9505021291

19
® Academy of Management Journal 1995. Vol. 38, No. 2, 555-572. A TYPOLOGY OF DEVIANT WORKPLACE BEHAVIORS: A MULTIDIMENSIONAL SCALING STUDY SANDRA L. ROBINSON New York University REBECCA J. BENNETT University of Toledo In this study, we developed a typology of deviant workplace behaviors using multidimensional scaling techniques. Results suggest that deviant workplace behaviors vary along two dimensions: minor versus serious, and interpersonal versus organizational. On the basis of these two di- mensions, employee deviance appears to fall into four distinct catego- ries: production deviance, property deviance, political deviance, and personal aggression. Theoretical and empirical implications are dis- cussed. Had I a hundred tongues, a hundred mouths, a voice of iron and a chest of brass, I could not tell all the forms of crime. Virgil, Aeneid Employee deviance and delinquency produce organizational losses es- timated to range from $6 to $200 billion annually (Murphy, 1993). Of all employees, 33 to 75 percent have engaged in some of the following behav- iors: theft, computer fraud, embezzlement, vandalism, sabotage, and absen- teeism (Harper, 1990). Almost daily, there are media reports of workplace deviance, whether it be corruption among police officers, violence in the post office, or illegal activity on Wall Street. The prevalence of workplace deviance and its associated organizational costs necessitates a specific, systematic, theoretically focused program of study into this behavior. To date, relatively little empirical research has directly addressed the darker side of employee behavior (Vardi & Wiener, 1992). The organization behavior literature has shown a disproportionate emphasis on desirable phenomena such as organizational citizenship behav- ior (e.g.. Organ, 1988), commitment (e.g., Mowday, Porter, & Steers, 1982), and adaptation (e.g., Hulin, 1991). An Academic Challenge Grant from the University of Toledo supported this study. We gratefully acknowledge the assistance of Daniel Lopuch and Timothy Murphy. We also thank Karen Jehn, for her advice and assistance during the various stages of the project, and Blake Ashforth, Aela Boyum, Dafna Eylon, Daniel Farrell, and Elizabeth Morrison, for their comments on earlier drafts of this article. 555

Upload: jules-iulia

Post on 06-May-2017

213 views

Category:

Documents


0 download

TRANSCRIPT

® Academy of Management Journal1995. Vol. 38, No. 2, 555-572.

A TYPOLOGY OF DEVIANT WORKPLACE BEHAVIORS: AMULTIDIMENSIONAL SCALING STUDY

SANDRA L. ROBINSONNew York University

REBECCA J. BENNETTUniversity of Toledo

In this study, we developed a typology of deviant workplace behaviorsusing multidimensional scaling techniques. Results suggest that deviantworkplace behaviors vary along two dimensions: minor versus serious,and interpersonal versus organizational. On the basis of these two di-mensions, employee deviance appears to fall into four distinct catego-ries: production deviance, property deviance, political deviance, andpersonal aggression. Theoretical and empirical implications are dis-cussed.

Had I a hundred tongues, a hundred mouths, a voice of iron and a chestof brass, I could not tell all the forms of crime.

Virgil, Aeneid

Employee deviance and delinquency produce organizational losses es-timated to range from $6 to $200 billion annually (Murphy, 1993). Of allemployees, 33 to 75 percent have engaged in some of the following behav-iors: theft, computer fraud, embezzlement, vandalism, sabotage, and absen-teeism (Harper, 1990). Almost daily, there are media reports of workplacedeviance, whether it be corruption among police officers, violence in thepost office, or illegal activity on Wall Street.

The prevalence of workplace deviance and its associated organizationalcosts necessitates a specific, systematic, theoretically focused program ofstudy into this behavior. To date, relatively little empirical research hasdirectly addressed the darker side of employee behavior (Vardi & Wiener,1992). The organization behavior literature has shown a disproportionateemphasis on desirable phenomena such as organizational citizenship behav-ior (e.g.. Organ, 1988), commitment (e.g., Mowday, Porter, & Steers, 1982),and adaptation (e.g., Hulin, 1991).

An Academic Challenge Grant from the University of Toledo supported this study. Wegratefully acknowledge the assistance of Daniel Lopuch and Timothy Murphy. We also thankKaren Jehn, for her advice and assistance during the various stages of the project, and BlakeAshforth, Aela Boyum, Dafna Eylon, Daniel Farrell, and Elizabeth Morrison, for their commentson earlier drafts of this article.

555

^^^ Academy of Management Journal April

Some research has addressed behaviors that could be considered devi-ant, although they have not been conceptualized as such: absenteeism (e gGoodman & Atkins, 1984), withdrawal (e.g., Gupta & Jenkins, 1980), with-holding effort (e.g., Kidwell & Bennett, 1993), and behaviors that lead toprocedural or distributive injustice or both (e.g., Sheppard, Lewicki, & Min-ton, 1992). However, these research efforts have not focused on the deviantnature of the behaviors themselves. Thus, although such research may ex-amine the same behaviors as the study of employee deviance and be usefulfor understanding it, workplace deviance needs to be examined as a distinctand important organizational phenomenon in its own right.

