njury Prevention'

Personality Types& Injuries

A statistical study and effective strategiesBy F. David Pierce

F. David Pieree, M.S.P.H., CSP,CIH, is a professional member andtwo-time president ofASSE's Utah

Chapter. He has been involved inthe SH&E profession for more than

30 years, fulfilling many rolesin corporate management, safety

and environmental leadership,academia and research. His current

professional focus is directed atdiscovering the true causes ofsuccess, injuries and behavior.

PERSONALITY TYPES are receiving increasedattention within the occupational safety and healthprofession as a possibly significant injury-causingfactor. As a clarifying and constructive effort for thisexpanding debate, the study described in this articleused best statistical means and comparative casestudy approaches to find answers to three progres-sive questions:

1) Is an individual's personality type a causal orassociated factor to increased or decreased injuryexperience?

2) If a statistically significant correlation existsbetween certain personaGty types and increased in-juries, do like organizations or companies experi-ence the same or similar injury experience becauseof them?

3) If a strong correlation exists between personal-ity types and increased injuries in some companiesbut not in others, what successful injury preventivestrategies are used?

This study found that a statistically significantcorrelation exisis between a certain personality typeand increased injuries but that some companiesstudied had implemented successful preventivemeans which seemed to negate this asstxiation.

Personality and its importance in many aspects oflife is certainly not a new subject. Personalit}''simpact on the safe behavior of individuals and on thesafety performance of organizations seems to be,however (Olivares; Cooper). How much do SH&Epractitioners truly know about personality theoryand how does personality apply to injury risk? If per-

sonality has a detrimental impact on anorganization's safety performance (associ-ated with higher injury experience), howcan safety and health programs minimizeor control this affect? With the economicfuturists' warning of a thinning workforcein the near future, how can the safety of allemployees {regardless of personality type)be positively impacted—while negatingthe need to even consider applicant per-sonality screening as a way to reduce thenumber of at-risk individuals cor\sideredfor employment?

A good starting point is to level the "knowledgefield" concerning personality typing. To do a decentjob in this expanding safety area of application andtheory, one needs a basic understanding of personali-ty theories and modeling, and especially their history.

Introduction to Personality TypologiesIdentifying differences in people is as old as

mankind. Aristotle {384-322 BC) wrote about the dif-ferent kinds of people who attended the OlympicGames, separating tiiem by the roles they chose toplay Nicolo Machiavelli (1469-1527) also dissecteddifferent personalities, dividing people by the waythey thought. Arthur Schopenhauer (1788-1860)explored the metaphysical aspects of personality.

CaH JungTo a large extent, the scientific and systematic

concepts of personality truly started with the workof Sigmund Freud (1856-1939). His concepts neverreally gained impetus until Swiss psychologist CarlJung (1875-1961) carefully studied Freud's work andbegan to structure a foundation of personality theo-ry (Jung).

Like Freud, Jung believed in the existence of the"collective unconscious," which he called "domi-nants, imagos, mythological or primordial images."Jung later used the term "archetype." He then gaveeach archetype a name that characterized behaviorsof individuals having that archetype. Jung theorizedthat some inner-wiring exists in each individualwhich forms an "organizing principle on the thingswe see and do." He further theorized that this inner-wiring began early in life and became basically unal-terable throughout the individual's life.

Jungian typology, as it was termed, was derivedfrom the theory that basic personality "functions"could be used to describe how each individual firstperceived things, then from that information, howthey made decisions. In essence, Jung created a four-square personality grid that was defined (top axis)by what he felt was the most important binary divi-sion—introversion versus extroversion—and by asecondary binary division (side axis)—^being ration-al versus irrational. Each of the grid's four squares

42 PROFESSIONAL SAFETY MARCH 2005 www.asse.org

was further divided in half by two additional sets offunctions—thinking versus feeling, which wereascribed to the grid's rational band—and sensationN'ersus intuition, which were assigned to the irra-tional band. Jungian typology, therefore, includedeight different personality types or archetypes.

