dimensions of interpersonal relationships and safety in the steel industry

DIMENSIONS OF INTERPERSONALRELATIONSHIPS AND SAFETY IN

THE STEEL INDUSTRY

George W. WatsonBloomsburg University of Pennsylvania

Dow ScottLoyola University Chicago

James BishopNew Mexico State University

Treasa TurnbeaughMarsh Risk Consulting

ABSTRACT: This paper extends theories explaining the influence of socialdeterminants on workplace safety. Specifically, we applied social capital theory’semphasis of trust, shared norms, and faithfulness to obligations to the outcomesof at-risk behavior and perceptions of a safe work environment. Data provided by395 employees of a major steel company supported the hypothesis that sharedemployee norms predicted both perceptions of work environment safety andat-risk behavior, trust in supervisor predicted perceptions of a safe work envi-ronment, and belief in management’s safety values predicted at-risk behaviors.

KEY WORDS: workplace safety; social capital theory; safety culture.

Injuries and death resulting from workplace accidents remains oneof the most costly factors, both personally and financially, in transactingbusiness today. In 2001, there were over 3.9 million disabling injuriesand 5,300 deaths in American workplaces1 (National Safety Council,2003) costing over $132.1 billion—a figure exceeding the combined profits

Address correspondence to George W. Watson, Department of Management, College ofBusiness Administration, Bloomsburg University of Pennsylvania, Sutliff Hall Room 235,400 East Second Street, Bloomsburg, PA 17815. E-mail: [email protected].

This study received funding from Marsh, Inc. in support of their continuing interest increating safer more productive work environments. We would like to thank William Grimes,Dean Larson and Dennis Morajda for their support and contribution to this research.

1These figures do not include the events of September 11, 2001.

Journal of Business and Psychology, Vol. 19, No. 3, Spring 2005 (�2005)DOI: 10.1007/s10869-004-2230-2

303

0889-3268/05/0300-0303/0 � 2005 Springer Science+Business Media, Inc.

reported for the top ten Fortune 500 companies in the same year.Moreover, costs to those injured include lost wages, medical costs notpaid by the employer, potential limits on future employment andadvancement, and restricted ability to enjoy personal pursuits.

Though the causes of industrial accidents are many and varied,the leading cause is the at-risk behavior of employees themselves.While the direct importance of at-risk-behavior is made clear by thegeneral principle that people who engage in risky conduct are morelikely to be injured, injure others, or cause damage, the degree of riskassociated with the working environment is also important. Employeeswho perceive that their work environment is safe are likely to havemore positive attitudes toward their employer than those who do not(Hoffman, Morgeson, & Gerras, 2003). Given that at-risk behavior andwork environment safety are important issues in today’s workplace,then if follows that antecedents related to them are also important.Social capital theory (SCT) suggests several social dimensions of anorganization’s environment that both influence individual behavior andshape perceptions of the work environment (Coleman, 1988). Thesedimensions include employees’ direct supervision, normative examplesof co-worker practices, and management policy and practices (Geller,1991). Although some research has emerged addressing some socialfactors within these dimensions, their combined and relative effectshave not been adequately addressed (Zohar, Dov, Luria, & Gil, 2004).With this in mind we chose factors we judged to be of particularimportance to safety related behavior and perceptions of work envi-

ni tsurTrosivrepuS

ytefaS tnemeganaMseulaV

rekrow-oCsmroN ytefaS

ksir-tAroivaheB

kroW deviecrePtnemnorivnE

ytefaS

ξ3

ξ2

ξ1

η1

η2

γ 11

γ 21

γ 22

γ 32

1X

2X

3X

1Y

2Y

λ 11

λ 22

λ 33

λ 11

λ 22

δ1

δ2

δ3

ε1

ε 2

)53.(

)12.(

)24.(

)33.(

)42.(

Figure 1Hypothesized Model

+To simplify the presentation of the model, correlations among the exogenousvariables are not shown. Numbers in parentheses are values for constrained

paths

JOURNAL OF BUSINESS AND PSYCHOLOGY304

ronment safety: specifically, trust in supervisor, co-worker safetynorms, and management safety values.

The purpose of this study, then, is to employ SCT to proposerelationships between safety related components of the organization’ssocial environment and at-risk behavior and perceived work environ-ment safety and test these relationships using an appropriate sample.To accomplish our purpose, we first review several approaches to theproblem of workplace safety. We then articulate an approach usingsocial capital theory and use this theory to hypothesize relationshipsamong our safety related social factors and outcomes. See Figure 1.Finally, we test our hypotheses and discuss the implications of theresults and potential directions for future research.

