certainty of punishment. Given the comparative
rarity of police shootings, 11 it would appear that the absence of
the element of certainty is a major limitation upon the deter­
rent effect of police guns. Considering the well-publicized prolif­
eration of handguns and the relative infrequency of on-the­
scene police intervention, it would seem that the would-be
robber, burglar, or thief would be far more likely to be deterred
from his crime by the prospect of meeting armed resistance
from his victim than by the less likely prospect of being shot by
police. Like the electric chair, liberal use of deadly force by the
police is probably better described as a simplistic and politically
attractive approach to crime prevention than as an effective
deterrent.
Neither do the data suggest that the fleeing felon rule as­
sists police in their efforts to apprehend perpetrators of nonvio­
lent crimes. Even though police shootings are far more fre ­
quent in America than in other western countries, they are

301
extremely rare when measured against the number of nonvio­
lent crimes reported in the United States. According to Law­
rence Sherman and Robert Langworthy, American police shot
and killed approximately 700 persons during 1978. 12 If my own
figures on the ratio of woundings to killings resulting from New
York City police shootings 13 can be applied to the rest of the
country, it is likely that an additional 1,400 persons suffered
serious injury from police shootings during that year. Obvi­
ously, not all of these 2,100 shooting victims were unarmed
suspects fleeing nonviolent felonies; but, even were they, those
shot would have represented a very small percentage (.02 per­
cent) of the persons responsible for the 10,079,508 nonviolent
felonies (i.e., burglary, auto larceny, larceny over $50) reported
to American police in 1978. 14 Stated in another way, in order
for the police to have cleared even 1 percent of the nonviolent
felonies reported in 1978 through "apprehensions effected by
shooting," they would have had to increase the rate at which
they shot people during that year by at least fifty-fold. Doing so
would have resulted in approximately 35,000 fatalities and
70,000 woundings. Unless we are prepared to add tremendous
numbers of police to accomplish even this modest clearance ·
rate, and unless, we are prepared to increase dramatically our
various types of medical facilities , and unless we are prepared
to bear the social and moral costs of such carnage, any discus­
sion of the broad fleeing felon rule as an adjunct to police ap­
prehension efforts is futile.

POLICE "COVER-UPS"
Determining exactly what occurred at the scene of a police
shooting is a challenging task. Most shootings occur in low
visibility settings, and rarely are uninvolved witnesses present
at such events. As a result, the source of most of the informa­
tion made available is the police themselves. Since officers are
presumably not motivated to expose themselves to criminal or
civil liability, 15 there often is room for speculation that police
shootings did not occur as reported on official documents.

302
 Skepticism about the accuracy of police shooting reports is
reinforced by occasional violations of what Rodney Stark calls
the police "cult of secrecy." 16 Arthur Kobler documents two
instances in which officers testified that their colleagues had
falsely reported events at police shootings in order to make the
shootings appear justifiable. 17 Skepticism about official versions
of events is also supported by documentation of incidents in
which the motive and methods of those charged with investigat­
ing police shootings have been suspect. Richard Harding and
Richard Fahey studied fatal police shootings in Chicago, and
report finding thirteen shootings in which there was prima fa­
cie evidence of murder or manslaughter; in only two of those
cases were officers actually charged. The major reason for this
variance, they assert, is that police investigators overlooked or
failed to report nonpolice "alternate credible versions" of
shootings. 18 More recently, the Los Angeles Board of Police
Commissioners (a civilian board) overruled an internal depart­
ment finding that a shooting was justifiable under department
policy, and noted that ·

The Commission's review of the Department's in­

vestigation and evaluation of the shooting of Eulia Love
reveals that many of the factors on which the majority
of the Shooting Review Board relied in reaching its
conclusions were based on erroneous or misconstrued
facts. The Board's failure properly to exercise its fact­
finding function, and to obtain and assess all available
evidence, prevented it from giving due consideration to
all elements of Department policies and standards. 19