Of the few studies examining workplace deviance, most have been iso-lated attempts to answer specific questions about particular types of deviantacts. For example, studies have looked exclusively at theft (Greenberg, 19901993; Hollinger & Clark, 1982), sexual harassment (Gutek, 1985), and uneth-ical decision making (Trevino & Youngblood, 1990). Researchers have yet todevelop a comprehensive theory or set of theories regarding workplace de-viance. For empirical work to advance an area of knowledge, studies thatbuild upon one another are needed (Robertson, 1993). The development ofemployee deviance theories will direct the currently scattered research ef-forts and enable researchers to establish complementary research agendas Insum, a systematic, theory-directed study of deviance will ultimately increaseunderstanding of workplace deviance.

WORKPLACE DEVIANCE

Definition

Employee deviance is defined here as voluntary behavior that violatessignificant organizational norms and in so doing threatens the well-being ofan organization, its members, or both. Employee deviance is voluntary inthat employees either lack the motivation to conform to normative expecta-tions of the social context or become motivated to violate those expectations(Kaplan, 1975). Organizational norms, or those prescribed by formal andinformal organizational policies, rules, and procedures, are specified herebecause deviance must be defined in terms of the standards of a specifiedsocial group rather than in reference to a system of absolute moral standards(Kaplan, 1975). We focused on the violation of norms espoused by the dom-inant administrative coalitions of organizations rather than the norms ofwork groups or subcultures.

The study of workplace deviance is distinct from the study of ethics inthat the former focuses on behavior that violates organizational normswhereas the latter focuses on behavior that is right or wrong when judged interms of justice, law, or other societal guidelines determining the morality ofbehavior (Lewis, 1985). Thus, although a particular behavior can be bothdeviant and unethical, the two qualities are not inevitably linked. For ex-ample, dumping toxic waste in a river is not deviant if it conforms with thepolicies of one's organization. However, most people would probably agree

1995 Robinson and Bennett 557

that this act is unethical. Conversely, reporting this dumping to authoritiesmay be an ethical act, but it would also be a deviant act in this particularexample if it violated organizational norms.

Deviant behavior also has the potential to harm an organization, itsmembers, or both. The term deviant is usually reserved for acts that violatesignificant norms (Cohen, 1966) and result in an "unacceptable violation . . .believed to threaten society's well-being" (Best & Luckenbill, 1982: 4). Con-sistent with this focus, our definition of workplace deviance focuses onviolations of norms that threaten the well-being of an organization. Hence,employee deviance excludes minor infractions of social norms, such aswearing a suit of the wrong style to the office, that are not usually or directlyharmful to most organizations.^

A Typology

A typology of employee deviance is a useful starting point for develop-ing a systematic, theory-based study of employee deviance. We saw such atypology as enabling us to develop broader, more comprehensive theories ofdeviance (Rich, 1992) and as giving parsimony and order to the diverse setof behaviors that comprise workplace deviance, helping us to identify therelationships between these different deviant behaviors, and enabling us tomake connections between the different findings of studies that have ad-dressed specific types of deviant behavior. We also viewed a typology asuseful for developing broader measures of employee deviance and thus asenabling empirical tests of our theories of deviance. Aggregated measures aremore reliable and valid than specific measures (Rushton, Brainerd, & Press-ley, 1983) and also overcome the low base rate problems commonly associ-ated with measuring deviant behaviors (Hulin & Rousseau, 1980).

Few attempts have been made to classify employee deviance. Redeker(1989) developed a list of punishable offenses in organizations but did notintegrate the different behaviors into any meaningful pattern. Wheeler(1976), examining how arbitrators should punish rule-breaking behavior,classified forms of organizational rule-breaking into serious offenses andnonserious offenses. HoUinger and Clark (1982) categorized employee devi-ance into property deviance, or acquiring or damaging property belonging toone's employer, and production deviance, or violating organizational normsregarding the quantity and quality of work performed. Mangione and Quinn(1974) suggested two similar categories of deviance: counterproductive be-havior, defined as purposely damaging employers' property, and doing little,defined as producing output of poor quality or low quantity.

These categorical schemes of workplace deviance provide a useful start-ing point for creating an integrative framework of deviant behaviors. How-

^ It should be noted that although many behaviors might meet one of our definitionalcriteria, we focused only on behaviors that met all three criteria and were likely to be deviantin most organizational contexts.