APPROACHES TO WORKPLACE SAFETY

Historically, there have been four major approaches to improvingsafety in the workplace. The first approach has been to focus on thework environment itself: making it physically safer. The major thrust inthis approach has been to revise policies, redesign jobs, and engineerand ergonomically design tasks with the human factor in mind (Ram-sey, 1995). In recent studies, the importance of these factors has re-ceived increased attention commensurate with the surfacing ofcumulative trauma disorders, such as carpel tunnel syndrome (Benci-venga, 1996). In fact, HR researchers and practitioners expect newOSHA regulations regarding ergonomics to influence the way jobs aredone in several major industries (OSHA, 2000). The ergonomic approachemphasizes improving the fit between human physiological capacitiesand workplace demands. The psychological and sociological aspects ofworkplace safety are de-emphasized in an effort to better match jobrequirements with the behaviors that humans can perform safely andcomfortably.

A second focus has been to identify traits in employees that are morelikely to lead to accident. The underlying question is whether there arehuman traits that make some people more prone to accidents than oth-ers. In particular, neuroticism and extroversion are hypothesized to de-tract from an individual’s vigilance toward matters of safety (Hansen,1988; Smith & Kirham, 1981). Empirical findings support the link be-tween extroversion and ‘‘accident proneness’’ while the link betweenneuroticism and ‘‘accident proneness’’ is supported only in cases of ad-vanced neuroticism.

In contrast with the job design and trait models, a third approachattempts to convince employees to behave more safely. Behavioralmodification models of safety emphasize that specific conduct can be

G. W. WATSON, D. SCOTT, J. BISHOP, AND T. TURNBEAUGH 305

encouraged by organizations that enhances safety related outcomes. Forexample, Geller (1990) argues that management can demonstrate avisible commitment through the institutionalization of rewards andsanctions that enhance safe conduct. In addition, Salzer-Azaroll andDe Santamaria (1980) found positive behavior changes when safetyfeedback systems were integrated into the daily routine of production.Salzer-Azaroff, Loafma, Merante, and Hlavack (1990) found that multi-dimensional models including feedback, reinforcement processes andrepetitive goal setting have positive effects on the behaviors that improvesafety-related outcomes.

A fourth general framework for understanding safety improvementis from the interpersonal or social perspective. The central question inthe social approach is: what interpersonal factors enhance safety in theworkplace? In an apparent foreshadowing of Goldberg, Dar-EL, andRubin (1991), Zohar (1980) investigated the role of safety climates. Safetyclimate is considered one dimension of a multi-dimensional organiza-tional climate. Safety climate reflects employees’ perceptions about therelative importance of safe conduct. Perceptions of safety climate,according to Zohar, pivot on four dominant factors: perceived relativeimportance of the safety committee, importance of safety training, workpace, and the perceived status of the safety officer. Clarke (1999) arguedin a related vein—that organizational activities such as elevating thestatus of the safety officer and improving interpersonal communicationscan foster a positive safety culture. Zohar (2002), examined managerialbehavior in the form of safety related feedback and found that as safetygoal interactions increased, perceptions of the safety climate improved,specific safety behaviors increased (ear-plug use), and injury rates de-creased. In sum, Zohar’s contributions highlight the importance ofstructural factors in enhancing perceptions of the safety climate andreducing injury rates.

The role of the organizational culture has also been investigated forits effects on safety related perceptions and behaviors. O’Toole (2002)concluded that employee perceptions of the safety system are linked tomanagement’s perceived commitment to safety, which in turn, is linked toinjury rates. In addition, Clarke (1999) found that accurate inter-groupperceptions regarding safety issues were essential to the building of trustbetween levels of the organization.

Other social-theory based models have also been proposed. Geller(1999) for example, outlined the positive effects of interpersonal trust inthe establishment of effective observation and feedback systems. He ar-gued that trust is an important component in the open reception anddelivery of safety related performance feedback. Also, Roberts and Geller(1995) developed the approach of ‘‘actively caring’’ to describe a morecommunitarian organization safety atmosphere. Actively caring cowork-


ers behave in ways that optimize safety for themselves and theircoworkers. Geller (1991) proposed that three psychological factorsunderlie the formation of an atmosphere of active caring: positive self-esteem, a sense of belonging, and the social conditions of empowerment.