Social scientists have also expressed skepticism about po­

lice shootings, reporting a cult of secrecy at the administrative
levels of police agencies. Much of the research done on police
shootings has been conducted without the cooperation of the
agencies involved. Kobler, for example, employed as his data
source a newspaper clipping service that supplied him with
journalistic accounts of 1,500 police shootings. Newspaper ac­
counts also served as the basis for a study conducted by the
Public Interest Law Center of Philadelphia.20 Harding and

303
Fahey based their analysis of Chicago police shootings on re­
cords compiled by that city's coroner because the police depart­
ment refused to provide access to shooting-related data. Reti­
cence to provide such information on the exercise of the
ultimate police power may be a result of administrative fear of
simplistic or biased interpretations, and be justified, particu­
larly when administrators are dealing with clearly unfriendly
researchers. It may also indicate that administrators wish to
conceal potentially damaging or embarrassing research find­
ings, and imply that the police are "hiding something."
The belief that police may be hiding something also is en­
couraged by agency reluctance to make public the dispositions
of internal investigations of the justifiability of police shootings
or the criteria upon which those dispositions are based . These
perceptions may also be ill-founded, but their existence is dam­
aging to agency credibility and may reduce public confidence in
the police. Thus, as the Los Angeles Board of Police Commis­
sioners has observed, it is in the interests of the police to report
on the use of deadly force and to provide "for the communica­
tion of [results of investigations of use of deadly force] to the
community in a manner which merits public credibility and con­
fidence. " 21

DISCUSSION
Discussion of the limits of police power to use deadly force
may be framed in the context of traditional responsibilities of
the police . First, police manuals universally describe the pri­
mary poli-ce responsibility as protecting life. Nowhere is that
responsibility construed not to apply to the lives of suspected
criminals. Indeed, if we are to believe B-movies, much of the
work of western sheriffs required that they endanger their own
safety in order to protect the lives and due process guarantees
of villains against the "quick justice" sought by lynch mobs
composed of otherwise admirable townsfolk. Given that tradi­
tion and given the fact that, even if apprehended, tried, and
proven guilty beyond a reasonable doubt, no unarmed fleeing

304
 nonviolent felony suspect is likely to be executed by the state,
there arises the question of whether the fleeing felon rule con­
flicts with the police responsibility to protect life. If it could be
demonstrated that the loss of the lives of a few fleeing nonvio­
lent felony suspects would somehow preserve the lives of some
"good folks," one might conclude that no conflict exists. At­
tempting to demonstrate this point, however, involves argu­
ments even more tenuous than those that are made by persons
who favor capital punishment on the grounds that it preserves
life. As indicated above, there is little to support the notion that
the fleeing felon rule serves as a deterrent to criminality. But
even if it did deter, it is doubtful that it would preserve life.
Unlike the death penalty, the hypothetical deterrent power of
the fleeing nonviolent felony suspect rule affects only potential
nonviolent criminals.
A second police responsibility involves the preservation of
order. At the extreme, one might argue that the fleeing felon
rule serves as a symbol of society's determination to prevent
anarchy by granting police extreme power to act against those
who would violate the rights of others to possess and enjoy the
fruits of their labors. From this perspective, the fleeing felon
rule is an affirmation of traditional American values and an
indicator of the boundaries of acceptable behavior. At the other
extreme, the fleeing felon rule may be viewed as an indication
that society places a higher value upon property than upon
human life. Thus, the rule is likely to be regarded by the disen­
franchised, who are its most probable subjects, as an oppressive
and disproportionately harsh means of protecting the interests
of the prosperous against the peccadilloes of the young poor.
When contrasted with the minor penalties suffered by middle­
class offenders' whose crimes are usually far more profitable
than are those of the young poor, police shootings of unarmed
fleeing nonviolent felony suspects are likely not to be legiti­
mized by the disenfranchised, who know that shootings of flee­
ing white-collar-crime suspects are virtually unheard of. The
rule that allows such shootings is likely to be greatly resented,
and to contribute to public disorder rather than to public order.

