PROFESSIONAL DEVELOPMENT

Safety Management

2000 Our Strengths & Weaknesses

T
he 20th century witnessed the Research has also examined the types ment. How do safety professionals meas-
birth and growth of the safety of management systems that have led to ure their efforts and determine whether
movement around the world. catastrophes. The results can be best safety “programs” are effective? Historic-
One can trace its evolution described via the following six criteria for ally, practitioners have chosen ineffective,
from early attempts to guard safety excellence. The safety system must: inadequate and invalid measures.
machinery and address severe 1) Ensure daily proaction by supervi- In the early days of safety, accident
hazards, through passage of sors and teams which demonstrates that measures (e.g., number of accidents, fre-
workers’ compensation laws, to today’s safety is a core value of the organization. quency and severity rates) were used to
sophisticated approaches. At times, as 2) Involve middle managers as key assess progress (of a corporation, depart-
safety professionals have delved into con- players. It must require them to: ment or facility). Practitioners felt com-
cepts such as system safety, behavior- a) ensure subordinate, supervisor or fortable using these measures even
based safety, ergonomics, industrial team performance; though they offered little—they did not
hygiene, human factors and human error b) ensure quality of that performance; indicate whether the system was work-
reduction, it has often appeared that all the c) engage in actions that demonstrate ing; diagnose what was right or wrong;
bases have been covered. Despite these the importance of safety. nor indicate whether the system was in or
advances, one must ask, “To what extent 3) Require visibly demonstrated exec- out of control.
have safety professionals built systems utive action, not merely commitment. When problems with these metrics
that truly control losses?” 4) Ask for and obtain hourly involve- became obvious during the 1950s and
ment in meaningful daily activities. 1960s, safety professionals created a dif-
THE RESEARCH 5) Allow flexibility. Units and person- ferent measure—the audit. In theory, it
The research is relatively clear. It sug- nel must have options regarding what was reasoned, if a firm can dictate, in
gests that certain criteria are essential for actions they will take. advance, what actions it should take to
safety success. NIOSH studied several 6) Be perceived as positive by the prevent accidents, then it can measure
companies in a matched pair study (Cohen workforce. how well those predetermined actions
et al). National Safety Council conducted a are being executed.
study in 1967 and a follow-up study in DO TODAY’S SYSTEMS MEET THESE CRITERIA? Clearly, however, the practice of
1992 which identified safety system ele- Over the years, few methods have accepting audits as a valid measure of
ments that are used by top-performing been available to assess the true effective- excellence is questionable, unless the
U.S. companies (Plank et al; Plank and ness of safety systems. Although research audits have passed some rigorous tests. If
Fearn). Similarly, firms have historically has helped, it offers little advice to guide the audit an organization uses has been
benchmarked each other, often reaching efforts to assess where systems actually correlated to its accident record in large
similar conclusions. When one examines are in comparison to that research. enough numbers over time, it may be a
the step-change improvements made, key Perhaps the greatest problem in safety good indicator of performance; if not, it
criteria for success can be identified. has been—and continues to be—measure- should be considered suspect.
16 PROFESSIONAL SAFETY AMERICAN SOCIETY OF SAFETY ENGINEERS
 THE PERCEPTION SURVEY TABLE 1 20 Categories of a Safety System think and what works/does not work.
The perception (climate) survey is Although norms cannot be estab-
a third measure of safety system effec- lished based on overall scores from
tiveness. Such surveys have long been Communication Training surveys conducted within an array of
used in non-safety applications. Dr. Attitudes Meetings organizations (including furniture
Rensis Likert, a pioneer of this tech- manufacturers, railroads, chemical
Support Recognition
nique, used surveys to measure the and petrochemical facilities, paper
relationship among key factors of pro- Goals Inspections mills, food processing and construc-
ductivity. His research suggests that Motivation Discipline tion), the scores can be interpreted
a “high achievement” organization Regulations Investigations and shared with each organization.
generally exhibits a high degree of Scores are communicated as “%
supportive relationships; utilizes the Substance Abuse Hazard Correction positive” for each category. A survey
principles of group decision-making; Involvement Climate can measure by unit, location or craft
and has supervision in areas with Credibility Supervision within each company to reveal any
high performance aspirations. similarities/differences. By-category
Employment Supervisory Training
Attitudes toward the company, job maximum, minimum and mean
and superiors, as well as the level of scores are calculated for a number of
motivation, are also key factors. Good companies in order to provide a picture
performance in these areas results in against which an organization can com-
higher sales volume and production,
lower costs and better quality. In short,
Perhaps the greatest pare itself.