Social exchange theory has also formed a foundation for research insafety. Hoffman and Morgeson (1999) found significant links betweenleader–member exchange (LMX) and perceived organizational support(POS) to safety-related communications, perceptions of safety commit-ment and ultimately, accidents. For POS the authors used a generalized(non-safety specific) scale essentially tapping the extent to which theorganization was perceived as considerate toward and cared about themember. LMX was also measured with a generalized scale pertaining tothe perceived quality of the relationship the member has with thesupervisor. Hoffman, et al. (2003) examined the relationship betweenLMX and safety role definitions as moderated by the safety climate. Theyfound the LMX-Safety Role definitions were moderated by safety climateand that safety role definitions were positively related to safety citizen-ship behaviors. Their concept of safety climate included several dimen-sions: management attitude toward safety, social status, safety behavior,and safety reward. In this study the measures of these dimensions werecombined into one score.

A SOCIAL CAPITAL MODEL OF WORKPLACE SAFETY

While research in workplace safety has matured, new social contexttheories that emphasize the combined effects of multi-dimensional fac-tors in interpersonal relationships have emerged. One such theory, SCT,shows promise in explaining the influential processes that help indeveloping safe working environments or cultures (Nahapiet & Ghoshal,1998). Following the interpersonal dimensions of social capital presentedby Nahapiet and Ghoshal (1998), we examine the relationship of trust,shared norms, and management’s safety values with the safety relatedoutcomes of at-risk behaviors and perceptions of workplace safety.

Social capital is among the most recent in a series of theories ininterpersonal capital that have developed over the past 40 years. Eachtheory considers the social aspects of organizational life and insists thatinterpersonal associations provide dividends to the people that invest inthem. Furthermore, each theory claims that collectives fostering theseinterpersonal associations also derive benefits from them. The nature ofthese dividends varies among the theories as does the normative deter-mination about what constitutes an adequate organizational return foran individual’s contributions (Portes, 1998). Several of the more influ-ential of these theories include the predecessors of social exchange theory(Blau, 1964), human capital theory (e.g., Coleman, 1988), and psycho-


logical contracts (for a review see Rousseau, 1995). (For a more detailedperspective see: Belliveau, O’Rielly, & Wade, 1996; Burt, 1997; Coleman,1998; Chung & Gibbons, 1997; Friedman & Krackhardt, 1997; Fukuy-ama, 1995; Leanna & Van Buren, 1999; Pennings, Lee, & Van Witte-loostujin, 1998; Putnum, 1995.)

Relational Dimensions of Social Capital

Putnum (1995), argued that ‘‘we must sort out the dimensions ofsocial capital, which is clearly not a unidimensional concept’’ (1995,p. 72). Nahapiet and Ghoshal (1998) following Putnum’s admonition,crafted several theoretical dimensions they believe comprise socialcapital that include structural, cognitive, and relational dimensions. Ofparticular interest to us is their conceptualization of the relationaldimension. This dimension includes factors that characterize the nat-ure of interpersonal relationships within the organization. Specifically,these factors include meeting obligations, trust, and norms. As ourhypotheses unfold we assert that these factors are related to at-riskbehavior and perceived work environment safety. Thus, unlike previ-ous research we examine the simultaneous effects of these relationaldimensions on the safety outcomes of at-risk-behaviors and perceivedwork environment safety.

In general, obligations represent a commitment to undertake a duty.Rousseau’s (1995) psychological contract research emphasized theimportance of meeting implied or explicit expectations of employees.Although many organizations undertake a sincere effort to ensure a safework environment, it is not uncommon for some managers to simply paylip service to safety (Gavin, 1986). Because managers establish policies,fund safety programs, and oversee employee conduct, their impact onsafety is both pervasive and profound. Moreover, their engagement inmeeting the obligations of a safe working environment are likely toinfluence worker behavior and attitudes.

Indeed, employees’ perceptions of management’s sincerity and val-ues toward creating a safe workplace will be reflected in their perceptionsof the work environment. We believe that risky behavior, whether re-quired due to job tasks and design, or attributable to work pace, or simplyenacted through personal choice, is less likely to occur in organizationswhere management sincerely values safe conduct. In addition, if man-agement in perceived as insincere then employees may reasonably seeonly two alternatives; exit, or a fatalistic acceptance of whatever conse-quences occur. Neither of these options reduces risky behavior and may,at times, encourage it. Therefore:


Hypothesis 1: Managements safety values are negatively related toat-risk behavior.