^^^ Academy of Management Journal April

ever, these typologies also raise several questions. First, are the typologiescomprehensive? They capture acts against organizations, such as theft andslowed production, but they do not seem to be able to account for deviantacts of an interpersonal nature, such as physical aggression and sexual ha-rassment. An accurate typology of employee deviance should take into ac-count not only behavior directed at organizations, but also that directed atindividuals. Second, how are the different deviant behaviors, or the catego-ries themselves, related to one another? It would be useful to have a typologythat identified the dimensions underlying these categories. And finally arethe typologies valid? Although it is quite conceivable that these typologiesare all somewhat accurate classifications of employee deviance, they werenot inductively or empirically derived and have not yet been empirically

In this study, we sought to answer the above questions by developing aninductively and empirically derived typology of workplace deviance usingmultidimensional scaling (MDS) techniques (Kruskal & Wish, 1978) Theproduct was intended to be a comprehensive classification of' deviant be-haviors highlighting the similarities and differences between deviant behav-iors as well as their underlying dimensions.

METHODS AND RESULTS

Multidimensional scaling is a useful tool for producing inductive butempirically derived, typologies. MDS techniques enable a researcher to pro-duce a typology using the perceptions of a diverse set of individuals who areblind to the purpose of a given study. Hence, MDS-based typologies are lessprone to researchers' biases than typologies developed through other meth-ods. MDS involves several distinct phases of data collection and analysesThe procedures followed for each phase in this study and the results of eachphase are discussed below.

Phase 1

Sample. We recruited 70 respondents, 27 men and 43 women, from foursources m Toledo, Ohio: a university office (n = 7), a technical staff officewithm an industrial company (n = 10), a neighborhood (n = 38) and anevening master's in business administration (M.B.A.) class (n = 15) Allrespondents worked full-time. Their average age was 37 years (s.d. = 14 69)and their average number of years of work experience was 15.69 (s.d. ='l ^ t J.UJ.

Procedures. Respondents were asked to describe two incidents ofsomeone at work engaging in something considered to be deviant at their

workplace, i.e., something that is considered to be wrong." We provided thiscolloquial definition of employee deviance instead of our theoretical defi-nition because it was easier to understand. We also asked respondents todefine deviance in their own words. Over 81 percent made specific reference

1995 Robinson and Bennett 559

to violation of norms or rules, consistent with our definition. Both the de-scriptions and the definitions were in writing.

Next, the second author and a research assistant independently re-phrased the descriptions the respondents provided to simplify them, to re-move redundant words and phrases (most were a paragraph or longer), andto ensure that the descriptions were relatively generic and applicable acrossorganizations and occupations. Descriptions repeating another descriptionwere then removed. The final pool of statements described 45 deviant work-place behaviors; the Appendix lists these behaviors.

We also had 12 judges, professors of management, independently assesshow well each behavior fit our definition of employee deviance. They ratedeach behavior (yes or no) in terms of whether it was voluntary, potentiallyharmful to organizations, and likely to violate the norms of most organiza-tions. The judges unanimously agreed that most of the behaviors fit ourdefinition of deviance.

Phase 2Sample. There were 180 respondents, 86 men and 94 women, all of

whom were part-time evening students in an M.B.A. program at a midwest-ern university. All the respondents worked full-time. Their average age was29 years (s.d. = 12.57) and their average number of years of work experiencewas 11.12 (s.d. = 11.06).

Procedures. We gave each respondent a survey containing the list ot 45deviant workplace behaviors and a brief description of a target behavior,which appeared at the top of the first page. The respondents rated eachdeviant behavior in terms of its similarity to or difference from the targetbehavior, using a nine-point Likert-type scale (1 = very similar, 9 = verydifferent). We also asked respondents to specify the criteria they used todistinguish between the target behavior and each of the deviant behaviors.

Although MDS often involves having respondents compare every pos-sible pair of stimuli [n(n - l)/2], we asked our respondents to make only asubset of the 990 possible comparisons because having them address the fullset would have been too cognitively taxing and likely to have resulted infatigue, errors, and respondent attrition. A valid means by which to over-come the problems associated with comparing a large number of stimuli is tohave respondents make a subset of comparisons (Thompson, 1983).

To determine the visual configuration and underlying dimensions of thedeviant behaviors, we used the ALSCAL program (Young & Lewykcyj, 1979).This program derives spatial configurations of objects on the basis of theperceived differences between the objects. The greater the perceived differ-ence between the objects, the greater the distance between them in the spa-tial configuration. We first created a dissimilarities matrix by computing theperceived differences between the pairs of deviant behavior descriptions(Kruskal & Wish, 1978). We then employed Torgeson's (1952) metric MDSanalysis to create five different visual configurations of these deviant behav-ior descriptions, ranging from one dimension to five dimensions. An analogy

^^'^ Academy of Management Journal April

for this program is that it is like having a computer draw several maps ofvarious dimensions for a set of cities by relying on only information aboutthe distances between the cities. The resulting maps could be one-dimensional in that they would be placed along a single line- two-dimensional like a typical road map; three-dimensional like a globe withbas-relief reflecting elevation; and so forth. We used Kruskal's (1964) stressindex to determine which map configuration explained the most varianceThis stress index indicates how well data fit a particular configuration- thehigher the stress, the poorer the fit.

