U.S.

Army Research Institute

for the Behavioral and Social Sciences




Research Report 1927



The Roles of Perseverance, Cognitive Ability, and
Physical Fitness in U.S. Army Special Forces
Assessment and Selection






Scott A. Beal
U.S. Army Research Institute




July 2010




Approved for public release; distribution is unlimited.

U.S. Army Research Institute
for the Behavioral and Social Sciences

Department of the Army
Deputy Chief of Staff, G1

Authorized and approved for distribution:

BARBARA A. BLACK, Ph.D. MICHELLE SAMS, Ph.D.
Research Program Manager Director
Training and Leader Development
Division

Technical Review by

Peter Legree, U.S. Army Research Institute
Thomas R. Graves, U.S. Army Research Institute









NOTICES

DISTRIBUTION: Primary distribution of this Research Report has been made by ARI.
Please address correspondence concerning distribution of reports to: U.S. Army
Research Institute for the Behavioral and Social Sciences, Attn: DAPE-ARI-ZXM,
2511 Jefferson Davis Highway, Arlington, Virginia 22202-3926.

FINAL DISPOSITION: This Research Report may be destroyed when it is no longer
needed. Please do not return it to the U.S. Army Research Institute for the Behavioral
and Social Sciences.

NOTE: The findings in this Research Report are not to be construed as an official
Department of the Army position, unless so designated by other authorized documents.
i

REPORT DOCUMENTATION PAGE
1. REPORT DATE (dd-mm-yy)
July 2010

2. REPORT TYPE
Final
3. DATES COVERED (from. . . to)
December 2008 to October 2009


4. TITLE AND SUBTITLE

The Roles of Perseverance, Cognitive Ability, and Physical
Fitness in U.S. Army Special Forces Assessment and Selection



5a. CONTRACT OR GRANT NUMBER




5b. PROGRAM ELEMENT NUMBER
A792
6. AUTHOR(S)

Scott A. Beal (U.S. Army Research Institute)






5c. PROJECT NUMBER
633007
5d. TASK NUMBER
103


5e. WORK UNIT NUMBER


7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

U.S. Army Research Institute for the Behavioral and Social Sciences
Fort Bragg Scientific Coordination Office
DAPE-ARI-RB
Fort Bragg, NC 28310
8. PERFORMING ORGANIZATION REPORT NUMBER

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
U.S. Army Research Institute for the Behavioral and Social
Sciences
2511 Jefferson Davis Highway
Arlington, VA 22202-3926
Source Code: 597861
10. MONITOR ACRONYM
ARI
11. MONITOR REPORT NUMBER
Research Report 1927
12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release; distribution is unlimited.

13. SUPPLEMENTARY NOTES
Subject Matter POC: Scott A. Beal





14. ABSTRACT (Maximum 200 words):
The U.S. Army Research Institute for the Behavioral and Social Sciences (ARI) conducted research for more than a decade in
support of U.S. Army Special Operations Forces (ARSOF) assessment, selection, and training. This research was completed prior
to the events of September 11, 2001. Leaders at the John F. Kennedy Special Warfare Center and School (SWCS) at Fort Bragg,
NC, requested that ARI begin updating research in support of Special Forces. This report documents a new effort to understand
better the roles of cognitive ability, physical fitness, and performance events in the Special Forces Assessment and Selection
(SFAS) process. In addition, researchers identified a measure of perseverance, viewed as an individual Soldier characteristic, to
include in the investigation at the SWCSs request. The 824 Candidates who participated in this research completed a series of
cognitive ability tests, physical fitness measures, SFAS performance events, and the test of perseverance. The results showed
that almost all the tests and measures included in the analyses contributed to valid predictions of Soldier success with SFAS, but
that their individual strengths of prediction varied. The SFAS performance events provided the greatest predictive strength,
followed by the cognitive ability and physical fitness tests. While perseverance provided a unique contribution, its role was
incremental, at best, and should not be used as a criterion for selection decisions in isolation from the other measures. Taken
together, the tests and measures form an empirically-sound foundation upon which SFAS decisions can be based. Once obtained
and analyzed, the outcome data from the Special Forces Qualification Course (SFQC) and subsequent training will provide a more
complete view of how well the tests and measures included in this research predict long-term success in Army Special Forces.


15. SUBJECT TERMS
Army Special Forces Personnel Selection Perseverance Predictors of success SFAS



SECURITY CLASSIFICATION OF 19. LIMITATION OF
ABSTRACT

Unlimited
20. NUMBER
OF PAGES

24
21. RESPONSIBLE PERSON
Ellen Kinzer
Technical Publications Specialist
(703) 602-8049

16. REPORT
Unclassified
17. ABSTRACT
Unclassified
18. THIS PAGE
Unclassified

ii




iii

Research Report 1927





The Roles of Perseverance, Cognitive Ability, and Physical
Fitness in U.S. Army Special Forces Assessment and
Selection





Scott A. Beal
U.S. Army Research Institute





ARI Scientific Coordination Office (SOCOM/FORSCOM)
James W. Lussier, Chief



U.S. Army Research Institute for the Behavioral and Social Sciences
2511 Jefferson Davis Highway, Arlington, Virginia 22202-3926



July 2010



Army Project Number Personnel Performance
633007A792 and Training


Approved for public release; distribution is unlimited.

iv

ACKNOWLEDGEMENTS


The author would like to thank LTC Kris Woolley, Command Psychologist, John F.
Kennedy Special Warfare Center and School (SWCS), for initiating and guiding this research
effort. The author also thanks the leaders and cadre at 1
st
Battalion, Special Warfare Training
Group (SWTG) who provided subject-matter expertise and coordinated data collection events.



v
THE ROLES OF PERSEVERANCE, COGNITIVE ABILITY, AND PHYSICAL FITNESS IN
U.S. ARMY SPECIAL FORCES ASSESSMENT AND SELECTION


EXECUTIVE SUMMARY

Research Requirement:

Prior to September 11, 2001, the U.S. Army Research Institute for the Behavioral and
Social Sciences (ARI) completed more than a decade of research in support of the U.S. Army
Special Operations Forces (ARSOF) at Fort Bragg, NC. From these years of research, ARI
published reports on Special Forces manpower and personnel, assessment, selection, training,
and Soldier development. Army Special Forces leaders and trainers have maintained interest in
these specific areas of research since 9/11. However, sweeping changes at all levels of Special
Forces necessitated updating the research to provide more relevant and timely results.