Trust also is present in the relational dimension of Nahapiet’s andGhoshal’s (1998) model of social capital. The relational dimension stressesthe significance of interpersonal relationships in the facilitation ofmutually advantageous exchange. Interpersonal trust (e.g., Lewicki,McAllister, & Bies, 1998; Sheppard & Sherman, 1998) has been at thecenter of theorizing about cooperative and productive interaction for sometime (e.g., Barnard, 1938; Blau, 1964). Indeed, for sociologists, trust iscentral to successful social cooperation (e.g., Homans, 1964). Within thiscontext, and because we can never be assured that the effort we make tohelp others will ever be reciprocated, we are obliged to trust that they willbe (Blau, 1964). Coleman (1988, P.S103) argues that social capital dependsupon this trustworthiness and further concludes that: ‘‘without a highdegree of trustworthiness among the members of the group, the institutioncould not exist.’’

The importance of trust and its inter-relationships with othermanifestations of social capital, such as safe work environments, is notyet fully understood. In the classical sense trust is defined as a con-fidence in the integrity, character, ability, and truth of another(American Heritage Dictionary). In the organizational sciences trust islikely to be understood as an individual’s willingness to be vulnerableto others (Rousseau, Sitkin, Burt, & Camerer, 1998). The psychologicalpredisposition toward personal vulnerability can influence workplacesafety in at least two ways. First, Geller (1999) linked interpersonaltrust to the constructive feedback vital in sustaining and improvingsafety related behavior. Workers, Geller argues, must feel safe inreceiving and giving criticism of other’s safety related behavior. Sec-ond, shared perceptions of distrust can inspire active resistance to theestablishment of oversight programs. People will avoid, if possible,being evaluated by those they do not trust to make honest, unbiasedassessments of their behavior.

We believe that the psychological assurance of trusting that one’ssupervisors will act on behalf of one’s health and welfare will enhanceone’s perceptions of work environment safety. Similarly, trusting thatsupervisors will not politicize the safety program or goals for their owngains will also positively enhance perceptions of a safe work environ-ment. Therefore;

Hypothesis 2: Trust in supervisor is positively related to perceivedwork environment safety.


Norms constitute the relational factors by which people in organi-zations govern themselves and their interface with others (Nahapiet &Ghoshal, 1998). These frameworks generally inform members about whatis considered right or wrong, good or bad. Shared normative frameworksare second only to trust in their broad recognition as an underlying factorin the formation of social capital. In fact, when a norm is effective ‘‘itconstitutes a powerful, though sometimes fragile, form of social capital(Coleman, 1988, P.S104).’’ Shared norms have also been linked to moreeffective interpersonal cooperation (Starbuck, 1983), and are significantfactors in the development of trust because they help us to interpret andunderstand co-workers’ behavioral intentions, and predict their attitu-dinal and behavioral responses (Doney, Cannon, & Mullen, 1998).

For example, Roberts and Geller found that the norm of ‘‘activelycaring’’ enhanced the climate of organizational safety. Within this con-text, actively caring ‘‘… takes the form of looking for environmentalhazards and unsafe work practices and implementing appropriatecorrective actions …’’ (1995, p. 53). Such shared norms should be stronglyrelated to the organization’s safety climate.

Hence we posit that safety norms, those that govern expectations forothers’ behaviors and establish boundaries for appropriately safe per-sonal conduct, are powerful organizational assets and should be reflectedin reduced-at-risk behavior. Furthermore, norms that shape what isappropriate and inappropriate conduct should also enhance one’s per-ceptions of work environment safety. Therefore:

Hypothesis 3: Co-worker safety norms are positively related to per-ceived work environment safety.

Hypothesis 4: Co-worker safety norms are negatively related toat-risk behavior.

We also believe that certain relationships in our hypothesizedmodel should be compared with others. In particular, it is important toexamine the relative strength of the relationships that safety norms ofco-workers has with the dependent variables to those of the otherindependent variables. This variable is particularly important sinceco-worker interaction is among the most salient features of theworkplace. Field theory asserts that individuals’ reactions to anenvironment are determined to a great extent by both the proximityand salience of the perceived elements (Mathieu & Hamel, 1989). Mostof the impact an organization has on its members is derived jointlyfrom people with whom members associate in the course of organiza-tional activities and from the tasks they perform (Porter, Lawler, &Hackman, 1975). Furthermore, workers are usually in the presence of


their co-workers more than their supervisors or upper management.Hence, they are exposed to co-worker norms continually and receiveconstant feedback as to how their own behavior conforms with orvaries from established norms. This suggests that,

Hypothesis 5: The relationship between co-worker safety norms andperceived work environment safety is stronger than that betweentrust in supervisor and perceived work environment safety(c12 > c11).