305
 A third police responsibility pertinent to this discussion is
the mandate to fight crime through prevention and apprehen­
sion. As noted, the literature offers little evidence that the
fleeing felon rule has contributed to lowering the crime rate.
The clear incongruities between the rule that police may
shoot any fleeing felony suspect and police obligations to pro­
tect life, preserve order, and fight crime give rise to two ques­
tions. First, why does the fleeing felon rule remain the law in
more than half the states? Second, what are the obligations of
police in those states to remedy those incongruities?
The answer to the first question is probably to be found in
the nature of the legislative process. Law making and law revo­
cation are political processes, and lawmakers are politicians.
Legislators are accountable to their constituents, who ar~, by
definition, the franchised. The franchised are understandably
concerned about any legislative action that appears to diminish
the ability of police to protect society. Consequently, it takes a
courageous legislator to introduce or vote for a bill that is likely
to be perceived by voters as a "handcuff" on police powers. The
introduction of such a bill (or even a vote for it) almost certainly
would be raised during subsequent election campaigns as an
indication that the incumbent was "soft on crime."
Because of the apparent reluctance of legislators to deal
with the fleeing felon rule, police chiefs may face the choice of
negating the rule administratively, perhaps by establishing an
internal policy on deadly force with more stringent restrictions
than are found in state laws. Here again, however, police chiefs
who fear that implementing such policies will cause them to be
perceived as "soft on crime" show considerable reluctance to
act. Others fear that committing a restrictive policy to paper
will increase the civil liability of their departments and individ­
ual officers. This fear is apparently based on the questionable
logic that one cannot be sued successfully for shooting an un­
armed fleeing nonviolent felony suspect if there is no policy
formally prohibiting such an act. Finally, some chiefs are reluc­
tant to restrict police use of deadly force more tightly than does
the law, on the grounds that doing so would contradict the will
of the people as expressed through the legislative process. 22

306
 The danger that police chiefs who implement restrictive
deadly force policies will be perceived as soft on crime is real.
On the other hand, many chiefs have avoided this perception by
gaining the participation of community leaders and rank-and­
file officers in policy formulation, and by carefully explaining
the new policies produced as well as their purposes to both
community and department members. Still others have issued
such policies during periods of tension caused by specific in­
stances of deadly force, 23 with the timing of the announcement
immediately succeeding a manifestly unjustified use of deadly
force mitigating the potential for opposition. Finally, in some
states, police chiefs' associations have promulgated model
deadly force policies that represent their collective professional
opinions regarding proper use of police weapons. 24 The exist­
ence of such a model policy makes it possible for the individual
police chief who wishes to restrict deadly force to do so without
appearing arbitrary and without being singled out as soft on
crime. Instead , he can point out that he is acting in accord with
the consensus of his peers.
Chiefs who do not formulate policy on the ground that do­
ing so will increase the civil liability of their agencies and their
officers are interesting case studies. They are administrators of
justice who permit police to use unrestricted discretion in the
decision whether to shoot an unarmed fleeing nonviolent felony
suspect because their jurisdictions and their personnel might be
on the wrong end of wrongful death litigation if such shootings
were prohibited. In addition to the question of whether these
chiefs have correctly defined the "administration of justice"
(and have correctly identified the wrong end of wrongful death
litigation), such a policy omission involves the curious assump­
tion that officers will expose themselves and their agencies to
civil liability by violating a prohibition on use of deadly force to
apprehend unarmed fleeing nonviolent felony suspects. Two in­
ferences may be drawn from this assumpton. First, it is possi­
ble that this is a conscious assumption, that these chiefs believe
that their officers are not capable of staying within administra­
tively prescribed bounds of behavior. Given the plethora of reg­
ulations that govern other aspects of police behavior, however,
307
 such an inference is probably not valid. A second inference is
that this assumption is unconscious, and that these chiefs have
not given significant thought to the issue.
The second inference is probably more valid because it ac­
cords with the problem-solving model that pervades many po­
lice agencies. Police administrators generally are career per­
sonnel who have worked their way up through the ranks from
the patrol officer level. During the period in which they are
socialized and trained, new patrol officers are typically schooled
in the notion that "no two situations are alike." Thus, planning
for police field problems is a fruitless endeavor, and police
should be prepared to follow an ad hoc approach, solving prob­
lems as they arise. As a consequence, police usually become
very skilled at responding to crises, but rarely develop compa­
rable skill in devising plans to reduce or avoid future problems.
Police manuals, for example, are replete with detailed prescrip­
tions for such administrative behavior as the preparation of
reports on incidents that have already occurred. Rarely, how­
ever, do they offer much guidance for response to such predict­
able police field problems as armed robbery calls or traffic
stops. Instead, patrol officers are encouraged to assess situa­
tions as they occur, and to postpone formulation of an appropri­
ate response until that time.
Recent history offers considerable evidence that this crisis
management mode is operative at all police organizational lev­
els. The riots in Watts, Newark, and Detroit are examples of
large-scale police field problems for which administrative plan­
ning was either absent or inadequate. In these cases as in the
1971 Attica prison rebellion and the 1977 New York City black­
out looting, real attempts to develop adequate plans of response
occurred only after the fact. Thus, just as patrol officers rarely
give much thought to hypothetical field problems until they
materialize, police administrators rarely plan for more infre­
quent, but potentially cataclysmic, events until they have al­
ready taken place. Similarly, they often do not seriously con­
sider methods of controlling police behavior until a specific
behavior becomes a major public issue. Many police chiefs, in