Results. We conducted a scree test by plotting the stress indexes for allfive map configurations (Cattell, 1986). The plot of stress indexes producesa curve. The appropriate configuration is determined on the basis of where(at which configuration) the stress index values begin to level off to form analmost horizontal slope. The one-dimensional solution had a stress index of.494. For the two-dimensional solution, the index made a considerable dropto .27 suggesting a better fit with the data. The amount of reduced stressleveled off for the three-, four-, and five-dimensional solutions with valuesof .199, .159, and .137, respectively. Hence, the scree results suggested thatthe two-dimensional solution provided the most parsimonious and accuratedescription of the data. Figure 1 shows the two-dimensional configuration.

Phase 3

Procedures. We derived a list of potential labels or attributes to describethe dimensions from the criteria that respondents said they used in phase 2to compare the deviant behaviors and the target behavior. We selected theSIX most frequently cited criteria before looking at the two-dimensional so-lution. Four judges, senior candidates for doctorates in management, ratedhow well each attribute described each deviant behavior. Working indepen-dently and blind to the purpose of the study, they used five-point bipolarscales with the following attribute anchors: unintentional/intentional notserious/serious, not harmful to company/harmful to company, not harmful toindividuals/harmful to individuals, very unethical/ethical, and covert/overt.

Results. The judges' ratings were averaged for each attribute-behaviorcomparison. We used muhiple regression analysis to assess relationshipsbetween the attributes and the two-dimensional configuration The labelschosen were based on examination of the multiple correlation coefficientsf-values, and beta weights from the regression analysis (Kruskal & Wish1978). Table 1 provides those statistics.

Dimension 1. Relationships between the attributes and the first dimen-sion suggested a label reflecting the seriousness or harmfulness of the devi-ant acts. The attributes that explained the most variance for dimension 1were not serious/serious (ß = .66, p < .001) and very unethical/ ethical (ß =- .67, p < .001). Also significantly related to dimension 1 were not harmfulto conipany/harmful to company (ß = .42, p < .001) and not harmful toindividuals/harmful to individuals (ß = .27, p < .05). Thus, one end of this

1995 Robinson and Bennett 561

dimension reflected deviant behaviors that were not serious, not harmful tothe company, and not harmful to the individuals targeted, and the other endreflected deviant behaviors that were serious, harmful to the company, andharmful to the individuals targeted. Consequently, we labeled this first di-mension "minor versus serious deviance."

Other attributes were also related to this dimension in ways consistentwith the minor-versus-serious label. Covert/overt was negatively related tothe not serious/serious attribute (r = - .36, p < .01) and positively related tothe very unethical/ethical attribute (r = .47, p < .01). This relationship is notsurprising given that harmful behavior is typically more covert than innoc-uous behavior. Similarly, the unintentional/intentional attribute was posi-tively related to the not serious/serious attribute (r = .57, p < .01) andnegatively related to the unethical/ethical attribute (r = - .60, p < .01). Thisrelationship is consistent with the fact that behavior is perceived to be in-tentional to the extent that it is considered harmful (Hamilton, 1980). Ex-amination of the two-dimensional configuration (Figure 1) also supports useof the label minor versus serious deviance. The less serious, less harmfuldeviant behaviors fell on the negative end of this dimension, and the moreharmful or serious deviant behaviors fell on the positive end of this dimen-sion. J

Dimension 2. The relationships between the attributes and the seconddimension suggested a label reflecting the extent to which deviant behaviorsare interpersonal and harmful to individuals rather than non-interpersonaland harmful to organizations. The attribute not harmful to company/harmfulto company was positively related to dimension 2 (ß = .30, p < .05), and theattribute not harmful to individuals/harmful to individuals was negativelyrelated to dimension 2 (ß = - .84, p < .001). The covert/overt attribute wasalso negatively related to this dimension (ß = - .74, p < .001). This rela-tionship makes sense given that interpersonal behaviors are more likely to beovert than covert since at least one other person is usually present in thecontext of interpersonal behavior. Hence, one end of this dimension re-flected behaviors that were harmful to individuals, not harmful to the organ-ization, and overt, and the other end reflected behaviors that were harmfulto the organization, not harmful to individuals, and covert. Consequently,we chose the label "interpersonal versus organizational deviance" for di-mension 2. Observation of the two-dimensional configuration (Figure 1) sup-ports use of this label. Deviant behaviors that fell on the negative end of thisdimension were overt, interpersonal behaviors, directly harmful to individ-uals and deviant behaviors that fell on the positive end of this dimensionwere non-interpersonal behaviors that were directly harmful to organiza-tions. . 1 f

Quadrant labels. A closer inspection of the two-dimensional configu-ration suggests that deviant acts not only vary along two dimensions but canalso be classified into four categories. Figure 2 exhibits these categories.