Hypothesis 6: The relationship between co-worker safety norms andat-risk behavior is stronger than that between management safetyvalues and at-risk behavior (c22 > c23).

METHOD

Sample and Survey Administration

This study was conducted in the sheet rolling division of steelmanufacturing plant owned by a Fortune 100 company in the Midwest. Atotal of 408 production employees took part in the survey, which wasapproximately 49% of the Division’s employee population. This crosssection consisted of those employees who were randomly pulled fromproduction during a one-week administration period. This cross-sectionwas reflective of the demographics of the workforce in which the majoritywere men (approximately 85%) with an average age of 48. Approximately80% had finished high school and 3% had some college. The sample wasproportionally representative of all facilities and departments in thedivision. The work involved the use of cranes to dip coils of steel into vatsof chemicals for the finishing process.

Table 1Means, Standard Deviations, Correlations, and Measures of Reliability among the

Variables (Coefficient Alphas are on the Diagonal)

Construct M SD 1 2 3 4 5

1. Trust in supervision 3.85 1.20 (.76)2. Co-worker safety norms 4.47 .94 .18** (.76)3. Management safety values 3.35 1.27 .43** .14* (.74)4. Safe work environment 4.16 1.03 .43** .27** .32** (.69)5. At-risk behavior 2.60 .95 ).17* ).38** ).23** ).23** (.73)

Note: * p < .01; ** p < .001.


Employees left the main manufacturing area in groups of 20–40 tocomplete the survey in a company conference room. An outside researchteam administered the survey, and no managers or supervisors were inthe room when the survey was completed. The survey was confidentialand employees were paid for the time it took to complete it. A total of 13surveys had to be dropped because they were incomplete.

Table 2Factor Loading of Scale Items

Factor Loadings

Variables and Items 1 2 3 4 5

Trust in Supervisor1. If I make a mistake, my supervisor is not

willing to ‘‘forgive and forget’’ (R).63

2. My supervisor is friendly and approachable .733. My supervisor can be trusted .82

Safety Norms of Co-workers1. If we see someone working in a unsafe

manner, members of my work crew willremind them to work safely

.71

2. Members of my work crew try to improvesafety in our work area

.86

3. Members of my work crew almost alwayswear their safety equipment

.43

4. If members of my work crew noticed asafety hazard, they will take correctiveaction

.67

Belief in Management’s Safety Values1. Management will compromise safety for

increased productivity (R).53

2. Management does not listen to employees’suggestions for improving safety (R)

.73

3. The only thing that matters to Manage-ment is getting the job done (R)

.84

Perceptions of a Safe Work Environment1. My work area is very safe .692. I have the protective equipment I need to

work safely.56

3. The equipment in my work area is wellguarded

.72

At-risk Behavior1. I always follow the safety rules .6632. I take short-cuts to get my work done (R) .693. I often break safety rules to get the job done(R)

.64

4. I never take risks at work .58


Measures

The participants responded to items measured by six-point Likertscales. The response options ranged from ‘‘strongly disagree’’ (1) to‘‘strongly agree’’ (6). The means, standard deviations, correlations, andreliabilities of our scales are reported in Table 1. All items appear inTable 2.

Analysis

Prior to testing our hypotheses, we performed a confirmatory factoranalysis (CFA) on the 17 items of our scales. The indices used to assessmodel fit were the root mean square error of approximation (RMSEA), thenon-normed fit index (NNFI), and the comparative fit index (CFI). These fitindices are recommended based on sample size and number of parametersestimated (Gerbing & Anderson, 1992; Medsker, Williams, & Holahan,1994; Rigdon, 1996). The results of the CFA indicate that the measurementmodel fit the data well: v2

(109) ¼ 242.47; RMSEA ¼ .056; NNFI ¼ .92;CFI ¼ .93. All items loaded significantly on their intended factors. SeeTable 2.

RESULTS

In order to test the hypothesized model (See Figure 1), a covariancematrix was used as input to LISREL 8.54 (Joreskog & Sorbom, 1993).

Trust inSupervisor

Management SafetyValues

Co-WorkerSafety Norms

At-riskBehaviour

Perceived WorkEnvironment

Safety

-.46**

-.23**

+.54**

+ .25**

ξ3

ξ2

ξ1

η1

η2

R2= .42

R2= .27

Figure 2Standardized Path Estimates for the Hypothesized Model

**p < .01, one-tailed


Following the procedures outlined by Settoon, Bennett, and Liden (1996),we created manifest indicators for each latent construct by averaging theitems for each scale. Creating single indicators from the scales enhancesthe subject-to-degrees-of-freedom ratio and allows a more rigorous test ofthe structural portion of our model. Because, a covariance matrix wasused as input, we set the error variance for each manifest indicator to theproduct of the variance of the scale, and the quantity one minus thereliability of the scale. The values to which the error variances were setappear in Figure 1. The exogenous variables were allowed to correlate.Figure 2 displays the completely standardized path coefficients for therelationships in the model. Completely standardized path coefficients arereported because of their suitability in comparing relative contributionsexplained variance (Bagozzi, 1980).