The leaders at the John F. Kennedy Special Warfare Center and School (SWCS), who are
responsible for the growth and maintenance of Special Forces, were particularly interested in the
extent to which the current tests and measures used for assessing and selecting Soldiers predicted
their success. They recognized also the need for assessing perseverance as an individual Soldier
characteristic that contributed to selection following the Special Forces Assessment and
Selection (SFAS) process, and ultimately with completing the Special Forces Qualification
Course (SFQC). This report documents one effort at updating this research. More specifically, it
analyzes the extent to which Soldiers cognitive ability, physical fitness, ability to complete
difficult performance events, and perseverance contributed to valid and reliable predictions of
SFAS selection.

Procedure:

Candidates admitted to SFAS (N=824) completed a battery of cognitive ability, physical
fitness, and perseverance tests prior to attempting to complete a series of SFAS performance
events. During the SFAS process, a proportion of the Candidates were withdrawn due to pre-
existing or emerging physical injuries or medical problems. The data from these Candidates
were removed from the analyses to eliminate their impact on the overall results. A larger
proportion of Candidates withdrew from SFAS voluntarily, most citing insufficient physical
fitness as the primary reason for their withdrawal. The Candidates who completed SFAS
successfully were brought before a board of leaders who made the final selection decisions for
admitting the Candidates to the subsequent SFQC. The ARI obtained and analyzed the data to
determine the predictive strengths of the variables contributing to SFAS completion and
selection for the SFQC. In addition, the data were used to analyze the potential impact of
selected variables on Special Forces recruiting. Finally, ARI explored the extent to which a
more efficient model of tests could be identified and used to predict success.





vi

Findings:

The results showed that almost all of the current tests and measures used to assess SFAS
Candidates were significant predictors of selection for training following the SFAS performance
events, although the strength of their individual contributions varied. The SFAS performance
event scores were strongest for predicting selection, followed by scores from cognitive ability
tests, the Army Physical Fitness Test (APFT), and finally perseverance. Taken together, the
current tests and events accounted for a robust amount of variance in our Soldier sample. The
Special Forces recruiters could use the measures available to them to increase the stringency of
current recruiting standards, which may increase the probability of choosing successful SFAS
Candidates, but may also reduce the overall force structure. A more efficient model of five
variables was derived from the analyses that accounted for slightly more variance than the
current model of tests and measures. However, using the more efficient model would reduce the
amount of information on which board leaders base their selection decisions.

Utilization and Dissemination of Findings:

This report represents an effort to update ARIs previous research in support of U.S.
Army Special Forces, and supports the overall mission of the U.S. Army Special Operations
Command (USASOC). The results of this research were briefed to members of the Directorate
of Special Operations Proponency (DSOP) Command Group and continue to influence decisions
made by leaders who are responsible for maintaining and increasing the force structure of U.S.
Army Special Forces.























vii
THE ROLES OF PERSEVERANCE, COGNITIVE ABILITY, AND PHYSICAL FITNESS IN
U.S. ARMY SPECIAL FORCES ASSESSMENT AND SELECTION

CONTENTS

Page

PREFACE ..................................................................................................................................... 1

BACKGROUND AND PURPOSE .............................................................................................. 1

SFAS DATA COLLECTION ....................................................................................................... 4

RESULTS ..................................................................................................................................... 5

DISCUSSION ............................................................................................................................... 10

CONCLUSIONS AND RECOMMENDATIONS ....................................................................... 13

REFERENCES ............................................................................................................................. 17

LIST OF TABLES

TABLE 1. COGNITIVE ABILITY TESTS, PHYSICAL FITNESS TESTS, AND SFAS
EVENTS FROM WHICH DATA WERE COLLECTED FOR PRELIMINARY
ANALYSES ................................................................................................................. 4

TABLE 2. THE DISTRIBUTION OF RANKS AMONG THE PARTICIPATING
CANDIDATES .......................................................................................................... 5

TABLE 3. GRIT SCALE SUBSCALE RELIABILITIES .......................................................... 6

TABLE 4. BINARY LOGISTIC REGRESSION ANALYSES USING INDIVIDUAL GRIT
SUBSCALES ............................................................................................................. 6

TABLE 5. BINARY LOGISTIC REGRESSION ANALYSES USING COGNITIVE ABILITY
TEST SCORES .......................................................................................................... 7

TABLE 6. BINARY LOGISTIC REGRESSION ANALYSES USING PHYSICAL FITNESS
AND SFAS PERFORMANCE EVENT SCORES.................................................... 8

TABLE 7. POTENTIAL IMPACT OF IMPLEMENTING MORE STRINGENT
REQUIREMENTS FOR SFAS ADMISSION USING GT, APFT, AND GRIT
SCALE SCORES ....................................................................................................... 10

TABLE 8. IMPACT OF VARIABLES FROM MOST EFFICIENT MODEL ON SFAS
SELECTION .............................................................................................................. 13

viii



1

THE ROLES OF PERSEVERANCE, COGNITIVE ABILITY, AND PHYSICAL FITNESS IN
U.S. ARMY SPECIAL FORCES ASSESSMENT AND SELECTION

Preface

The U.S. Army Research Institute for the Behavioral and Social Sciences (ARI)
conducted research for more than a decade in support of Army Special Operations Forces
(ARSOF). The results of these research efforts were published in the following areas of interest:
(a) supporting manpower and personnel, (b) improving selection, (c) developing tools for
assessment, (d) enhancing training, and (e) Soldier and leader professional development (see
Brooks & Zazanis, 1997 for a complete review). While these research areas continue to be
relevant and of interest to ARSOF leaders, ARIs investigations were conducted prior to the
events of September 11, 2001. Given the sweeping changes in Special Forces (SF) training and
combat operations in response to these events, a need arose to replicate some of the earlier
research conducted by ARI. This report presents the results of a systematic replication of ARIs
earlier efforts to assist in developing a better understanding of the impact of cognitive ability and
physical fitness on SF recruiting and selection. In addition, it summarizes the role of
perseverance as a newly investigated individual characteristic in SF assessment and selection.