308
other words, exercise few of their management prerogatives
unle ss they are pressured to do so from outside the organiza­
tion.
Ironically, the courts themselves may well become the
source of pressure for administrative modification of broad
statutory police deadly force justifications. The Eighth Circuit
Court of Appeals, in a decision later overturned by the United
States Supreme Court on grounds not related to the merits of
the case, ruled that Missouri's statutory justification for the use
by police of deadly force in order to apprehend any fleeing
felony suspect was in violation of Fourteenth Amendment due
process guarantees. 25 In a civil rights action based on a Pennsyl­
vania police shooting, a federal court indicated that the mere
flight of a forcible felon did not justify the use by police of
deadly force, even though such action was permissible under
Pennsylvania statutes. Instead the court ruled that the validity
and necessity of police decisions to use deadly force would be
evaluated case by case in terms of the danger posed by individ­
ual escapees.26 The Supreme Court has not yet ruled on the
constitutionality of the rule that police may shoot any fleeing
felon, and the federal decisions cited above are relatively iso­
lated cases. But these decisions nonetheless provide some evi­
dence that the courts are becoming increasingly concerned
about and involved with use of deadly force by state and local
police. Some police chiefs known to this author have, in fact,
instituted restrictive firearms policies in response to judgments
against them for shootings of unarmed fleeing nonviolent fel­
ony suspects that were justified by law. Thus, the trend of
court decisions is likely to make the institution of a restrictive
deadly force policy the best insurance against wrongful death
litigation.
The question of cover-ups is also a difficult one for the
police administrator to resolve. As indicated earlier, the low­
visibility milieux of most police shootings and the interests of
reporting officers in avoiding criminal, civil, or administrative
action frequently precipitate conjecture that official r eports do
not accurately reflect actually events. To be credible, therefore,

309
the police administrator must do everything possible to avoid
even the appearance of impropriety in matters pertaining to
use of deadly force. He must investigate instances of police
deadly force as intensively as his agency investigates the most
heinous crimes. Moreover, he must adjudicate shootings objec­
tively, and he should make public the results of his investiga­
tion, as well as the reasons for it.

RECOMMENDATIONS FROM THE LITERATURE

Because of the effects of police shootings, and because even
justifiable shootings and subsequent proceedings are so easily
perceived as exercises in injustice, many recommendations
have been made to reduce both the frequency of police shoot­
ings and the doubt that often accompanies them. 27 The consen­
sus of these recommendations may be summarized as follows:
1. Police departments should institute clear policy guide­
lines to limit the use of deadly force.
2. Policy guidelines should be related to the dangerous­
ness of suspects, and should prohibit use of deadly force to
apprehend nonviolent suspects.
3. Police departments should investigate thoroughly all in­
cidents in which police weapons are discharged, whether or not
anybody or anything is actually struck by bullets.
4. Reports of these investigations should minimize uncer­
tainty about what has happened by making as complete use of
witnesses and forensic techniques as possible.
5. All police shootings should be reviewed and adjudicated
by police departments, with contributions from as many per­
spectives and levels of the police organization as possible. These
adjudications should consider both the circumstances of the in­
stant shooting and the record of the officer involved.
6. The findings of investigations and adjudicatory bodies
should be made available to the public , and police departments
should react to questioning and criticism openly and responsibly
rather than defensively.