Myers & BriggsBuilding on Jung's work, personality theory

moved into high gear with the expansive work ofKatharine Briggs and Isabel Briggs Myers (Myersand Briggs). TTieir epic work formed the foundationof one of the best-recognized personality typologiesused today Differing from Jung, Myers and Briggsbegan not by establishing a primary function but bytreating all as equals. They also added another bina-ry function—judging versus perceiving—thus creat-ing 16 different personality types (4 x 4).

They furthered the theory of a developmentsequence and timeline for superior, secondary, terti-ary and inferior functions. Their sequential develop-ment "stack" of functions included: 1) the flow ofenergy (from inside or outside—introversion versusextroversion); 2) how information is absorbed—viasensing or by intuition); 3) how decisions aremade—using thinking or feeling; and 4) how theindividual deals with the external world on a day-to-day basis—using judging means or perceivingmeai\s. This became known as the Myers-Briggstypology.

Myers and Briggs also developed a testing means,the Myers-Briggs Type Indicator (MBTI) to determinethe personality type of individuals (e.g., ESTJ, ESTP,INFJ, INFP). This test has been used for many years—primarily in academic and research settings.

Like Jung, Briggs and Myers visited the importantquestion concerning where personality functionscome from, which was expanded by the work ofW. Harold Grant. All suggested that personalities arebuilt step-by-step, beginning at a young age and fin-ishing at or near maturation. During this process, a"boilerplate" personality encoduig was created.

The primary complaints lodged against theMyers-Briggs approach are significant and two-fold.Becaused it originated in a research environmentand was developed and validated almost totally inacademic settings, the typing and tool have beengenerally shunned by those seeking practicalapproaches and answers. In other words, the per-sonality typing is interesting but not applicable oruseful in the real world.

Furthermore, only limited descriptive correla-tions have been drawn between Myers-Brigg's fourfunctions or 16 personality types and actual individ-ual behavior patterns. In other words, testing says ateam of four employees has an ESTJ, an ISFP, anINTP and an ESTP, but what can be expected ofthem and how can this information be used to makethis team stronger or to make decisions better?

Keirsey & BatesThis was certainly not the end of this Freud-initi-

ated evolution of personality theory. David Keirsey

Principles <»fPersonalities1st PrincipleThere exist liistinct and definable personalitytypes in pec^ple that influence the way theyrelate to the world and to others, and the waythey absorb and process information.

2nd PrincipleEach person's personality type can be deter-mined accurately by a testing means.

3rd PrinciplePersonality types are developed very early inlife and remain principally intact throughout aperson's entire life.

and Marilyn Bates took a different scientificapproach, dissecting personality into two sides: tem-perament and character (Keirsey and Bates).Temperament is first developed early in life and ispredisposition—that is, precharacter. According toKeirsey and Bates, "temperament is the brain's hard-ware and character is the brain's software"—oneobviously set in cement and the other modifiable.Keirsey and Bates viewed Myers-Briggs four typesas "dimensions of temperament."

Keirsey also created a testing means, the KeirseyTemperament Sorter, to identify the personality ofindividuals and to divide them into the four tem-peraments: rationals, idealists, artisans or guardians.This naming of temperaments was a throwback toJung's character descriptions (archetypes), but thisremains only a minor thrust of this typology. Themajor thrust of the Keirsey-Bates approach has beento connect or align the different temperaments withwell-known historical figures.

To many, however, although Keirsey's work iscertainly interesting, it seems to fail the "so what?"test. In other words, from a practical perspective, ifone individual aligns with same temperament asDwight Eisenhower did and another aligns with thetemperament that Mother Teresa had, then "sowhat?" How does knowing these various tempera-ment aligrunents help an organization or team suc-ceed, build relationships or define expectedbehavior patterns? , ,

The EnnegramAnother personality typology also exists that at

first seems non-Freudian-based. Usually referred toas the Ennegram (or Enneagram), this theoryemphasizes psychological motivations as the root ofpersonality. It has its roots in ancient Sufi (ancientPersian myslics) traditions that were describedaround the tum of the 20th century by GeorgeGurdjieff (1866-1949) and emerged via the writingsof Claudio Niiranjo. However, it was not until DonRichard Riso wrote Personality Types in 1987 that the

www.asse.otg MARCH 2005 PROFESSIONAL SAFETY 43

Personalitytheory is

interesting,but is itof value

to SH&Epractitioners?