Background and Purpose

The U.S. Army John F. Kennedy Special Warfare Center and School (SWCS) at Fort
Bragg, NC, forms and implements the standards for selecting Army Special Forces (SF) Soldiers.
These standards ensure that the Soldiers entering SF can demonstrate physical toughness, mental
sharpness, and enough operational readiness potential to meet the demands of current and future
special operations. The SWCS meets their requirements for maintaining the overall SF force
structure, in part, through a Special Forces Assessment and Selection (SFAS) process. This
process includes recruiting Army Soldiers (and a limited number of civilians), assessing their
cognitive and physical appropriateness for SF, testing their ability to complete a rigorous series
of performance-driven events, and then selecting some of them for additional SF training.

Because the events during SFAS are difficult, as many as 40% of the recruited candidates
in a given class will voluntarily withdraw (VW) during the process. The vast majority of the
Soldiers who VW cite their inability to complete the events due to insufficient physical fitness.
However, recent data analyses have shown that about 15% of the Soldiers who completed SFAS
successfully and were selected for additional SF training had similar or even lower physical
fitness and IQ scores than some of the Soldiers who VW during SFAS. These results suggest
that there is an element of individual perseverance that reinforces a Soldiers willingness to stay
the SFAS course, in spite of perceived or real limitations. It also suggests that, while mental and
physical achievement and ability are important, the probability of success increases when a high
level of performance is sustained over time and under increasingly difficult conditions; that is,
when Soldiers persevere.

Leaders and trainers at SWCS, particularly those directly responsible for SFAS
outcomes, wanted to explore the extent to which an individual characteristic like perseverance

2

contributed to Soldiers completing SFAS and being selected for SF training. In addition, SWCS
wanted to better understand the potential consequences of modifying their recruiting practices if
empirical evidence suggested that Soldiers with higher levels of reported perseverance had a
greater tendency to complete the SFAS process successfully. The SWCS leaders asked ARI at
Fort Bragg to identify a tool appropriate for measuring perseverance in a specialized military
context, and then to validate the tool with SFAS candidates. In addition, leaders tasked ARI to
develop a most efficient predictive model that combined existing SF cognitive ability tests and
physical fitness test scores with the measure of perseverance to predict Soldiers who might VW
from SFAS, versus those who would be selected for additional training. This report documents
the results of this research effort. The findings suggest that a test of perseverance adds
incremental strength to predict SFAS success when combined with existing cognitive ability
tests, physical fitness tests, and SFAS performance event scores.

Identifying the Grit Scale as an Appropriate Measure of Perseverance

The belief that exceptional accomplishments can be produced by individuals that
demonstrate something other than inherent ability is not new (see Terman & Oden, 1947). In
large measure, SF operates, trains, and manages its expectations of current and future capabilities
according to the widely held assumption that successes are obtained through consistent hard
work and perseverance. Selection guidelines and training programs in support of SF are written
from this vantage point. Yet, to date, military researchers have done little to establish an
empirical basis for perseverance as a measurable, individual Soldier characteristic. The
requirements to maintain the SF force structure and reduce the percentage of Soldiers who VW
from SFAS provided the impetus for exploring the influence of perseverance in a SF context.

Duckworth, Peterson, Mathews and Kelly (2007) combined results from a series of
studies that measured the extent to which the Grit Scale, a survey of perseverance, predicted
educational attainment, grade point average, and military retention in adults. The Grit Scale
includes five subscales that are each represented by a combination of items from the survey. The
subscales (and the number of items that comprise them) are as follows: Grit Score (12 items),
Consistency of Interest (six items), Perseverance of Effort (six items), Brief Grit (eight items),
and Ambition (five items). The authors defined the construct grit as:

perseverance and passion for long-term goals. Grit entails working
strenuously toward challenges, maintaining effort and interest over years despite
failure, adversity, and plateaus in progress. The gritty individual approaches
achievement as a marathon; his or her advantage is staminaDisappointment or
boredom signals to others that it is time to change trajectory and cut losses,
[whereas] the gritty individual stays the course (p. 1087-1088).

The Army, to include SF, has relied heavily on indicators of cognitive ability (e.g.,
Armed Services Vocational Aptitude Battery General Technical (ASVAB-GT) score) to
inform decisions about selection, job assignment, and training. One finding discussed by
Duckworth, et al., showed that grit was greater than IQ in predictive validity and magnitude
when measuring achievement. In addition, the Grit Scale did not correlate with IQ. From these
findings, three important conclusions can be drawn about perseverance and its potential

3

application to this research: (a) that it is a measurable predictor of accomplishment, (b) that it
can account for more variance than IQ, and (c) that it can account for variance independent of
IQ. It was for these reasons that we explored the efficacy of the Grit Scale, or perseverance, as a
component of our predictive model for validation with SFAS Candidates.

Testing During SFAS

During the SFAS process, Soldiers complete a battery of tests that includes measures of
language aptitude, general intelligence, and several other dimensions of cognitive ability, along
with the Army Physical Fitness Test (APFT). In addition, multiple types of biographical data
(e.g., age, years in active duty, education, etc.) are collected from each Soldier. Recruited
Soldiers are admitted as candidates into an SFAS class when they meet or exceed established
criteria for cognitive and physical fitness.