The quadrant containing serious and organizationally harmful deviancewas labeled "property deviance." This quadrant is consistent with Man-

562 Academy of Management Journal April

FIGURE 1Two-Dimensional Configuration of Deviant Behaviors

Employee making personalcalls or mailings

Boss leaving early or Employee wasting com-leaving his work for pany resources

employee to do Employee com- *• ing in late or

leaving earlyEmployee taking excessive breaks • Employee leaving job in

p , , . • Employee calling in sick when not ""^"^ïbmployee working unnecessary overtime .

Employee hiding in back Employee intentionally working slow0.5 -^ room to read newspaper *

Employee endan-gering self

• Employee talking with co-worker instead of working

Employee acting foolish in front of customer

* Employee starting negativirumors about company

0 -

- 0 .5 -

-„

- 1 . 5 -

Employee gossiping Employee blaming co-worker for mistakesabout manager ,

Boss asking employee to work beyond job description

Employee gossiping about co-worker

Boss blaming employee for own mistakes

Employee competing in nonbeneficial way

Boss showing favoritism• Boss gossiping about employees

- 2 -

1995 Robinson and Bennett 563

FIGURE 1 (continued)

Employee misusing Employee stealingdiscount privilege Employee accepting company equipment and merchandise

Employee lying kickbacks .u u 1 J • Employee steahng money from cash

ahout hours worked * '̂ -' ,drawer

Employee going Employee intentionallyagainst hoss's decision making errors Employee sabotaging

, merchandise•

Employee overcharging for Employee misusingservices for own profit expense account

• •Employee covering up mistakes

, Employee stealing customer's possessions

Employee sabotaging equipment

Boss refusing to give employee earnedbenefits or pay

Employee physically abusing customer

Employee stealing co-worker's possessions

Employee verbally abusing customerEmployee endangering co-workers •

by reckless behavior

Boss unjustifiably firing employee

Boss sexually harassing employee

, ,, , . * Employee sexually harassing co-workerBoss verbally abusing ^ ' •' °

employee

Boss following rules to letter of law

0.5 1 1-5

564 Academy of Management Journal April

I»3

a

a•3

C35 OCM m

CM1

«

CO

1

*«05CO

1

«COCO

I I

a5g11S3

I I

CO CM o ^ CO ^i-i t-H CO 00 O ^s

oI

I I

S:

COCO

2408

*

CO

38*

CO

14*

38

CO

CO

ldua

ompa

nin

div

*

85*

25

inlO

*

tr.

10

COCO

1995 Robinson and Bennett 565

gione and Quinn's (1974) counterproductive behavior and Hollinger andClark's property deviance, which they defined as "those instances whereemployees acquire or damage the tangible property or assets of the workorganization without authorization" (1982; 333). The quadrant reflectingrelatively minor but still organizationally harmful deviant acts was labeled"production deviance." This quadrant is similar to Mangione and Quinn's(1974) doing little or nothing and Hollinger and Clark's production devi-

FIGURE 2Typology of Deviant Workplace Behavior"

ORCANIZATIONAL

i

Production Deviance

• Leaving early• Taking excessive breaks• Intentionally working slow• Wasting resources

Political Deviance

• Showing favoritism• Cossiping about co-workers• Blaming co-workers• Competing nonbeneficially

\

Property Deviance

• Sabotaging equipment• Accepting kickbacks• Lying about hours worked• Stealing from company

Personal Aggression

• Sexual harassment• Verbal abuse• Stealing from co-workers• Endangering co-workers

SERIOUS

INTERPERSONAL

'These lists are not exhaustive. We provide a set of the most typical behaviors for eachcategory for illustrative purposes only.

566 Academy of Management Journal April

ance, which they defined as "behaviors that violate the formally proscribednorms delineating the minimal quality and quantity of work to be accom-plished" (1982: 333). The third quadrant contained minor and interperson-ally harmful deviant behavior. We labeled this quadrant "political devi-ance," defining the behavior as engagement in social interaction that putsother individuals at a personal or political disadvantage. The final quadrant,containing serious and interpersonally harmful deviant behavior, was la-beled "personal aggression," which we defined as behaving in an aggressiveor hostile manner toward other individuals.

The four quadrants appeared to represent four distinct but related typesof deviance. To garner empirical support for our post hoc interpretations ofthese categories, we had four judges who were blind to the study and itsresults independently code the behaviors. These judges, who were not thesame as the judges used in phase 3, were doctoral students in management.They coded each behavior into one of four categories on the basis of thelabels and definitions provided. Kappas were calculated for each of the fourcategories to measure the degree of interrater reliability. Overall, the kappas,which ranged from 74 to 89 percent, indicated high agreement between ourtypology and the raters' categorization of the behaviors. These results lendvalidity to our interpretation of the quadrant labels and their meanings.