310
 7. Police agencies should attempt to improve and use
state-of-the-art psychological devices and techniques to screen
and monitor personnel in order to identify those likely to use
violence without proper justification.
8. Police supervisors and field commanders should be held
accountable for monitoring and acting on the unjustifiable use
of force or violence by personnel reporting to them.
9. Police training programs should be based upon social
science principles to increase officers' skills in daily interaction
and crisis intervention in the community, and to reduce the
possibility that police actions will cause the escalation of vio­
lence.
10. Police firearms training programs should consider le­
gal, administrative, and moral questions concerning use of the
gun, and should encourage the use of less drastic alternatives
where possible.
11. Police deployment policies and practices should be for­
mulated in such a manner that the potential for police-citizen
violence is reduced .
12. Police internal reward systems should be structured in
such a manner that quantitative measures of police work (e.g.,
numbers of arrests) and involvement in violent activities are
not the predominant or exclusive means of obtaining recogni­
tion for outstanding performance.
13. Police agencies should encourage citizens to make com­
plaints about officers' use of abuse or unnecessary force, should
thoroughly investigate all such allegations, and should advise
complainants of findings and the action taken.

CONCLUSION
Police have a major stake in determining the limits imposed
on the use of deadly force. Indiscriminate use of weapons is a
discredit to police departments and a disservice to the citizens
they protect. Broad mandates to use police deadly force, such
as the fleeing felon rule, have no place in the administration of
justice in a democratic society.

311
 Notes
1. Most definitions of "deadly force" describe it as force likely to kill
or capable of taking life. Since deadly force is almost always exercised
with a firearm, the most appropriate quantitative measure of its fre­
quency is the number of times police shoot at others, rather than
simply the number of persons killed. This is so because police bullets
do not always kill, and because missed shots, woundings, and fatalities
often are merely varying results of equally grave decisions. Stated
simply, police deadly force is best measured by counting decisions to
employ it: "Body counts" represent only the consequence of some
percentage (approximately 9 percent in New York City) of those deci­
sions. See James J. Fyfe, "Shots Fired: an Examination of New York
City Police Firearms Discharges" (Ph .D. diss., State University of
New York at Albany, 1978), p. 301.
2. See, respectively, President's Commission on Law Enforcement
and Administration of Justice, Task Force Report: The Police (Wash­
ington, D. C.: Govt. Printing Office, 1967); Jerry V. Wilson, "Deadly
Force," Police Chief, December 1972, pp. 44-56; Southern Methodist
Law School, "Report on Police Shootings" (Report to Dallas Police
Department, 1974); Paul Takagi, "A Garrison State in 'Democratic'
Society," Crime and Social Justice. Spring-Summer 1974, pp. 27-33;
John S. Goldkamp, "Minorities as Victims of Police Shootings: Inter­
pretations of Racial Disproportionality and Police Use of Deadly
Force," Justice System Journal, Winter 1976, pp. 169-83; Kenneth B.
Clark, "Open Letter to Mayor Abraham D. Beame and Police Commis­
sioner Michael J. Codd," New York Times, Sept. 17, 1974; Lawrence
W. Sherman, "Execution without Trial; Police Homicide and the Con­
stitution," Vanderbilt Law Review, January 1980, pp. 71-100; Robert
Sherrill, The Saturday Night Special (New York: Charterhouse, 1973).
3. President's Commission, The Police, pp. 189-90; Sherman, "Execu­
tion without Trial."
4. Sherman, "Execution without Trial," pp. 71, 72.
5. Robert M. McKiernan, "Police Shotguns : Devastating to the Ani­
mals," New York Times, Feb. 7, 1973, p. A-35.
6. Gerald D. Robin, "Justifiable Homicide by Police Officers," Jour­
nal of Criminal Law, Criminology and Police Science, June 1963, pp.
225-31.
7. Michael Kiernan, "Shooting of Civilians by District Police Declines
by Nearly Half," Washington Star, Sept. 2, 1979, p. B-1; Catherine H.
Milton et al., Police Use of Deadly Force (Washington, D.C.: Police
Foundation, 1977); Gerald F. Delman, "Varieties of Police Policy: A