Following initial testing, the SFAS process requires that admitted candidates attempt to
complete a series of difficult performance-based events. These events test the candidates
problem solving abilities, their physical strength and endurance, and their willingness to endure
hardship under conditions that mirror the operational circumstances during SF combat missions.
The SFAS Cadre collect and maintains the candidates performance data. The candidates who
complete the SFAS performance events successfully are presented before a board of trainers and
leaders who make final selection decisions. The board members make selection decisions
according to a general statement of risk that represents a composite of data and test scores for
each candidate.

The ARI analyzed data from several previous SFAS classes to understand how the tests
and events impacted selection. The results of statistical analyses (e.g., linear regression, binary
logistic regression) showed that all the tests and events in current use had properties that
contributed to predicting selection, though the results highlighted variations in the relative
predictive strengths of each test and event. Table 1 shows the tests and events from which data
were collected during the recruiting and SFAS processes, and then analyzed.
1
















1
Our intent was to publish this research report with unlimited distribution. As such, we omitted detailed results of
these data analyses and the details of the SFAS events. We also omitted the specific procedures for collecting data.

4

Table 1

Cognitive Ability Tests, Physical Fitness Tests, and SFAS Events from which Data Were
Collected for Preliminary Analyses
Cognitive Ability Tests Physical
Fitness
SFAS
Event
Armed Services Vocational
Aptitude Battery General
Technical score (ASVAB-GT)
Army Physical
Fitness Test
(APFT)
1
st
Ruck
March
Defense Language Aptitude Battery
(D-Lab)
Number of
Pull-ups
2
nd
Ruck
March
Test of Adult Basic Education
(TABE)
1
st
Run
Wonderlic Personnel Test 2
nd
Run
General Ability Measure for Adults
(GAMA)
Obstacle
Course
Years of School Completed

To meet the purposes of this research, all the tests and events shown in Table 1 were
combined with the Grit Scale to determine their respective and collective roles in predicting
selection for SF training. Another purpose was to understand better their potential impact on SF
recruiting as if modifications to recruiting practices had already been implemented.

SFAS Data Collection

Candidates from four SFAS classes (N=824) participated in this research. The data were
collected November 2008 through February 2009. The procedures for recruiting, testing, and
performance events, and the selection standards and procedures, were held constant across the
four classes to decrease the potential for confounding variables. The participating candidates
completed all the cognitive ability and physical fitness tests that were required of previous SFAS
classes. They completed the Grit Scale in conjunction with all other paper-and-pencil tests,
which were administered at the beginning of each SFAS class. Following the testing, candidates
began the process of completing the SFAS performance events in preparation for the selection
board.

The data from previous SFAS classes showed that candidates experienced one of the
following outcomes:

Candidate VW from SFAS.
Candidate was withdrawn during SFAS due to a pre-existing physical injury or medical
condition.
Candidate was withdrawn during SFAS due to a physical injury or medical condition that
emerged during the process.
Candidate completed the performance events, but was not selected by the SFAS board.
Candidate completed the SFAS performance events and was selected for SF training.


5

Results

Biographical Data

The average age of the participating candidates was 26 years. Fifty-two percent of the
candidates were high school graduates. Twenty-nine percent of the candidates had completed
some college, 17% held Bachelors degrees, and 2% had completed some post-graduate work.
Seventy-two percent of the candidates reported that they had experienced combat during
Operation Iraqi Freedom/Operation Enduring Freedom (OIF/OEF). Approximately 90% of the
candidates were noncommissioned officers (NCO), and the remaining 10% were commissioned
officers. Table 2 shows the distribution of ranks among the participating candidates.

Table 2

The Distribution of Ranks among the Participating Candidates
NCO Ranks Frequency Percentage (%) of
Total Candidates
E1 3 <1
E2 3 <1
E3 86 10
E4 256 31
E5 239 29
E6 138 17
E7 18 2
Officer
Ranks
Frequency Percentage (%) of
Total Candidates
O1 1 <1
O2 8 <1
O3 72 9

Grit Scale

From the four SFAS classes, we retained for analysis the data from the candidates
(N=758) who VW (298/39%), as well as the data from those who completed the SFAS
performance events but were not selected by the board (110/15%), and from those who
completed SFAS and were selected by the board for SF training (350/46%). By doing so, we
eliminated the potential impact of physical injuries and other medical conditions on our results
and our ability to develop an efficient model to predict success. Therefore, 46% (350) of the
candidates were selected, and 54% (408) were not selected.

Scale Reliability. Table 3 shows the reliability estimates for the Grit Scale subscales.
These results show that the subscales have sufficient reliability for this Candidate sample.





6

Table 3

Grit Scale Subscale Reliabilities
Subscale
Reliability (Cronbachs )
Grit .81
Consistency of Interest .81
Perseverance of Effort .70
Brief Grit .78
Ambition .69

Predictive Strength of Grit Subscales. To examine the effects of the individual Grit
subscales on SFAS selection, the scores were standardized for ease of interpretation. The SFAS
selection variable was coded as a dichotomous variable where 1 = selected and 0 = not selected
or VW. The results from binary logistic regression analyses, one analysis for each of the Grit
subscales, are shown in Table 4.

Table 4

Binary Logistic Regression Analyses Using Individual Grit Subscales
Grit
Subscale
N B (beta) p value Nagelkerke R Odds Ratio
Grit Score 758 .198 .008 .013 1.22
Consistency
of Interest
758 .115 .114 .004 1.12
Perseverance
of Effort
758 .189 .010 .012 1.21
Brief Grit 758 .194 .009 .012 1.21
Ambition 758 .219 .003 .016 1.25

The values for beta (a change in the log odds of selection as a function of a one unit
change in a subscale score) and the odds ratios (a change in the odds of selection as a function of
a one unit change in a subscale score) are included in the table. A p value less than or equal to
0.05 suggests that the subscale is a significant predictor of SFAS selection. The Nagelkerke R
represents the variance accounted for by each subscale.