DISCUSSION

The study reported here integrated numerous deviant workplace behav-iors into a parsimonious framework. The results suggest that workplace de-viance varies along two dimensions and can be classified into four types.The typology derived here makes a contribution to the literature by empir-ically validating Wheeler's (1976) distinction between serious and nonseri-ous workplace offenses as well as Mangione and Quinn's (1974) andHollinger and Clark's (1982) typologies, which distinguish between produc-tion and property deviance.

A more significant Contribution of this study is that it builds upon theseprevious categorical schemes and produces a more accurate and compre-hensive typology of workplace deviance. First, our typology identifies theunderlying dimensions of deviance and thus clarifies not only the differentcategories of deviance but also how these categories are related to one an-other. Second, the typology incorporates two previously neglected forms ofemployee deviance, political deviance and personal aggression. To date, thelimited literature on workplace deviance has focused primarily on produc-tion and property deviance (acts directed at organizations), ignoring inter-personal forms of deviance. As this study demonstrates, numerous behaviorsdirected at individuals, including verbal abuse, physical assault, and polit-ical behavior, are also perceived as deviant. These findings suggest thatworkplace deviance research should address social as well as organization-directed forms of deviance. This refocusing is important in light of growingconcerns about reducing social injustice, discrimination, and interpersonalviolence in workplaces.

1995 Robinson and Bennett 567

This typology should prove useful in the development of general theo-ries of workplace deviance. First, it creates meaningful patterns out of thewide range of deviant behaviors and enables researchers to look at the gamutof deviant behaviors as a parsimonious whole. Second, it identifies the re-lationships between these behaviors as well as their underlying constructs.This framework paves the way for creating an integrated theory or theoriesencompassing the behaviors found within each quadrant or across quad-rants. Rather than continue to expend efforts on separate, unconnected, andpotentially redundant studies of specific deviant acts, researchers can beginto develop and test theoretical models of the basic forms of deviant behavior,such as production deviance and personal aggression. Such theories willdirect research attention and enable systematic exploration of this phenom-enon and cumulation of findings.

The results of this study have implications for theories addressing theantecedents of workplace deviance. Opinions on the causes of devianceabound; some have argued that deviance results from individual attributes,such as low moral standards (Merriam, 1977), and others have argued forsituational explanations of deviance, such as organizational inequities(Greenberg, 1990) and group norms (Siehl, 1987). This study suggests thatdifferent variables may explain different types of workplace deviance. Forexample, organizational variables might be more likely to influence devi-ance directed at harming organizations, and individual variables may bemore likely to explain interpersonal forms of deviance.

The results of this study may also have implications for theories ad-dressing the outcomes of workplace deviance. Deviance may be dysfunc-tional and threatening to the well-being of a social system (Best & Luckenbill,1982), but it may have several positive outcomes as well: providing a safetyvalve, alerting group members to their common interests, and providingwarning signals to organizations. This study, which clarifies the differenttypes of workplace deviance, suggests that different forms of deviance mayhave different consequences. For example, interpersonal deviance may servesocial functions for organization members—building group cohesiveness,for example—and organization-directed deviance may be more likely toserve signaling functions for organizations.

This typology, which incorporates a wide range of behaviors from var-ious domains of organizational behavior, is also valuable in that it connectsthese previously unrelated domains of study. For example, the employeedeviance literature, which has focused almost exclusively on organization-directed forms of deviance such as sabotage and theft, has remained largelyseparate from the growing literature on discrimination and sexual harass-ment in the workplace (Kahn & Robbins, 1985; Levinger, 1987), which ad-dresses some behaviors that could be considered interpersonal forms of de-viance. Our typology may provide a bridge between these two currentlyunrelated bodies of research by cohceptually integrating organizational andinterpersonal deviance. Similarly, this typology may offer a conceptualbridge between the traditional study of absenteeism and withdrawal (re-

568 Academy of Management Journal April

fleeted in production deviance) and other, thus far unrelated behaviors thatwe also identified as forms of production deviance, such as wasting com-pany resources and intentionally working slowly.

These findings also have managerial implications. Procedural justiceresearch has shown that employees perceive consistent punishment as fair(Bennett, 1993); that is, punishment is perceived to be fair to the extent thatsimilar behaviors are punished similarly and to the extent that it matches theseriousness of the offense committed (Wheeler, 1976). This study revealsemployees' perceptions of the similarity between potentially punishable be-haviors and their perceptions of the seriousness of these behaviors. Theprocedural justice findings, combined with the results of this study, suggestthat managers who are seeking to be fair should apply similar types of pun-ishments to deviant acts that are physically close to one another in thespatial configuration, such as intentionally making errors and engaging insabotage. Managers should also match the severity of punishment to theperceived seriousness of a deviant act; for instance, employees acceptingkickbacks and employees intentionally making errors should receive simi-lar, severe punishments, whereas employees intentionally working slowlyshould receive less severe punishments.