Direct Relationships

The fit indices for the hypothesized structural model werev2

(3) ¼ 3.86, RMSEA ¼ .027; NNFI ¼ .99, and CFI ¼ 1.00. These resultsindicate that the data fit the hypothesized model well (Medsker et al.,1994). Based upon the strength of the path coefficients, all hypotheseswere supported (Hypothesis 1: c23 ¼ ).23, t ¼ )3.59, p < .01; Hypothesis2: c11 ¼ .55, t ¼ 8.49, p < .01; Hypothesis 3: c12 ¼ .25, t ¼ 3.86, p < .01;Hypothesis 4: c22 ¼ ).46, t ¼ )7.23, p < .01).

Comparative Relationships

In order to test comparative relationships we used the alternativemodels approach (Joreskog, 1993) in which we constrained the paths tobe compared to be equal and employed the v2 difference test to determineif the differences were significant. Hypothesis 5 was not supported sincec11 > c12. On the other hand, Hypothesis 6 was supported when c22 > c23

(|).46| > | ).19| and the difference was significant, Dv2(1) ¼ 5.16,

p < .05.

DISCUSSION AND CONCLUSION

Accidents that result in death, injury, and property damage areimportant problems in the workplace. Although safety has been ap-proached from a variety of different frameworks, new and more com-prehensive theories can assist us to better understand and managesafety issues. This study extends theories that argue that characteristicsof the organization’s social environment are related to workplace safety.To accomplish this we investigated the relationships between specific


relational factors of social capital have with at-risk behavior and per-ceived work environment safety. We found support for the relationshipbetween co-worker safety norms and both at-risk behavior and perceivedwork environment safety. In addition, trust in supervisor was related toperceived work environment safety and management safety values wasrelated to at-risk behaviors.

The results illustrate the importance of management polices andpractices that emphasize safety, and equally important, the degree towhich management follows through on the implementation of suchpolicies and practices. In addition, the degree to which employeesbelieved that management is truly committed to safety was signifi-cantly related (negatively) to the degree to which they engaged inbehavior that would put them at risk of an accident. Studies of first-level leadership have produced a plethora of theories and empiricalfindings that emphasize its importance. The results of this studyfurther support the notion that trust in first-level leadership is sig-nificantly related to perceptions of a safe working environment(Hypothesis 2). Like leadership, the behavior and norms of one’s peershas received much study and its importance has been well established.Our results add to the breadth of this knowledge with respect to thesafety dimension. Safety norms of co-workers were significantly relatedto both perceptions of a safe working environment (Hypothesis 3) andat-risk behaviors (hypothesis 4), respectively.

This study, similar to other field research, has limitations. Onelimitation is that the subjects are mostly male (85%). Although there isno reason to believe this condition has detrimental effects on generaliz-ability of the study, it should, nonetheless, be noted. As with all types ofstudies of this type, common method variance, or mono-method bias, is aconcern. That is, it can be argued that relationships among the variablescould be inflated since they were all taken from a single source, indi-vidual employees. The nature of the constructs, as well as the theoreticalexplanations of their relationships, made it necessary to assess thevariables from the perspective of the same individual. As individualratings are required to assess support, commitment, and felt responsi-bility, we could not use ratings from other sources to control for methodvariance. Spector claims self-reports can be useful for deriving insightsabout how people feel about and react to their jobs as well as under-standing relationships among feelings and perceptions. He notes that‘‘the reasonableness of using self-reports depends upon the purpose of thestudy’’ (1994, p. 387).