Based on these results, the Consistency of Interest subscale was not a significant
predictor of SFAS selection and was removed from further analyses. The odds ratios for the
remaining four subscales show that the candidates who scored a standard deviation above the
subscale mean scores were 21% to 25% more likely to be selected. The amounts of variance
accounted for by the four subscales were relatively small, which suggested that, though the
subscales were significant predictors of SFAS selection, their contribution was slight.
Perseverance Model. The four Grit subscales retained for further analyses were included
in a single binary logistic regression (with selected vs. not selected or VW as the dependent
variable) to determine the effect of a Perseverance model for predicting selection. The results
showed that the Perseverance models effect on SFAS selection was significant, (5, 758) =

7

11.23, p <.024; -2 Log likelihood = 1035.140; Nagelkerke R = .020. However, the amount of
variance accounted for by this model was relatively small and should not be interpreted in
isolation from additional tests reported below.

Cognitive Ability Tests

To determine and ease interpretation of the effects of individual cognitive ability tests on
SFAS selection, test scores were standardized. The same dichotomous dependent variable (i.e.,
selected vs. not selected or VW) was used and binary logistic regressions were performed for
each cognitive test. Table 5 shows the results of these analyses.

Table 5

Binary Logistic Regression Analyses Using Cognitive Ability Test Scores
Cognitive
Ability Test
N B
(beta)
p value Nagelkerke R Odds Ratio
Entry GT 721 .620 <.001 .110 1.86
Wonderlic 758 .574 <.001 .097 1.78
GAMA 758 .455 <.001 .062 1.58
D-Lab 732 .631 <.001 .113 1.88
TABE 756 .872 <.001 .139 2.39
Years of school
completed
754 .513 <.001 .076 1.67
All five cognitive ability tests and Years of school completed were significant predictors
of SFAS selection and were retained for further analyses. The Candidates who scored one
standard deviation above the mean for Entry GT were 86% more likely to be selected, whereas
those who scored a standard deviation above the mean Wonderlic score were 78% more likely to
be selected. Those who scored a standard deviation above the means for GAMA and D-Lab
were 58% and 88% more likely to be selected, respectively. Those who scored a standard
deviation above the TABE mean, and those who had one standard deviation above the mean for
Years of school completed were 139% and 67% more likely to be selected, respectively.
Cognitive Ability Model. The standardized scores from the cognitive ability tests and
Years of school completed shown in Table 5 were included in a single binary logistic regression
(with selected vs. not selected or VW as the dependent variable) to determine the effect of a
Cognitive Ability model for predicting SFAS selection. The results showed that this models
effect on SFAS selection was significant, (6, 691) = 105.45, p <.001; -2 Log likelihood =
846.352; Nagelkerke R = .189. The Cognitive Ability Model accounted for approximately 19%
of the variance, which suggests that the predictive power of this model is relatively robust.

Physical Fitness and SFAS Performance Events

Binary logistic regressions were performed for each physical fitness measure and each
SFAS performance event to establish the magnitude of their individual effects on selection. All
test scores were standardized to ease interpretation of results. Because lower scores for the ruck
marches, runs, and obstacle course misses represented better performance than higher scores, the

8

standardized scores were factored by -1.00 for ease of interpretation and to maintain consistency
with results from the two models shown above. The dichotomous variable, selected vs. not
selected or VW was used as the dependent variable. Table 6 shows the results of these analyses.

Table 6

Binary Logistic Regression Analyses Using Physical Fitness and SFAS Performance Event
Scores
Physical
Fitness
N B (beta) p value Nagelkerke R Odds Ratio
APFT Score 758 .614 <.001 .107 1.85
Pull-ups 600 .438 <.001 .056 1.55
1
st
ruck march 732 1.77 <.001 .326 5.85
2
nd
ruck march 581 2.32 <.001 .199 10.15
1
st
run 703 .639 <.001 .111 1.90
2
nd
run 544 .453 <.001 .061 1.57
Obstacle
Course
488 .375 <.001 .043 1.46

The results for these regressions showed that the two physical fitness variables (i.e.,
APFT score and number of pull-ups) and the five SFAS performance event variables were
significant predictors of SFAS selection. As such, all seven variables were retained for
additional analyses. The SFAS candidates whose APFT Scores and number of pull-ups were a
standard deviation above the mean of these two variables were 85% and 55% more likely to be
selected, respectively. Those whose 1
st
and 2
nd
run times were a standard deviation faster than
the mean were 90% and 57% more likely to be selected, respectively. The candidates whose
total number of misses on the Obstacle Course was a standard deviation below the mean were
46% more likely to be selected. Those whose 1
st
ruck times were one standard deviation faster
than the mean had almost a five-fold greater likelihood of being selected. The candidates whose
2
nd
run times were one standard deviation faster than the mean had more than a nine-fold greater
likelihood of being selected.
Physical Fitness Model. The standardized scores from the variables shown in Table 6
were included in a single binary logistic regression (with selected vs. not selected or VW as the
dependent variable) to determine the effect of a Physical Fitness model on SFAS selection.
This models effect was significant, (7, 468) = 76.66, p <.001; -2 Log likelihood = 464.511;
Nagelkerke R = .220. The Physical Fitness Models effect on selection was fairly robust,
accounting for 22% of the variance.

An Efficient Overall Model for Predicting SFAS Selection

All 17 variables shown in Tables 4, 5, and 6 that were significant predictors of SFAS
selection (only Consistency of effort was not) were included in a single binary logistic regression
using the same dichotomous dependent variable. This inclusive models effect was significant
and accounted for more than 39% of the variance, (15, 421) = 132.46, p <.001; -2 Log
likelihood = 357.309; Nagelkerke R = .393.


9

One purpose for this research was to determine, statistically, the most efficient combination
of variables with the greatest magnitude of predictive strength, as indicated by the amount of
variance accounted for. We entered every possible combination of variables that was
conceptually meaningful into binary logistic regressions. The following combination of
variables accounted for the most variance and is rank ordered, starting with the variable that
produced the greatest predictive strength ( [5, 727] = 274.838, p <.001; -2 Log likelihood =
731.675; Nagelkerke R = .420):

1
st
ruck march
TABE
Ambition subscale
Years of school
APFT score

This particular model can be considered most efficient because it consists of just five variables
that accounted for more variance than the regression model with all 17 variables.