How canat-risk

personalitiesbe identified

and howcan that

informationbe used to

improvesafety and

healthprograms?

Ennegram got much notice {Riso}. Since then, manyresearchers have contributed to this theory and pro-vided a nine-type matrix based on "long-termdrive" matched against "short-term drive," which intheory diagnoses one's emotional outlook on lifefwww.enneagraminstitute.com).

These nine personality types include the re-former, the helper, the motivator, the romantic, thethinker, the loyalist (sometimes called the skeptic),the enthusiast, the leader (also called the confronterand the boss) and the peacemaker (sometimesreferred to as the mediator). As with other Freudian-based personahty theories, most agree that theEnnegram is "fixed" in each individual early in life.

Two primary problems are encountered in apply-ing the Ennegram personality model to practicalapplications, however. First, it fails the "so what?"test. It may be interesting but it does not seem to bepractical or applicable. Second, no accepted and val-idated personality-testing tool is available. Althoughthe RtffiTI test and the Essential Ennegram Test areavailable, both struggle with determination accuracy.What is significant about the Ennegram is that itreturns personality typing back to Jung's archetype-like thiriking in that character descriptions are usedto describe personalities.

Correlating Personality to BehaviorsPersonality theory is interesting, but is it of value

to SH&E practitioners? How can at-risk personali-ties be identified and how can that information beused to improve safety and health programs? Thesequestions lie at the crux of tlie "so what?" issue formany reasons. Theoretically, on the conservativeside, personality is a major contributor to behavior.Viewing this relationship from the other side of thependulum, personality may well be a predisposing,individual hard wiring for behavior.

Factually, from a safety perspective, the hardwiring of the subconscious (personality) plays amajor role in automatic (auto-pilot) or default deci-sion making and, thereby, is a major influence on anindividual choosing at-risk or risk-avoidance behav-iors, compliant or noncompliant behaviors. How canpersonality be correlated with automatic or defaultbehavior and decision making? Furthermore, howcan those individuals with at-risk personalities beidentified with certainty? Finally, what can be doneto mirumize or control both behavior-<aused risk-taking and the resulting increased injuries?

This is where current work in psychology coupledwith currently used business concepts and tools canhelp. Keirsey began an interesting divergence in the-ory progression when he simplified the MBTI andidentified four fundamental temperaments. Healigned this approach with those used by Hippoc-rates (430-377 BC), Immanuel Kant (1724-1804) andFriedrich Nietzsche (1844-1900). This divergencestarted an effort to align temperaments with defin-able and predicable behaviors.

More recently, this line of correlating personalitytheorv has been described in The Hero and the Outlaw:

Building Extraordinarif Brands Through the Pmvcr ofArchet\fpes (Mark and Pearson). Tliis book focuses onbuilding products tbat appeal to specific personalitytypes due to hard-wired preferences and decision-making patterns. Unfortunately, as this effort pushedtoward adding practicality and applicability to per-sonality types, it offered no testing process or tool.

Social Style of IndividualsThe most significant bridge between personality

and behavior has become known as the social styleof individuals. Begun more than 25 years ago, thisconcept also begins with four basic "temperaments,"but instead of focusing on identifying differences,this approach focuses on understanding them. It rec-ognizes an early development of social style (per-sonality) hard wiring as well as the existence of bothprimary and secondary social styles.

Social styles were first identified and correlatedwith behavior by Merrill and Reid. Thrtiugh thework of two primary sources, the Tracom Group(www.tracomcorp.com) and Wilson Learning Corp.(www.wilsonleaming.com). an extensive amount ofvalidation research on social styles has been accom-plished over the past 20 years—much of it ftxrusedon practical business applications. This scientific yetbusiness-focused approach provides a personalitytyping approach that easily passes the "so what?"test because personality is tied to behavior and deci-sion-making patterns. Additionally, several sourceshave developed highly validated tests that willdetermine both the primary and secondary socialstyles of individuals with great accuracy.