There are several directions to take SCT with regard to safety. First,it would be fruitful to understand the relationships between social capitaland actual accident rates and costs. This study may reveal more aboutthe importance of the social climate relative to workplace safety. One of


the realities of organizational life is that not all supervisors or managersare proactive about safety. Social capital’s role in combating fatalism andensuring the presence of worker voice in the safe production processwould be a worthwhile investigation. Because organizations differ con-siderably, the importance of the dimensions of social capital are likely tovary from one organization to the next. For example, trust in one’ssupervisor or co-workers may be critical in the steel industry, but per-haps not so important for college professors, where the environment isnot as threatening to one’s health. Social capital’s role in helping toovercome resistance to change, or its role in creating resistance to changealso can be of importance. Positive social capital may enhance the pos-sibilities of improvements, but negative social capital—that is, socialcapital that encourages pockets of shared behaviors and perceptions thatare contrary to the collective’s best interests—may inhibit organizationalimprovements. Furthermore, there are likely to be cross-cultural differ-ences in the formation and maintenance of social capital and how thesedifferences affect management approaches to problems and opportuni-ties. In sum, SCT appears to be a promising vein of research in our effortsto reduce accidents and foster a safe working environment.

REFERENCES

Bagozzi, R. P. (1980). Causal modeling in marketing. New York: John Wiley & SonsBarnard, C. (1938). The functions of the executive, Cambridge, MA: Harvard University

Press.Belliveau, M., O’Rielly, C., & Wade, J. (1996). Social capital at the top: Effects of social

similarity and status on CEO compensation. Academy of Management Journal, 29(6),1568–1593.

Blau, P. (1964). Exchange and power in social life. New York: Wiley.Bourdieu, P. (1977). Outline of a theory of practice. Cambridge, UK: Cambridge University

Press.Burt, R. (1997). The contingent value of social capital. Administrative Science Quarterly, 42,

339–365.Chung, L., & Gibbons, P. (1997). Corporate entrepreneurship: The roles of ideology and

social capital. Group and Organization Management, 22(1), 10–30.Clarke, S. (1999). Perceptions of organizational safety: Implications for the development of a

safety culture. Journal of Organizational Behavior, 20(1), 185–189.Cobb, A., & Wooten, K. (1998). The role social accounts play in a ‘‘justice intervention.’’ In

R. Woodman & W. Passmore (Eds.), Research in organizational change and develop-ment (Vol. 11, pp. 243–295) Greenwich, CT: JAI Press.

Coleman, J. (1988). Social capital in the creation of human capital. American Journal ofSociology, 94(Suppl), 95–120.

Craske, S. (1968). A study of the relationship between personality and accident history.British Journal of Social Psychology, 41, 399–404.

Doney, P., Cannon, J., & Mullen, M. (1998). Understanding the influence of national cultureon the development of trust. Academy of Management Review, 23(3), 601–620.

Eysenck, H. (1962). The personalities of drivers and pedestrians. Medicine, Science and theLaw, 3, 416–423.

Fine, B. J. (1963). Introversion–extroversion and motor vehicle driver behavior. Perceptualand Motor Skills, 16, 95–100.


Friedman, R., & Krackhardt, D. (1997). Social capital and career mobility. Journal of Ap-plied Behavioral Science, 33(3), 316–334.

Fukuyama, F. (1995). Trust: The social virtues and the creation of prosperity. New York:Free Press.

Gavin, J. (1986). Observations for a long-term, survey-guided consultation with a miningcompany. Journal of Applied Behavioral Science, 21(2), 201–220.

Geller, E. (1990). Performance management and occupational safety: Start with a safetybelt program. Issues in performance management, New York: The Haworth Press.

Geller, E. (1991). If only more would actively care. Journal of Applied Behavior Analysis, 24,607–612.

Gerbing, D. W., & Anderson, J. C. (1992). Monte Carlo evaluations of goodness of fit indicesfor structural equation models. Sociological Methods and Research, 21, 132–160.

Goldberg, A., Dar-El, M., & Rubin, A. (1991). Threat perception and the readiness to par-ticipate in a safety program. Journal of Organizational Behavior, 12(2), 109–122.

Hansen, C. P. (1988). Personality characteristics of the accident involved employee. Busi-ness and Psychology, 2(4), 346–365.

Hoffman, D., & Morgeson, F. (1999). Safety-related behavior as a social exchange: The roleof perceived organizational support and leader–member exchange. Journal of AppliedPsychology, 84(2), 286–296.

Hoffman, D., Morgeson, F., & Gerras, S. (2003). Climate as a moderator of the relationshipbetween leader–member exchange and content specific citizenship: Safety climate as anexemplar. Journal of Applied Psychology, 88(1), 170–178.