Considerations for SF Recruiting

The Special Forces Recruiting Battalion (SORB) requires potential SFAS candidates to
report their General Technical (GT) scores and their most recent APFT scores at the recruiting
station. When Soldiers meet or exceed minimum GT and APFT scores, they are admitted as
potential candidates into SFAS. Upon arriving at Fort Bragg, just prior to the beginning of a new
SFAS class, Soldiers are required to complete another APFT. The Soldiers whose scores do not
meet or exceed the minimum APFT score are not admitted as SFAS candidates, but can return at
a later date to take the test again. Those who meet the APFT requirement, but whose GT scores
are lower than the minimum cut-off can request a waiver to be admitted as candidates into SFAS.

The leaders at the SWCS and SORB wanted to understand better the potential impact of
implementing more stringent requirements for SFAS admission, thus reducing the number of
lower-scoring Soldiers who are admitted. Using the data from the four SFAS classes, we
calculated the impact of cut-off thresholds on SFAS selection as if they had been enforced during
the classes from which we collected data. The impact of these cut-off thresholds were analyzed
using GT, APFT, and Grit Scale scores because these would be readily available to recruiters.

Based on conversations with SWCS and SORB leaders, the cut-off thresholds for the
lowest scoring Soldiers were set at 10%, 15%, 20%, and 25%. The percentages of SFAS
Candidates who were selected or who VW, along with the odds ratios, were calculated for each
cut-off threshold. Table 7 shows the impact of implementing more stringent SFAS admission
requirements.







10

Table 7

Potential Impact of Implementing More Stringent Requirements for SFAS Admission
Using GT, APFT, and Grit Scale Scores
Measure 10% Cut-off 15% Cut-off 20% Cut-off 25% Cut-off
Entry GT
Score
31 (47%)
Selected
35 (53%) VW
*1.13
50 (47%)
Selected
57 (53%) VW
*1.13
70 (47%)
Selected
78 (53%) VW
*1.13
79 (45%)
Selected
92 (54%) VW
*1.20
APFT
Score
12 (17%)
Selected
58 (83%) VW
*4.88
26 (24%)
Selected
84 (76%) VW
*3.17
35 (24%)
Selected
112 (76%) VW
*3.17
48 (27%)
Selected
132 (73%) VW
*2.70
Grit
Scale
Score
29 (45%)
Selected
36 (55%) VW
*1.22
45 (45%)
Selected
55 (55%) VW
*1.22
79 (47%)
Selected
88 (53%) VW
*1.13
120 (48%)
Selected
128 (52%) VW
*1.08
*Odds ratio for VW/Selected.

The results show that implementing more stringent requirements using GT and Grit Scale
scores would have prohibited the recruiting of almost half the Candidates who completed SFAS
successfully and were selected for SF training. By comparison, using APFT scores would have
been far more beneficial than using GT and Grit Scale scores for eliminating Soldiers at the
SORB who would have VW during the SFAS process. This is highlighted particularly at the
10% cut-off threshold for APFT scores where the odds ratio of VW/Selected is almost four times
greater than for GT and Grit Scale scores. (The odds ratio is referring to the likelihood that
Soldiers will VW rather than be selected, fail rather than pass).

Discussion

This research explored the strengths with which perseverance, (measured by the Grit
Scale), cognitive ability tests, physical fitness, and SFAS performance events impacted selection
of Soldiers for SF training following SFAS. Another purpose was to develop the most efficient,
conceptually appropriate combination or model of these tests for predicting selection.
Finally, we wanted to better understand the consequences of modifying SF recruiting criteria
using the Grit Scale, GT, and APFT scores, all of which are readily available to the SORB.

Predictive Strength of Perseverance

We found that four of the five Grit Scale subscales predicted selection. Results showed
also that perseverance, represented generally by a model comprised of the four subscales, was
not significantly correlated with IQ scores. This suggested that the Grit Scale added unique
value to the existing cognitive ability, physical fitness, and performance event measures in
current use at SFAS. However, the strength of prediction for perseverance, though statistically
significant, was relatively weak. As such, it would be unwise to use the Grit Scale in isolation
from the other tests and measures that accounted for most of the variance in this Soldier sample.

11

A multisource approach to gaining information about Soldier performance to inform selection
decisions is recommended.

Though the impact of perseverance on SF selection was slight, it is difficult to ignore the
fact that about 15% of the Soldiers who completed SFAS scored lower on cognitive ability and
physical fitness tests than a similar percentage of Soldiers who VW. This suggested that, at least
for some Soldiers, the drive to finish SFAS in spite of limitations in cognitive and physical
ability played a role in their eventual success. If perseverance paid off for these lower scoring
Soldiers, then we can speculate that at least some of the Soldiers who VW might have been
selected had they managed to remain firmer in their resolve to complete SFAS.

The SFAS performance events are difficult enough to tax even the fittest Soldiers. Those
who complete the process can be confident in their mental and physical toughness, and in their
ability to endure hardship. However, the toughness required of selected Soldiers to complete the
SFQC successfully is even greater, by comparison. With regard to predicting SFQC success for
our remaining sample of Soldiers, we believe that the Grit Scale will increase in predictive
strength proportional to the increase in the mental and physical demands placed on the Soldiers
who will endure the SFQC. Upon obtaining the SFQC data, we intend to test this hypothesis and
publish an addendum to this report that describes the role of perseverance in SFQC success.

Predictive Strength of Cognitive Ability Tests

The Army relies heavily on cognitive ability tests to inform decisions about recruiting,
selection, occupation, and training. The value of the cognitive ability tests used in this research
was demonstrated in their ability to predict Soldier selection and to account for a substantial
amount of the total variance in our Soldier sample. This suggested that a Soldiers cognitive
ability to maintain awareness, to think, to judge, and to adapt, played a key role in completing
SFAS and being selected for SF training.