Two of the associated behavioral opposites iden-tified within the social style grid are; risk-taking ver-sus risk-avoiding, and ttiinking (determine actionsby carefully and meticulously thinking thingsthrough) versus feeling (going with the flow, auto-matically reacting). Grounded fundamentally in per-sonality theory, these two behavior extremesprovide an ideal approach for use in this studybecause they help identify individuals who are bothrisk-takers and feeler decision-makers (reactors).Individuals witb these two traits are the personali-ties who would be most likely to choose to hangglide off El Capitan, drive fast, play chicken withreal knifes or be more accepting of higher-risk situa-tions (Merrill and Reid). \n addition, these risk-tak-ing automatic-reacting individuals are identifiableusing simple and validated accurate tests.

Study MethodologyTwenty-five employers, representing moderate-

to higher-injury-potential industry sectors, wereselected and asked to participate in a study thatsought a statistical correlation between worker per-sonahty types and injury performance. As a study-specific deviation from classical cohort studyformats, no matching of employers was attemptedin that the different social styles existing within eachemployer's workforce would constitute a matchingpopulation in itself. Only employees of study em-ployers whose jobs were involved directly with

44 PROFESSIONAL SAFETY MARCH 2005 www.asse.org

product manufacture, warehousing and/or distri-bution, or those involved directly in providing serv-ices were included in the study and were tested forindividual social styles. The rationale for selectingonly these employees was to focus primarily on jobsthat were most at risk of injury due to increasedexposure to hazards.

A social styles assessment tool developed byAlliance for Training Inc. (Salt Lake City) was used.Tliis tool was selected based on four factors: 1) easeand quickness of the assessment prtKess; 2) ability toself-grade and determine social styles on-slte in realtime; 3) extensive validation of the tool with othersocial style assessment tools successfully used formore than 20 years; and 4) lower expense.

The results of the social style testing provideddata pools that were divided into four definablesocial styles or temperaments:

•driver (risk-takers and deep-thinkers);•analytical (risk-avoiders and deep-thinkers);•amiable (risk-avoiders and feet ing-reactors);• expressive (risk-takers and feeling-reactors).Theoretically, based on characteristic behavior

(hard-wiring) factors associated strongly with eachsocial style, one would expect (the null hypothesis)that expressive individuals would have the highestpotential for injuries and analytical individuals thelowest, leaving individuals with driver and amiablesocial styles somewhere in between. For this reason,expressive and analytical individuals became thecomparative focus populations in this study.

Each study company provided detailed injurystatistics and data that covered in some cases morethan 10 years. All injury data were initially used. Alldata—both social style testing results and the pro-vided injury data from employers—used an individ-ual identifying means, a simple numerical coding.This coding was comprised of a leading alphabeticalcharacter (A through Y) to distinguish the individ-ual's place of employment, followed by the last fourdigits of the individual's social security number(e.g., H3209). This coding was used to eliminate anypossibility of compiler-bias that may arise concern-ing differing gender, race or other factors.

Data ConsiderationsAs the study progressed, several data considera-

tions were identified, including the lack of consistentinjury data periods reported by participatingemployers, possible data impacts caused by employ-ee turnover, and possible impacts of personality-caused work-preference issues.

Thirteen (52 percent) of the study employers pro-vided injury data that extended beyond 10 years,while the remaining employers reported data forperiods between five and nine years. The concernwas not knowing what impact these differenceswould have on the power of the statistical inferencesmade using all acquired data. In most compilers'minds, this deviation provided too many unknownsand inserted possible biases into the data pool.Therefore, all injury data past five years were ex-

cluded from the data pool and all employee testswith employment longevity past five years wereexcluded from the analyses.

Most statistical studies (e.g., epidemiologicalstudies) have captive or identified populationswhere all necessary data for comparison can beacquired. One major issue in this case was that theinjury data covered many years, but personality typ-ing (social style) involved only those currentlyemployed by the study companies. Three concernswere identified,

1) Turnover is a normal ongoing component ofany business. The greater the time period covered,the greater the potential impact of turnover.

2) Depending on industry sector, managementvalues, work and work environment issues, andorganizational culture, turnover rates varied widelyamong employers—and were not calculated orknown by many employers in the study.

3) In addition to and paired with turnover, theinfluence of different personality tynpes on both vol-untary and "for cause" termination was unknownand highly suspected as being a potential data bias.