Homans, G. C. (1964). Social behaviour: Its elementary forms, New York: Harcourt Brace.Joreskog, K. G. (1993). Testing structural equation models. In K. A. Bollen & J. S. Long

(Eds.), Testing structural equation models. Newbury Park, CA: Sage.Joreskog, K. G., & Sorbom, D. (1993). LISREL 8: Structural equation modeling with the

SIMPLIS command language. Hillsdale, NJ: Lawrence Erlbaum Associates.Kawachi, I., Kennedy, B., Lochner, K., & Prothrow-Smith, D. (1997). Social capital, income

inequality and morality. American Journal of Public Health, 87(9), 1491–1498.Leanna, C., & Van Burren, H. (1999). Organizational social capital and employment

practices. Academy of Management Review, 24(3), 538–555.Lewicki, R., McAllister, D., & Bies, R. (1998). Trust and distrust: New relationships and

realities. Academy of Management Review, 23(3), 438–458.Mathieu, J. E., & Hamel, K. (1989). A causal model of the antecedents of organizational

commitment among professionals and nonprofessionals. Journal of Vocational Behav-ior, 34, 299–317.

Medsker, G. J., Williams, L. J., & Holahan, P. J. (1994). A review of current practices forevaluating causal models in organizational behavior and human resource managementresearch. Journal of Management, 20, 439–464.

Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organiza-tional advantage. Academy of Management Review, 23(2), 242–285.

National Safety Council, (2002). Report on Injuries in America, 2001. Website Resource.URL:http//:www.nsc.org/library/rept2000.htm

O’Toole, J. (2002). The relationship between employee’s perceptions of safety and organi-zational culture. Journal of Safety Research, 33(2), 231–243.

Pennings, J., Lee, K., & Van Witteloostujin (1998). Human capital, social capital, and firmdissolution. Academy of Management Journal, 41(4), 425–440.

Portes, A. (1998). Social capital: Its origins and applications in modern sociology. AnnualReview of Sociology, 24, 1–24.

Portes, A., & Landolt, P. (1996). The downside of social capital. The American Prospect.May–June, 1996, 18–26.

Putnam, R. (1995). Bowling alone: America’s declining social capital. Journal of Democracy,10(January), 24–35.

Rigdon, E. E. (1996). CFI versus RMSEA: A comparison of two fit indexes for structuralequation modeling. Structural Equation Modeling, 3, 369–379.

Roberts, D., & Geller, E. (1995). An actively caring model for occupational safety. AppliedPreventive Psychology, 4(1), 53–59.


Robinson, S. (1996). Trust and breach of the psychological contract. Administrative ScienceQuarterly, 41(3), 574–599.

Rousseau, D. (1995). Psychological contracts in organizations: Understanding written andunwritten agreements. Thousand Oaks, CA: Sage.

Rousseau, D., Sitkin, S., Burt, R., & Camerer, C. (1998). Not so different afterall: A crossdiscipline view of trust. Academy of Management Review, 23(2), 393–404.

Settoon, R. P., Bennett, N., & Liden, R. C. (1996). Social exchange in organizations: Per-ceived organizational support, leader–member exchange, and employee reciprocity.Journal of Applied Psychology, 81, 219–227.

Shaw, L. (1971). The validity and usefulness of the concept of accident proneness. In L.Shaw & H. S. Sichel (Eds.), Accident proneness, New York: Pergamon.

Sheppard, B., & Sherman, D. (1998). The grammars of trust: A model and general impli-cations. Academy of Management Review, 23(3), 422–437.

Smith, D. J., & Kirkham (1981). Relationship between personality characteristics anddriving records. British Journal of Social Psychology, 20, 229–231.

Spector, P. E. (1994). Using self-report questionnaires in OB research: A comment on theuse of a controversial method. Journal of Organizational Behavior, 15, 385–392.

Starbuck, W. (1983). Organizations as action generators. American Sociological Review,48(1), 91–102.

Sulzer-Azaroff, B., Loafma, B., Merante, R., & Hlavack, A., (1990). Improving occupationalsafety in a large industrial plant: A systematic replication. Research Studies in Per-formance Management, 3, 557–580.

Sulzer-Azaroll, B., & De Santamaria, M. (1980). Industrial safety hazrd reduction throughperformance feedback. Journal of Applied Behavior Analysis, 13, 287–295.

Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and appliedimplication. Journal of Applied Psychology, 65, 96–102.

Zohar, D. (2002). Modifying supervisory practices to improve subunit safety: A leadership-based intervention model. Journal of Applied Psychology, 87(1), 156–163.

Zohar, D., Dov, R., Luria, G., Gil, L. (2004). Climate as a social-cognitive construction ofsupervisory safety practices: Scripts as proxy of behavior patterns. Journal of AppliedPsychology, 89(2), 322–333.


dimensions of interpersonal relationships and safety in the steel industry

Documents