The role of cognitive ability, though important during the SFAS process, tends to be even
more important when Soldiers begin their advanced training following the completion of the
SFQC. Leaders and trainers at SWCS report that the greatest mental challenges are presented to
Soldiers during tactical training and language classes. The Soldiers who persevered and
demonstrated physical toughness during SFAS and the SFQC, but who scored relatively low on
cognitive ability tests, will be challenged most during language and other advanced training.
Though they may attempt to persevere through these new mental challenges, it is not uncommon
for Soldiers with less cognitive capacity to be withdrawn from SF training because of their
inability to grasp knowledge, concepts, and skills that require higher levels of cognitive ability.
This underscores the importance of the cognitive ability tests to help determine longitudinal
success throughout the entire SF selection, training, and operational pipeline.

Predictive Strength of Physical Fitness Tests and SFAS Performance Events

Historically, physical fitness was thought to be among the most important factors for
determining successful completion of SFAS events and the SFQC. Our data support this
assertion. Soldiers who do not demonstrate physical toughness during the difficult SFAS process

12

simply will not prove themselves ready for subsequent SF training and military operations. It is
for these reasons that Soldiers entering SF must meet or exceed minimum standard requirements
for physical fitness.

Although all of the physical fitness and SFAS performance events in Table 6 were shown
to predict selection, the majority of the variance in our Soldier sample was accounted for by the
1
st
ruck march. The SFAS cadre and leaders report that it is during this event, more than any
other, that candidates VW from SFAS on account of insufficient physical fitness and preparation.
Those candidates who do persevere and complete the 1
st
ruck march are soon faced with a 2
nd

ruck march, during which additional candidates VW for the same reported reason.

Unlike perseverance and cognitive ability, physical fitness can be improved rather
quickly with the appropriate amount and frequency of work. Many of the candidates who VW
during SFAS have the option of returning to a class after they have improved their physical
conditioning. This is an option that allows returning Soldiers to complete SFAS and the SFQC,
while supporting the growth of the overall SF force structure.

A More Efficient Model

The recruiters, trainers, and psychologists who support SF currently use all of the tests
and measures that were included in this research. These tests and measures require substantial
amounts of time and effort to administer and interpret, but ensure that a board of leaders is well
informed before making selection and training decisions. One purpose of this research was to
determine, statistically, if a more efficient model of tests would have a similar impact on SFAS
selection compared to the current model of tests and measures. The amount of variance
accounted for by all 17 of the tests and measures combined was fairly robust. However, the
model that included the five variables discussed above (i.e., 1
st
ruck march, TABE, Ambition
subscale, Years of school completed, APFT scores) accounted for slightly more variance.

To understand how this more efficient five-variable model would have affected selection,
we calculated the impact of the four cut-off thresholds used above as if they had been enforced
during the participating SFAS classes from which we collected data (see Table 7). The impact of
these cut-off thresholds were analyzed using 1
st
Ruck March, TABE, and Ambition subscale
scores. The impact of APFT scores is shown in Table 7 above. The Years of school completed
variable was not reported because the cut-offs did not discriminate differences using the
distribution of reported years of school completed (i.e., about 9% of participants reported 11
years of school completed, whereas 52% reported 12 years completed. No one reported between
11 and 12 years of school completed). The results of these calculations are shown in Table 8.









13

Table 8

Impact of Variables from Most Efficient Model on SFAS Selection
Measure 10% Cut-off 15% Cut-off 20% Cut-off 25% Cut-off
1
st
Ruck
March
9 (12%)
Selected
69 (88%) VW
15 (13%)
Selected
99 (87%) VW
25 (15%)
Selected
143 (85%) VW
35 (15%)
Selected
194 (85%) VW
TABE 8 (12%)
Selected
59 (88%) VW
20 (19%)
Selected
86 (81%) VW
29 (21%)
Selected
108 (79%) VW
38 (22%)
Selected
133 (78%) VW
Ambition
subscale
20 (31%)
Selected
44 (69%) VW
38 (37%)
Selected
64 (63%) VW
N/A 60 (35%)
Selected
112 (65%) VW

In general, all three variables shown in Table 8 would have served recruiting and SFAS
fairly well in discriminating selected candidates from those who VW. Relative to the Ambition
subscale, the 1
st
Ruck March and TABE scores were slightly better in predictive accuracy. For
all three measures, the 10% cut-off would have prohibited the least amount of recruited
candidates who would have been selected otherwise.

From a purely statistical point of view, using the more efficient five-variable model to
screen out Candidates at the SORB who are less likely to be selected following SFAS seems
reasonable. Yet, in practice, it would be difficult for recruiters to require candidates to complete
a ruck march, administer and score the TABE and the Grit Scale, and then combine and analyze
all the scores before making final recruiting decisions. Though the 1
st
ruck march emerged as
the variable with the greatest predictive strength, it would be among the most difficult events to
execute at the SORB. On the other hand, including a ruck march at the SORB would greatly
increase the models overall efficacy to predict SFAS success.

The predictive strength of the remaining 12 variables that were excluded from the more
efficient model may increase as a function of the SFQC. This has yet to be determined. Board
leaders who make decisions about admitting SFAS Candidates into the SFQC would be better
informed by the rich information represented by all 17 variables, as opposed to having their
information limited to the reduced number of variables in the more efficient model, even though
the latter accounts for slightly more variance. The costs of eliminating information should be
weighed against the benefits of reducing the number of tests and measures by which data are
collected, analyzed, and reported.

Conclusions and Recommendations

This updated analysis of the existing SF cognitive ability, physical fitness, and
performance event measures provided empirical support for their use as valid and reliable
sources of information upon which selection decisions can be based. In addition to the tests and
measures used currently to support selection, perseverance emerged as a unique contributor,
though the magnitude of its impact was incremental, at best. Based on the results of this
research, we offer the following conclusions and recommendations:

14

Perseverance

The Grit Scale is easy to administer and score.
The Grit Scale provides an empirically-valid measure of perseverance that is
independent from all other measures.
The Grit Scale can be used for professional development as a marker of individual
levels of perseverance.
The Grit Scale can be combined with other tests and used to inform and support SF
recruiting and selection decisions.
It is expected that the value of the Grit Scale will increase when analyzed as a
predictor of SFQC success, though this has yet to be determined, empirically.
The Grit Scale should not be used in isolation from other tests and measures as a
single criterion for decision making.