The major unanswered questions regarding thistliird concern were two-fold. First, does having a risk-taking personality raise the probability or individualacceptability of using voluntary termination (quitting)as a means of dealing with job, pay or other work-related dissatisfactions above that of a risk-avoidingpersonality? Second, does having a risk-taking and/orfeeling-reactor personality raise the probability that anindividual will be in\'oluntarily terminated (fired) dueto personality conflicts, job perfonnance or decision-making issues beyond that of those with a risk-avoid-ing and/or deep-thinking personality'?

To address these concerns, compilers made the fol-lowing decisions. With respect to concerns 1 and 2, itwas determined that their only impact on the data andanalyses would be to reduce the number of individu-als caused by their leaving employment during thefive-year data period and, therefore, would have noinjuries assigned to them within the study. This wouldlead to fewer injuries being assigned to the ftx'uscdsocial styles, \vliich would generate softer inferences;it would definitely not result in more injuries beingassigned or in more powerful inferences.

With respect to the third concern, the compilersgenerally agreed that individuals who possess risk-taking and/or feeling-reactor personalities would bemore probable candidates for both vokmtary andinvoluntary termination. However, because thefocus of this study was on the association betweeninjury experience among individuals with risk-tak-ing and feeling-reactor personalities, the only possi-ble impact would be that fewer injuries would beassigned to the focused group.

The compilers also concurred that all possibleimpacts from these concerns would be to reduce thestatistical gap between populations in the study.Therefore, these possible impacts were noted but notconsidered in the data and inferences.

The possible impacts or data biases caused by per-

www.asse.org MARCH 2005 PROF£SSIONAL SAFETY 45

Expressive Individuals: Injuries v. Longevity

•o

u

12

11

10

J S

107] \0S} [03] [03]1 2 3 4

Years of Employment (At-Risk Period)

sonality-caused work-preference issues was also dis-cussed. The question was whether an individual'spersonality type contributes either substantially orpartially to his/her choice of a job and decision tokeep that job? In other words, is a risk-taker more like-ly to choose and stay in a job that has a greater hazardpotential or contact than a risk-avoider? The compil-ers felt this question was most likely true based on thewell-dtKumented higher prevalence of analytical per-sonalities in occupations such as accountants andengineers and the higher prevalence of driver person-alities in management. Tlierefore, it was agreed thatthis personality-caused work-preference issue maywell be a significant source of bias in a study of alltypes of jobs. Howe\'er, this study only includedworkers employed in hands-on production, ware-housing, distribution and service roles. Based on this,it was determined that personality-caused work-pref-erence issues would most likely provide only a minoror possibly a negligible impact on data and inferences.

Study FindingsTJI total, the 25 participating companies reported

more than 10,500 injuries from 1999 to 2003. Thereported injuries included first-aid cases throughmore-severe incidents. Additionally, more than10,000 employees were given personality tests (socialstyle typing) to determine their dominant social style(personality). Only those employees who testedexpressive (risk-taking and feeling-reacting)—a pop-ulation of 2,771 individuals—and those who testedanalytical (risk-avoiding and deep-thinking)—1,429individuals—were included in the study and statisti-cal analyses.

46 PROFESSIONAL SAFETY MARCH 2005 www.asse.org

Al! injuries thatoccurred over theprevious five yearswhich were report-ed by study compa-nies were thencompared againstthe individuals in-cluded in the studyby matching identi-fying codes. Re-gardless of injuryseverity, a largedata matrix wascreated in whichindividuals werepooled first byyears of employ-ment (at-risk peri-od); the number ofinjuries tliese work-ers had experi-enced over theirat-risk period wasthen used to sepa-rate them.

Figures 1 and 2plot total injuriesexperienced versus

at-risk period (longevity of employment) for thosewho tested expressive and those who were found tobe analytical. The resulting regression lines shown ineach figure establish the best-fit linear correlations(statistical relationship) between the maximumamplitudes for each year's distributions.

One need not use complex statistical tools todetermine that the linear regression lines for bothplots (expressive versus analytical) are very differ-ent. In fact, the resulting regression line for injuryexperience of the expressive study population(Figure 1) is about double the slope of the analyticalinjury experience line (Figure 2). Simply trani^lated,this study strongly infers that those individuals whohave expressive social styles (at-risk and feeling-reactors) are nearly twice as likely to experienceinjuries as those individuals who have analyticalsocial styles (risk-avoider and deep-thinker). In fact,the plotted results \'alidate what many years of safe-ty experience would surmise to be true: Withoutcontrols or with inadequate controls, risk-takershave more injuries than risk-avoiders.