The remainder of this article discusses
Likert’s research showed a high positive
correlation between scores in these areas
problem in safety has survey data from 56 companies, which
cumulatively employ 1.657 million people
and the bottom line (e.g., profitability,
growth, return on investment).
been—and continues at all levels. These results depict how
these firms have succeeded—or failed—
Stemming from this research, safety
practitioners began to examine whether a to be—measurement. in their safety efforts as judged by the peo-
ple who truly count—hourly employees.
perception survey might work as an indi-
cator of safety system “health.” After Historically, practi- SURVEY RESULTS
many years of development and testing, A score below 70% positive in a
it has been determined that such surveys
are a much better predictor of a compa-
tioners have chosen category at the hourly employee level
suggests the need to examine the organ-
ny’s future safety performance than other
indicators tested. Perception surveys
ineffective, inadequate ization’s safety activities. Such a score
indicates that three of 10 employees do
have been found to be invaluable in diag-
nosing what actions are needed to im-
and invalid measures. not believe the system is working well. A
score below 60% positive is a red flag indi-
prove safety systems. cating that the system needs help.
Data accumulated through these sur- within an individual organization be- Analysis of this data reveals some sim-
veys may be descriptive of how truly cause it reveals just how removed man- ilarities in safety system element effec-
effective safety systems are. If a pattern agement is from shopfloor employees. tiveness. For example, as a rule, firms are
exists in the surveys of many organiza- In recent years, some patterns have not highly successful in categories that
tions over time, it might suggest that emerged—patterns which suggest that score, on average, below 60% positive
overall approaches to safety need to be safety professionals have been successful (hourly employee only):
reassessed—and perhaps changed. in some efforts, not so in others. Thus, it •recognition - 56.9% positive
A fundamental difference exists be- is time to reassess and reformulate •discipline - 58.4% positive
tween what perception surveys tell users approaches for the new millennium.
and what research or benchmarking The same can be said for categories
efforts reveal. Most safety research asks INTERPRETING THE SURVEYS that score below 70% positive:
what safety professionals or managers In recent years, many organizations •inspections - 60.0%
think works/does not work. Similarly, in have performed perception surveys pro- •supervisory training - 60.3%
the search for “best practices,” bench- vided by various sources. While it is not •substance abuse - 63.7%
marking efforts tend to ask safety profes- possible to accumulate data from differ- •employee training - 64.6%
sionals and managers what they feel ent surveys, data from companies using •quality of supervision - 65.4%
works/does not work. the same survey (the Minnesota Percep- •employee involvement - 66.4%
Conversely, perception surveys query tion Survey) can be discussed. •operating procedures - 67.6%
hourly employees. This suggests that the This survey asks 74 yes/no questions, Borderline categories include:
key reality is hourly employee percep- then clusters these questions into 20 cate- •attitudes toward safety - 70.0%
tion. Supervisor and/or manager percep- gories of a safety system (Table 1). •support for safety - 70.0%
tion is measured only to determine how Questions in each category have been sta- •management credibility - 70.0%
far from reality it is; this exercise is useful tistically validated to show what people •goal setting - 70.7%
JANUARY 2000 17
 TABLE 2 Corporate Highs & Lows
CRITERIA CATEGORIES MEAN LOW HIGH
Supervisory Recognition 56.9 40 76
Performance Discipline 58.4 29 82
Supervisory
Training 60.3 31 82
As these results indicate,
Middle- Quality of
surveyed firms fare poorly Management 65.4 42 85
Supervision
in four of the six criteria for Performance
safety excellence. When one Inspection 60.0 37 78
examines corporate highs Average 60.2 35.8 80.6
and lows, not merely Top Management Management
the means, the picture Performance Credibility 70.0 52 86
depicted in Table 2 emerges. Support for Safety 70.0 54 84
Goals 70.7 52 96
Overall, companies are rel- Operating
atively good in the following Procedures 67.6 45 82
categories:
Average 69.6 50.8 87.0
•accident investigation - 78.0% Employee Involvement of
•communication - 75.7% 66.4 59 83
Involvement Hourly Employees
•motivational programs - 73.4%
Employee Training 64.6 44 86
•new employee orientation - 72.7%
•hazard correction - 72.7% Average 65.5 51.5 84.5
•safety contacts - 72.7%
•climate - 72.4%
employee surveys only. In all cases, super- A fairly wide discrepancy was noted
Comparison of these data to the criteria visors and managers are also polled, and in perceptions about 1) employee train-
for safety success reveals that the cate- the differences in perception computed. ing, where managers think they are 16
gories of recognition, discipline, superviso- Within a good organization, the difference percent better than employees do; and
ry training, quality of supervision and is typically 10 to 12 percent; a similar dif- 2) inspections, where managers think
inspections—activities typically carried ference exists between managers or execu- they are 15 percent better than employees
out by supervisors or teams (criteria #1 tives and hourly employees. A composite do. In the remaining 18 categories, differ-
made to happen by criteria #2)—would picture shows considerably wider discrep- ences ranged between -1% and +11%.