Although this study focused on deviant workplace behavior rather thanon unethical behavior, it does suggest that the study of ethics may benefitfrom an analogous multidimensional scaling study of unethical behavior.Robertson (1993), discussing ethics research, offered a number of recommen-dations. Specifically, she recommended that researchers clarify their oper-ational definition of ethics, emphasize behavior as the key dependent vari-able, focus on theory development, and build broader predictive models ofbehavior. An MDS study like the one employed here would begin to addressthese exact issues.

To summarize, the present study makes a variety of theoretical andpractical contributions. Its findings empirically validate previous categoricalschemes of deviance and extend those typologies by identifying the thus farneglected interpersonal forms of deviance. Further, this typology shouldprove useful for developing much-needed general theories of workplace de-viance, particularly theories concerning the antecedents and outcomes offorms of workplace deviance. It should also aid in creating aggregated mea-sures of workplace deviance to be used to test those theories. More broadly,the typology generated here may provide a valuable conceptual bridge topreviously unrelated domains of study and serve as a model for relateddomains of study, such as organizational ethics. Finally, this typology haspractical implications, suggesting guidelines for ways in which managerscan fairly allocate punishments for deviant behavior.

Several limitations of this study should be noted. First, it was not tech-nically or conceptually feasible to use an exhaustive list of deviant behav-iors. The purpose of our study was to develop a typology from which wecould classify most deviant behaviors (not only those used to create thetypology). That goal was accomplished. Second, although we attempted to

1995 Robinson and Bennett 569

reduce researcher biases, they were not eliminated. Our biases may, forexample, have entered the study when we selected and rephrased the be-havioral descriptions to be used in the analysis. However, we only shorteneddescriptions, changing no content, and we removed only descriptions ofalready included behaviors. Second, we may have also been biased in ourselection of the attributes used in interpreting the dimensions. However, weselected only the most frequently used attributes, prior to looking at theconfiguration. Further, since we based the interpretation of the dimensionson a large set of attributes, the inclusion or exclusion of an additional at-tribute is unlikely to have altered this interpretation. And finally, in derivingthe names and meanings of the quadrants, we relied on our post hoc inter-pretation of the clusters. We attempted to offset the potential bias in thisinterpretation by validating our results with the judgments of independentand blind judges. In sum, we made a conscious effort to reduce the influenceof our biases in the development of this inductively derived typology.

We hope this study will bring attention to the darker side of organiza-tional behavior. It is our intention that this study serve as a springboard foradditional empirical research into workplace deviance. Some future re-search directions include the development of aggregated measures of devi-ance based on the "four p's" (property deviance, production deviance, po-litical deviance, and personal aggression) and the development of predictivemodels of deviance that take into account the wide range of deviant behav-iors. Management research has developed a fairly comprehensive under-standing of extrarole, prosocial, organizational citizenship behavior (e.g..Organ, 1988) but has largely neglected subrole, antisocial, deviant behavior.Researchers and practitioners need to understand not only behavior that isbeneficial to organizations, but also behavior that is detrimental to them.

REFERENCES

Bennett, R. 1993. Reactions to fair punishment for unethical behavior. Working paper. Uni-versity of Toledo, Toledo, OH.

Best, J., & Luckenbill, D. F. 1982. Organizing deviance. Englewood Cliffs, NJ: Prentice-Hall.

Cattell, R. B. 1986. The meaning and strategic use of factor analysis. In R. B. Cattell & J. R.Nesselroade (Ed.), Handbook of multivariate experimental psychology (2d ed.): 131-203.New York: Plenum.

Cohen, A. K. 1966. Deviance and social control. Englewood Cliffs, NJ: Prentice-Hall.

Goodman, P., & Atkins, R. 1984. Absenteeism: New approaches to understanding, measuringand managing employee absence. San Francisco: Jossey-Bass.

Creenberg, J. 1990. Employee theft as a reaction to underpayment inequity: The hidden cost ofpay cuts. Journal of Applied Psychology, 75: 561-568.

Greenberg, J. 1993. Stealing in the name of justice: Informational and interpersonal moderatorsof theft reactions to underpayment inequity. Organizational Behavior and Human Deci-sion Processes, 54: 81-103.

Gupta, N., & Jenkins, G. 1980. The structure of withdrawal: Relationships among estrange-ment, tardiness, absenteeism and turnover. Springfield, VA: National Technical Informa-tion Service.

^'^^ Academy of Management Journal April

Cutek, B. A. 1985. Sex in the workplace. San Francisco: Jossey-Bass.

Harper, D. 1990. Spotlight abuse—Save profits. Industrial Distribution, 79: 47-51.

Hamilton, V. L. 1980. Intuitive psychologist or intuitive lawyer? Alternative models of the

attribution process. Journal of Personality and Social Psychology, 39: 767-772.