Figures 3 and 4 show comparative distributionsof injury experience at two different points. Figure 3shows a comparison of injury distributions for twoyears of employment exposure while Figure 4 showsthe same for five years of exposure, for both expres-sive and analytical individuals.

Comparing arithmetic means for each distribu-tion at two years (Figure 3) showed 3.97 injuries (anannual injury experience of almost two injuries perperson on average) for expressive individuals versus1.91 injuries (an annual injury experience of about

one injury per person on aver-age) for analytical individuals.Additionally, the +2 standarddeviation (-\-2 sigma) ranges forthese distributions were 0.65 to7.29 injuries for individualswith expressive scxrial stylesand 0.00 to 5.33 for the graphedanalytical injury distribution.

According to the statisticsof distributions, these shouldrepresent the range of injuriesinto which 95 percent of allindividuals with expressive oranalytical social styles, givenan infinite database (and allother factors being the same).Translated, this means thatgiven any representative pop-ulation of employees in high-er-risk work settings, it wouldbe unreasonable to expect thata risk-taking and reacting indi-vidual could go without injuryfor two years; at the same time,it would be reasonable toexpect them to have as manyas four injuries each year.

Comparing the arithmeticmeans for each distribution atfive years (Figure 4) showed6.12 injuries (an annual injuryexperience of more than oneInjury per person on average)for expressive individuals ver-sus a value of 3.74 injuries (anannual injury experience ofmuch less than one injury perperson on average) for analyti-cal individuals. Additionally,the +2 standard deviation (-f-2sigma) ranges for these distri-butions were 0.92 to 11.38injuries for individuals withexpressive social styles and0.00 to 8.18 for the graphedanalytical injury distribution.

According to the statistics ofdistributions, these should rep-resent the range of injuries intowhich 95 percent of all individ-uals with expressive or analyt-ical social styles, given aninfinite database (and all otherfactors being the same). Againtranslating the statistical infer-ences from this data, it wouldbe reasonable if not a near-cer-tain expectation for a workerwith a risk-taking and reactingpersonality to have at least oneinjury each year and also rea-

Figure 2

Analytical Individuals: Injuries v. Longevity

.£ 9

— s"(5

[M] m

Ml

Ml m

[28] [20] RTl-2 3 4

Years of Employment (At-Risk Period)

Figure 3

Distribution of Injury Experienceat Itoo Years Exposure

40

•aI 30

0S 20

2

I _ I Expressive IndividualsArithmetic Mean = 3.97Two Sigma Range = 0.65 to 7.29

I I Analytical IndividualsArithmetic Mean = 1.91Two Sigma Range = 0.00 to 5.33

4 6 8Total Injuries Experienced

10 12

www.asse.org MARCH 2005 PROFESSIONAL SAFETY 47

Figure 4

Distribution of Injury Experienceat Five Years Exposure

30

•o

e 20

I10

I j Expressive IndividualsArithmetic Mean = 6.12Two Sigma Range = 0.92 to 11.38

I I Analytical IndividualsArithmetic Mean = 3.74Two Sigma Range = 0.00 to 8.18

n8

Total Injuries Experienced

10 12

sonable to expect this worker to average more thantwo injuries per year if s/he worked five years.

Figure 5 shows the total injuries experienced byyears of exposure (total injuries divided by the num-ber of years of exposure) for both expressive andanalytical individuals identified in the study. Best fitlinear regression lines for both data pools are alsoshown. This data presentation appears to show thatas employment longevity increases, individualshave fewer injuries. That is, their injury rates aregreatest in tlieir first years of employment and loweras they gain experience, adapt to the company's cul-ture and expectations and/or are more impacted bythe company's injury-preventive strategies. For ex-ample, using the expressive individual's regressionline, the percentage of injuries decreases 35 percentfrom year one to year five.