average a score of 60.2% positive—barely ancies in these categories:
above red-flag level. This suggests that •Employee training: Supervisors SAFETY STRENGTHS
hourly employees believe their supervi- think they are 31 percent better than Composite scores indicate that compa-
sors either do not know how to satisfy employees do. nies are, in employees’ eyes, doing a good
their safety responsibilities or that no sys- •Quality of supervision: Supervisors job: investigating accidents, talking about
tem requires them to do so. think they are 25 percent better than safety; and administering motivational
The categories of management credi- employees do. and awareness programs. Some would
bility, support for safety, goal setting and •Inspections: 25 percent better. interpret this to mean that safety profes-
operating procedures, which are means •Supervisory training: Nearly 25 per- sionals spend most of their time on meet-
of judging upper and mid-management cent better. ings, reaction and gimmicks.
(criteria #2 and #3), average a score of •Accident investigation: 18 percent Supervisors generally scored many
69.6% positive. This is still below the 70% better. categories high (nine categories in the
cut-off where a firm can begin to feel •Hazard correction: 17 percent better. 80s, one in the 90s), resulting in an overall
comfortable about safety system effec- •Attitude toward safety: 14 percent score of 77.2% positive—eight points
tiveness. Criteria #4 (employee involve- better. higher than hourly workers. Upper man-
ment) scores 66.4% positive. •Support for safety: 13 percent better. agers were considerably closer to reality.
As these results indicate, surveyed •Communication: 13 percent better.
firms fare poorly in four of the six criteria •New employee orientation: 13 per- SAFETY MANAGEMENT 2000
for safety excellence. When one examines cent better. If this small sample is descriptive of
corporate highs and lows, not merely the •Management credibility: 13 percent safety system effectiveness, then these
means, the picture depicted in Table 2 better. results reflect the current state-of-the-art
(above) emerges. In seven categories, the difference was in safety. As safety professionals strive to
less than 12 percent. In the category of address apparent weaknesses, they must
DIFFERENCES IN PERCEPTION discipline, employees perceived the situ- assess the following areas.
The cited scores reflect results of hourly ation to be better than supervisors. 1) Most companies score notably low
18 PROFESSIONAL SAFETY
 Safety excellence only occurs when supervisors, managers and
executives demonstrate their values through actions and then, being
credible, ask hourly workers to help improve the system.
on recognition. This category refers to that many managers say safety is a top requires daily proaction by line managers
whether people are recognized daily (reg- priority, yet their actions (downsizing, and supervisors—a missing link that can
ularly) for doing a good job and working outsourcing, overtime) say otherwise. only be corrected when the system holds
safely. This is a measure of whether people 7) Employee involvement is another these managers, supervisors and execu-
are positively reinforced, which is known area of concern. It received a composite tives accountable.
to be the best way to foster safe behaviors. score of 66.4% positive, which suggests Research and benchmarking clearly
Recognition was the worst-rated cate- that at least one-third of the total work- indicate where safety performance should
gory in more than 41 percent of companies force wishes to be more involved. be. Surveys reveal where performance lev-
surveyed. The composite score of 56.9% 8) The differences in perception, par- els actually are. As these data show, a large
positive indicates that nearly one-half of ticularly between supervisors and hourly discrepancy exists between the two. 䡲
the workforce feels it is being ignored. workers, have created a void. To be suc-
2) Not only are employees ignored cessful leaders, supervisors must be close REFERENCES
when they do a good job, they are also to their workers—or at least understand Bailey, C. “Managerial Factors Related to
ignored when they engage in unsafe acts. each worker’s needs; to motivate, super- Safety Program Effectiveness.” Professional
Discipline was the lowest category in 23 visors must understand “where their Safety. Aug. 1997: 33-35.
percent of those firms surveyed—receiv- workers are.” Bailey, C. “Improve Safety Program Per-
ing a composite score of 58.4% positive. In 11 of 20 categories, composite scores ception.” Professional Safety. Oct. 1993: 28-32.