Cognitive Ability

Assessed individually, the five existing cognitive ability tests and Years of school
completed serve as empirically-sound predictors of SFAS selection.
By comparison, the strength of the cognitive ability measures to predict SFAS
success is far greater than perseverance.
It is expected that the role of cognitive ability, in general, will increase during the
SFQC and subsequent advanced and language training, though this has yet to be
determined, empirically.
The TABE (Test of Adult Basic Education) emerged with greater value in
predicting SFAS success than the ASVAB-GT or Wonderlic, although greater
emphasis is typically placed on the latter two measures.
It is unlikely that the Wonderlic and Gama tests add value to predicting SFAS
selection and SFQC success independent of the TABE, D-lab, and GT.
Unless there are compelling reasons to use the Wondelic and GAMA tests to
provide information to selection boards, these tests could be eliminated without
jeopardizing the overall contribution of cognitive ability measures.

Physical Fitness and SFAS Performance Events

Physical fitness forms the primary basis upon which SFAS success rests.
Compared to perseverance and cognitive ability, the SFAS performance events and
the corresponding physical stamina required to complete them serve as the best
predictors of SFAS selection and VW.
It is expected that the SFAS candidates who are selected for the SFQC will be
required to demonstrate higher levels of physical toughness to complete the latter
course successfully. It is expected also that physical fitness will become less of an
issue during advanced training subsequent to the SFQC. Both hypotheses remain
to be supported, empirically.
Physical fitness stands out as a measureable proficiency that can be improved
rather quickly, compared to cognitive ability and perseverance, which cannot.

15

Implementing a cut-off APFT score at the SORB can produce both positive and
negative effects on the overall SF force structure. As a reliable standard for
detection, it helps to eliminate Soldiers who are more prone to VW from SFAS.
Yet, it also serves to eliminate some Soldiers who could improve their fitness fairly
rapidly, and who demonstrate higher cognitive ability and perseverance, but who
are not recruited because of immediate, insufficient levels of physical fitness.
Though the data exist, an analysis has not been conducted to determine how many
Soldiers fail the APFT, but then return later and pass it. Nor have we determined
how many of these Soldiers who return and pass the APFT go on to complete
SFAS and the SFQC successfully. Knowing these results would increase our
understanding of the role of perseverance beyond what we gain using the Grit
Scale.
The results of this research support the current emphasis on SFAS candidates
meeting or exceeding minimum levels of physical fitness. However, the number of
potential candidates who are rejected because of not meeting the APFT standard,
who do not return, yet who might have completed SFAS and the SFQC
successfully had they received a waiver, remains unknown. Our analyses suggest
that at least a small percentage of Soldiers would have completed SFAS
successfully, in spite of beginning with lower than standard APFT scores.

A More Efficient Model

The regression model that included the 1st Ruck March, TABE, Ambition
subscale, Years of school completed, and APFT score accounted for more variance
than the model of 17 variables in use, currently. Viewing this model from the
vantage point of efficiency, it requires far less time and effort to utilize than the
current model.

If the primary purpose of a model of variables is to predict an outcome with the
least amount of effort, then the more efficient five-variable model is superior.
However, if the purpose of a model is to provide a selection board with as much
information as can be made readily available, then the current model should be
perceived as more appropriate.

Other Variables Impacting Selection

This research showed that the tests and measures currently in use accounted for a fairly
robust amount of the variance in our Soldier sample. While commendable, from a practical
viewpoint, more variance still remains unaccounted for. This begs for continued development of
assessment techniques that do not restrict capable Soldiers from entering SFAS, but that restrict
those who VW or who are not selected for SF training.

In an effort to gain a better understanding for the reasons why SFAS candidates VW,
SWCS psychologists conducted exit interviews with all the candidates who VW from a recent
SFAS class (SWCS Unpublished Manuscript, 2009). Excluding medical reasons, when asked
why they VW, the vast majority stated that they were not physically prepared for the

16

performance events, which was consistent with the results reported in this paper. When asked if
they would return if given the opportunity, the majority of the candidates stated that they would
and that they believed their SFAS experience was a positive one. Additional reasons for VW
that candidates listed included the following:

Family reasons (e.g., wife not supportive, sudden illness in family, divorce, child
custody issues).
Decided not to enter SF after all.
Mentally not prepared.
Cannot gain proficiency with land navigation.
Cadre did not meet expectations of what a SF Soldier should be.
Choose an SFAS class when weather is not so hot.
SF is too different from Branch or Career Management Field.
Not sure I want to stay in Army.
SFAS was much more difficult than I expected.

Though not an exhaustive list of additional variables affecting SFAS success, it represents a
good starting point from which assessment tools can be developed and validated to improve the
predictive accuracy with which Soldiers are recruited and selected for SF. In addition to these
variables, researchers should continue to search for the determinants of SFAS success or VW
that have not yet emerged.


17

References

Brooks, J. E. & Zazanis M. M. (1997). Enhancing U.S. Army Special Forces: Research and
Applications. (ARI Special Report 33). Alexandria, VA. U.S. Army Research Institute for
the Behavioral and Social Sciences.

Duckworth, A. L., Peterson, C., Mathews, M. D., & Kelly, D. R. (2007). Grit: Perseverance and
passion for long-term goals. Journal of Personality and Social Psychology, 92, 1087-1101.

Terman, L. M., & Oden, M. H. (1947). The gifted child grows up: Twenty-five years follow-
up of a superior group. Oxford, England: Stanford University Press.

The John F. Kennedy Special Warfare Center and School (JFKSWCS) (2009). Unpublished
Manuscript. Psychological Applications Directorate, Fort Bragg, NC.