One might argue that this 35-percent decreasecould also be explained by a near-linear turnover rateof around nine percent each year. This also wouldaccount for fewer counted and assigned injuries ineach year past the first. Htjwever, this argument failsto explain the near parallel nature between theexpressive and analytical regression lines. If it is trulymore probable for expressive individuals to termi-nate employment voluntarily or to be terminateddue to personality {as one would conjecture fromexperience), one would expect the lines to not be par-allel. Instead, one would expect to see a statisticalclosing of the data each year with it reaching the clos-est point in year five. This would be the result of theslope of the analytical line being less than that forexpressives. However, this Is clearly not the case. Theparallel nature of both regression lines led compilers

48 PROFESSIONAL SAFETY MARCH 2005 www.asse.oig

to suggest that tliis reductionin injuries with longevity wastruly an obser\'fd plienome-non in tliis study (Figure 5).

Injury Performance:Individuals withExpressive Social Styles

This stLidy .iltempted tocompare the injury perform-ance of expressive individu-als in each of the 25 firms thatparticpated. Tliis comparisonltx>ked strictly at the calculat-ed injury rates of those em-ployees identified to haveexpressive social styles todetermine whether ct)nsisten-cy existed between employ-ers. Figure 6 sbows thedistribution of injury ratesfrom companies with the bestexpressive injury rates to theworst. This low-to-high dis-tribution shows a centralinjury rate plateau (range 32to 68 with an average at 47)that described more than half

of the participating companies. Figure 6 also showsthat five companies had much lower injury rates(range 4.5 to 21 with an average at 9.5) for theiremployed expressive individuals and that threecompanies had much Hgher rates (range 81 to 100with an average at 93) for this same social style.

Impact of Management & CultureFrom an injury pre\ention locus, these results

demanded the answer to a constructive question thatlaid at the purpose of this study. Why did five compa-nies show not only lower injury rates for expressiveemployees, but much lower rates? What commoninjury prevention strategies were used by these com-panies or what other common management systemelements were absent or minimally used/effective inthe other companies included in the study?

Of significant note, primary differences were ob-served in management philosophy; this was immedi-ately noted by the researchers when identifyingemployers for inclusion in the study. The largest potilof employers was primarily interested in the study'sfindings as a means of validating personality causali-ty for injuries. A smaller pool viewed their involve-ment as a way to measure or baseline their efforts anddiscover ideas for improving their safety programs.

Perhaps predictably, these two approaches weredistinguished by their total injury performance andinjury experience with expressive workers, Thoseemployers with interest in personality causality hadhigher injury rates, while those focused on measure-ment and ideas had much lower injury rates.

These observed and correlated differences led theresearchers to visit the issue of company culture, andin particular the work of Don Eckenfelder (a; b). He

Figure 5

Itotal Injuries Per Year of Employment TVendsY

ear

Per

•o

cV

a.MLUin0}'w3

'F"m1

240

220

200

180

160

140

120

100

80

60

40

20

"^ Expressive Individuals

(O) Analytical Individuals

2 3 4

Years of Employment (At-Risk Period)

Figure 6

Distribution of Injury Rates for Expressive Workers by Firm100

80

60

40

20

WM

O

V

Employers in Study (in order of expressive injury rates)

www.asse.org MARCH 2005 PROFESSIONAL SAFETY 49

has established a high correlation between positivesafety cultures and safety excellence. He has also provided a means of measuring safety culture (the SafetyCulture Barometer) that was used to evaluate themanagement system differences between study com-panies with excellent and poor safety performance.

From a search for these common strategiesamong study companies with excellent safety per-formance among potentially at-risk employees, thefollowing 10 commonalities were identified.

•These companies had more than just an execu-tive management commitment to safety. They had amandate for safety by which all management wasmeasured and held accountable.

•The safety program emphasized leading indica-tors (injury prevention efforts and activities) versustrailing indicators (injury rates and statistics).

•The workforce was fully involved in the safetyprogram and day-to-day safety activities (somecompanies had such deep involvement that no for-mal safety committee existed).

•An active recognition and celebration programwas in place to accentuate safe performance andsafety participation.

• An aggressive quality improvement programwas in place; it included process control, quality con-trol and the use of statistical process controls.

•The organization had an equal dedication toother associated efforts such as efficiency and costminimization.