312
 Study of Police Policy Regarding the Use of Deadly Force in Los
Angeles County," Loyola of Los Angeles Law Review, 1973, pp. 1-61.
In addition, previous research by this author also supports the conten­
tion that there is little association between police shooting rates and
rates of reported crime. The 1972 imposition of restrictive police
shooting guidelines in New York City resulted in precipitous drops in
shooting by that agency's personnel, with no significant increases or
decreases in either crime or arrest rates. See Fyfe, "Shots Fired," pp.
218-329.
8. National Advisory Commission on Civil Disorders, Commission Re­
port (New York: Bantam, 1968).
9. John P. Adams, "Police Deadly Force and Church Concern," in A
Community Concern: Police Use of Deadly Force, Robert N. Brenner
and Marjorie Kravitz, eds. (Washington, D.C.: Govt. Printing Office,
1979), pp. 13-18.
10. Donald J. Newman, Introduction to Criminal Justice, 2d ed. (Phil­
adelphia, J. B. Lippincott, 1978), pp. 430-31.
11. Kiernan, "Shootings of Civilians," reports that 1978 police shoot­
ing rates per 1,000 forcible felony arrests in ten American cities varied
from less than 1 per 1,000 arrests to more than 20 per 1,000 arrests.
12. Lawrence W. Sherman and Robert H. Langworthy, "Measuring
Homicide by Police Officers," Journal ofCriminalLaw and Criminol­
ogy, Winter 1979, pp. 546-60.
13. Fyfe, "Shots Fired."
14. Federal Bureau of Investigation, Crime in the United States: 1978
(Washington, D.C.: Dept. of Justice, 1979), p. 35.
15. See Paul Chevigny, Police Power: Police Abuses in New York City
(New York: Pantheon Books, 1969).
16. Rodney Stark, Police Riots (Belmont, Calif.: Wadsworth, 1972).
17 . Arthur L. Kobler, "Police Homicide in a Democracy," Journal of
Social Issues, Winter 1975, pp. 168-84.
18. Richard W. Harding and Richard P. Fahey, "Killings by Chicago
Police, 1969-70: An Empirical Study," Southern Californi a Law Re­
view, March 1973, pp. 284-315.
19. Board of Police Commissioners, "Part I - The Shooting of Eulia
Love," The Report of the Board of Police Commissioners Concerning
the Shooting ofEulia Love and the Use of Deadly Force (Los Angeles:
Board of Police Commissioners, 1979), p. 30.
20. Public Interest Law Center of Philadelphia, "A Study of the Use