In this context, discipline refers not only indicate that a major chasm exists between Bailey, C. and D. Petersen. “Using Percep-
to punishment, but also to whether peo- supervisors and workers. The primary tion Surveys to Assess Safety System Effec-
ple are allowed to work unsafely without problem involves the daily performance of tiveness.” Professional Safety. Feb. 1989: 22-26.
Cohen, H., R. Smith and S. Cohen. “Char-
being corrected. supervisors. In short, they simply are not
acteristics of Successful Safety Programs.”
3) When one considers that it has been as effective as they think they are. (In fair-
Journal of Safety Research. 1978.
30 years since passage of the OSH Act, ness, this is usually not their fault; their Likert, R. The Human Organization. New
with its emphasis on physical conditions, plates are more than full.) York: McGraw Hill, 1967.
it is disconcerting that inspections (the Petersen, D. Safety Management. Des
mechanism used to improve physical CONCLUSION Plaines, IL: American Society of Safety Engi-
conditions) is rated the worst category in These results suggest that many com- neers, 1998.
11 percent of companies and received a panies have in some way “missed the Petersen, D. Techniques of Safety Manage-
60.0% positive response from employees. boat” on safety. Historically, firms have ment. Des Plaines, IL: American Society of
In other words, a red flag. taken what can be called an “islands of Safety Engineers, 1998.
4) Supervisory training was rated the safety” approach. To comply with laws Planek, T., G. Driesser and F. Vilardo.
worst category in 11 percent of companies, and standards, they create various pro- “Industrial Safety Study.” National Safety
with an overall score of 60.3% positive. grams (islands), such as a process safety News. Aug. 1967.
5) These four categories (recognition, program, a lockout program, a fall pro- Planek, T. and K. Fearn. “Reevaluating
discipline, inspections and supervisory tection program, a HazMat program, an Occupational Safety Priorities: 1967 to
training) are solid indicators of a serious ergonomics program. By creating these 1992.” Professional Safety. Oct. 1993: 16-21.
problem in current approaches to safety “islands,” however, a company establish-
management. That problem? The perfor- es no main channel of solid management
mance of supervisors, middle managers performance—where everyone from
and teams. Add the quality of supervision CEO to first-line supervisor takes some Dan Petersen, Ph.D., P.E., CSP, is a consultant
category, with its 65.4% positive score, and action each day which reflects that safety specializing in safety management and organ-
it becomes clear that supervisory perfor- is a core value. izational behavior. He holds a B.S. in Industrial
mance is notably weak. Overall, these five As a result, many people come to Engineering, an M.S. in Industrial Psychology
categories received an average score of believe that these islands are “safety.” and a Ph.D. in Organizational Behavior. A fre-
60.2% positive—borderline red flag. They are not. These programs are impor- quent author and speaker, Petersen is a profession-
al member of ASSE’s Arizona Chapter and a
These scores reflect one of two condi- tant components of the overall safety member of the Society’s Management Division.
tions: 1) Supervisors do not know what it effort, but when a company believes that
is they are supposed to do with respect to they satisfy corporate safety responsibili-
safety—a training problem; or 2) no sys- ty, trouble is on the horizon.
tem requires them to take these actions—
an accountability problem. In the
As these data reveal, many safety
efforts have lost their focus. The true
READER FEEDBACK
author’s opinion, accountability for safe- focus should be integrated safety, not Did you find this article interesting
ty is a major problem throughout U.S. individual programs that staff can create, and useful? Circle the corresponding
industry. Managers, supervisors and thus relieving the line organization of its number on the reader service card.
teams simply are not held accountable for responsibilities.
safety performance. Safety excellence only occurs when YES 25
6) Management credibility does not supervisors, managers and executives SOMEWHAT 26
fare too well either. The categories that demonstrate their values through actions
comprise this criteria received an overall and then, being credible, ask hourly
NO 27
score of 69.6% positive. This indicates workers to help improve the system. This
JANUARY 2000 19