• Ail open communication climate was evident.Almost everything was shared with employees andevery employee had open communication pathwaysdirectly to the top.

•A strong and ongoing investment was made inbuilding employee skills and knowledge.

•Ethnic and cultural differences were seen as astrategic advantage. These differences were appreci-ated and celebrated by management.

•The culture was built on strong personal rela-tionships between management and employees.Mutual trust was also strong.

ConclusionsTlie data from this extensive case study establish-

es that a strong potential exists for causality orassociation between risk-taking and reacting perstm-alities and increased injury experience. This associa-tion appears to make individuals with suchpersonalities twice as likely to have injuries as theirrisk-avoiding deep-thinking counterparts. Addi-

tionally, the datainfer that it is rea-sonable to expectthat workers withrisk-taking andreacting personali-ties will have at

Your FeedbackDid you find this articleinteresting and useful?Circle the correspondingnumber on the readerservice card.

RSCn Feedback39 Yes40 Somewhat41 No

least one injuryper year and couldwell have as manyas four injuries peryear unless highly

effective prevention strategies are implemented.Further, the data inter that all workers—regardlessof personality type—experience fewer injuries peryear on average the longer they are employed.

A significant finding of this case study was thatwide differences existed in injury rates of risk-takingand reacting workers among the 23 participatingcompanies. Review of injury preventive strategiesused by those with lower injury rates validated thepositive impact of safety-culture-re la ted strategiessuch as executive leadership, high participation,effective leading measures, and a commitment tocommunication, building skills and trust, strongrelationships and gaining strength from workforcedifferences.

This study's findings has three possible uses. To amajor extent, these are a function of the companyculture/management system into which they wouldbe used and by the primary safety strategies/beliefsof the SH&E practitioner.

1) Tlie statistical correlation between at-risk per-sonalities and increased injuries might be used to val-idate safety emphasis or applicant selection/screeningfocused on individuals with such personalities. It mayeven support the use of perscmality testing of employ-ees and applicaiits as a means of reducing injuriesand/or the potential for injuries in these companies.

2) Second, the commonalities found among thosecompanies with excellent injury performanceamong at-risk personality employees could be a sig-nificant and important supportive argument forimproving a company's management systemand/or changing the culture with respect to safetyand highly related business aspects. In other words,the SH&E practitioner may be able to cite these com-monalities and their correlation to reducing injuriesto support the argument that significant changes areneeded in a company's culture.

3) The commonalities noted could serve as a val-idation for companies and SH&E practitioners whohave already embraced the importance of safety intheir cultures and perhaps add some new ideas forimprovement. •

ReferencesCooper, D. "Psychology, Risk & Safety: Understanding How

Personality' & Perception Can Influence Risk Taking." ProfessionalSafety. Nov. 2003: 39-46.

Eckenfelder, D. "Getting the Safety Culture Right " Occupa-tional Hiizardf. Clct. 2(X)3:

Eckenfelder, D, Values-Drivefi Safety. Rockviile, MD; Govem-menl institutes, 1998.

Jung, C.G. Pmliahgical 1]/pcs. Bollingen Series XX, Vol. 6.Princeton, NJ: PrinceUin University Press, 1971,1976.

Keirsey, D. and M. Bates. I'lcofi- ilmU-r^tmni Me: Character &Toiiih-rawcnt Ti/pcf-. Del Mar, CA: Promethfus Book, 1984.

Mark, M. and C.S. Pearson. The Hero ami the Outlaw: BuildingExtraoniinanf Bnmdi Through the Power of Archetypes. New York:McGMw-HiU, 2001.

Merrill, D. and R, Reid. Personal S/i//('s and EffectivePerforniaiicc. Boat R.iton, FL: CRC Press, 1999,

Meyers, I. and K. Briggs. Myt'rs-Bny,!;? Type Indicator, PaloAlto, CA: Consulting Psychologists Press, 1999.

Olivares, O.J. "Selecting for Safety." Occupational Health &Safety. Sept. 21X13:46-48.

Riso, R.R. PtTsoiialily Tyfts. New York: Houghton Mifflin, 1987.

50 PROFESSIONAL SAFETY MARCH 2005www.asse.org