313
 of Firearms by Philadelphia Policemen from 1970 through 1974"
(1975).
21. Police Commissioners, "Shooting of Eulia Love." p. 29.
22. At a panel on deadly force at the National Forum on Criminal
Justice held in Columbus, Ohio, in 1980, for example, a former police
chief who is directing a federally funded International Association of
Chiefs of Police study of police use of deadly force remarked that an
administrative policy that restricted deadly force more stringently
than state law was "unfair to officers," because it increased the
chance that they could be successfully sued for violating it. Evidence
that this sentiment is widespread among police also is provided by an
International Association of Chiefs of Police membership vote to reject
a resolution recommending that police agencies limit deadly force to
"defense of life" situations. Instead, at its 1980 annual convention, the
International Association of Chiefs of Police approved a resolution
calling upon police agencies to adopt deadly force policies "consistent
with state law."
23. Sherman, "Execution without Trial ," p. 91.
24. See, for example , Milton et al. , Police Use of Deadly Force, pp.
151-54; Delaware Police Chiefs' Council, "Administrative Policy State­
ment: Deadly Force" (Draft; Feb. 16, 1981).
25. Mattis v. Schnarr, 547 F.2d 1007 (8th Cir. 1976).
26. Phillips v. Ward, 415 F. Supp. 976 (D.C. 1976).
27. See James J. Fyfe, "Administrative Interventions on Police
Shooting Discretion: An Empirical Evaluation," Journal of Criminal
Justice, Winter 1979, pp. 309-23; Harding and Fahey, "Killings by
Chicago Police"; Kobler, "Police Homicide": Kenneth R. McCreedy
and James L. Hague , "Of a Policy to Limit the Use of Firearms by
Police Officers," Police Chief, January 1975, pp. 48-52; Milton et al.,
Police Use of Deadly Force: Gilbert G. Pompa, "A Major and Most
Pressing Concern" (Paper presented at the Annual Meeting of the
National Organization of Black Law Enforcement Executives, St.
Louis, June 22, 1978); President's Commission, The Police; Barbara
Rhine, "Kill or Be Killed: Use of Deadly Force in the Riot Situation,"
California Law Review, 1968, p. 829; Sherman, "Execution without
Trial"; Hans Toch, Peacekeeping-Police, Prisons, and Violence Lex­
ington, Mass.: D.C. Heath, 1977).

314
 ABOUT THE AUTHORS

Gary L. Abrecht is a patrol captain with the Metropolitan

Washington, D.C., Police Department. He has a B.A. in eco­
nomics from Yale.

J. Paul Boutwell, an attorney with the firm of Adair and

Goldthwait, is retired from the FBI.

Samuel G. Chapman is professor of political science at the Uni­

versity of Oklahoma and a member of the Norman City Council.
He has both a B.A. and an M.A. in criminology from the Uni­
versity of California at Berkeley, and was a police officer in
Berkeley for five years.

James J. Fyfe is an associate professor at The American Uni­

versity, Washington, D.C., and is a senior fellow at the Police
Foundation. He is a former lieutenant with the New York City
Police Department, where he served for 16 years. He holds a
Ph.D. from the State University of New York at Albany.

JohnS. Goldkamp is associate professor and chair of the De­

partment of Criminal Justice at Temple University in Philadel­
phia. He also is director of the NIC-NIF funded Bail Decision­
Making Project, studying bail practices and experimenting with
bail guidelines, and is the author of Two Classes ofAccused. He
holds a Ph.D. from the State University of New York at Al­
bany.

Jeanne Wahl Halleck, former staff director of National Advi­

sory Committee for Juvenile Justice and Delinquency Preven­
tion, currently is doing case studies on delinquency prevention
and juvenile justice. She holds both undergraduate and gradu­
ate degrees from the School of Criminology at Berkeley.

315
Robert H. Langworthy, a research analyst at the Criminal Jus­
tice Research Center, Inc., is completing a doctoral dissertation
on police organizational structure at the School of Criminal
Justice, State University of New York at Albany.

James Lardner, a former Metropolitan Washington, D.C., po·

lice officer, is a staff writer for the Washington Post.

Marshall W. Meyer is a professor of sociology at the University

of California, Riverside. He received an A.B. from Columbia
University and an M.A. and a Ph.D. from the University of
Chicago. His recent books include Change in Public Bureaucra­
cies and Environments and Organizations, and he has contrib­
uted numerous articles to scholarly journals.

Catherine Milton, a former assistant director of the Police

Foundation, has served on the staffs of the International Asso­
ciation of Chiefs of Police, the American Bar Association, and
the U.S. Department of Treasury. She currently is consulting
on criminal justice issues for the U.S. Senate.

Lawrence W. Sherman, an associate professor of criminal jus­

tice at the State University of New York , Albany, is director of
research for the Police Foundation. He holds a Ph.D. from Yale
University, and was senior author of The Quality ofPolice Edu­
cation. He is also the author of Scandal and Reform: Control­
ling Police Corruption, and is preparing a book on his two-year
study for the National Institute of Mental Health on homicide
by police officers.

Paul Takagi is associate professor at the School of Education ,

University of California- Berkeley, and is co-editor of the jour­
nal Crime and Social Justice.

316