Hollinger, R. C, & Clark, J. P. 1982. Formal and informal social controls of employee devianceSociological Quarterly, 23: 333-343.

Hulin, C. 1991. Adaptation, persistence and commitment in organizations. In M. D. Dunnette &I. M. Hough (Eds.), Handbook of industrial organizational psychology, vol. 2 (2d ed )•445-505. Palo Alto, CA: Consulting Psychologists Press.

Hulin, C, & Rousseau, D. 1980. Analyzing infrequent events: Once you find them, your troublesbegin. In K. H. Roberts & L. Burstein (Eds.), Issues in aggregation: 65-75. San Francisco-Jossey-Bass.

Kahn, E. D., & Robbins, L. (Eds.). 1985. Sex discrimination in academe. Journal of SocialIssues, 41(4): 135-154.

Kaplan, H. B. 1975. Self-attitudes and deviant behavior. Pacific Palisades, CA: Coodyear.Kidwell, R. E., & Bennett, N. 1993. Employee propensity to withhold effort: A conceptual model

to intersect three avenues of research. Academy of Management Review, 18: 429-456.

Kruskal, K. B. 1964. Multidimensional scaling by optimizing goodness of fit to a non-metrichypothesis. Psychometrika, 29: 1-27.

Kruskal, K. B., & Wish, M. 1978. Multi-dimensional scaling. Sage University paper series onQuantitative Applications in the Social Sciences, 07-011. Beverly Hills and London: Sage.

Lewis, P. V. 1985. Defining "business ethics": Like nailing jello to a wall. Journal of BusinessEthics, 4: 377-383.

Levinger, G. (Ed.). 1987. Black employment opportunities: Macro and micro perspectives Tour-nai of Social Issues, 43(1): 1-4.

Mangione, T. W., & Quinn, R. P. 1974. Job satisfaction, counterproductive behavior, and druguse at work. Journal of Applied Psychology, 1: 114-116.

Merriam, D. 1977. Employee theft. Criminal Justice Abstracts, 9: 380-386.

Mowday, R. T., Porter, L. M., & Steers, R. M. 1982. Employee-organizational linkages: The

psychology of commitment, absenteeism, and turnover. New York: Academic Press.

Murphy, K. R. 1993. Honesty in the workplace. Belmont, CA: Brooks/Cole.

Organ, D. W. 1988. Organizational citizenship behavior. Lexington, MA: Lexington.

Redeker, J. R. 1989. Employee discipline. Washington, DC: Bureau of National Affairs.Rich, P. 1992. The organizational taxonomy: Definition and design. Academy of Management

Beview, 17: 758-781.

Robertson, D. C. 1993. Empiricism in business ethics: Suggested research directions. Journal ofBusiness Ethics, 12: 585-599.

Rushton, J. P Brainerd, C. J., & Pressley, M. 1983. Behavioral development and construct va-lidity: The principle of aggregation. Psychological Bulletin, 94: 18-38.

Sheppard, B. H., Lewicki, R. J., & Minton, J. W. 1992. Organizational Justice: The search forfaimess m the workplace. New York: Lexington Books.

Siehl, E. W. 1987. Garment workers: Perceptions of inequity and employee theft. British Jour-nal of Criminology, 27{2]: 174-190.

Thompson, P. 1983. Some missing data patterns for multidimensional scaling. Applied Psy-chological Measurement, 7: 45-55.

Robinson and Bennett 571

Torgeson, W. S. 1952. Multidimensional scaling: Theory and method. Psychometrica, 17: 401-419.

Trevino, L. K., & Youngblood, S. A. 1990. Bad apples in bad barrels: A causal analysis of ethical

decision making behavior. Journal of Applied Psychology, 75: 378-385.

Vardi, Y., & Wiener, Y., 1992. Organizational misbehavior [OMB): A calculative-normativemodel. Paper presented at the annual meeting of the Academy of Management, Las Vegas.

Wheeler, H. N. 1976. Punishment theory and industrial discipline. Industrial Relations, 15:235-243.

Young, F. W., & Lewyckyj, R. 1979. ALSCAL-4: User's guide. Chapel Hill, NC: Data Analysisand Theory Associates.

APPENDIXDeviant Behavior Descriptions Used in the

Multidimensional Scaling Analysis

Employee stealing customer's possessions.Boss verbally abusing employee.Employee sabotaging equipment.Employee coming to work late or leaving early.Employee lying about hours worked.Employee gossiping about manager.Employee starting negative rumors about company.Boss sexually harassing employee.Employee physically abusing customer.Employee taking excessive breaks.Employee sabotaging merchandise.Employee overcharging on services to profit him- or herself.Employee intentionally making errors.Employee covering up mistakes.Employee leaving job in progress with no directions so the job is done wrong.Boss following rules to the letter of the law.Employee gossiping about co-worker.Employee intentionally working slowly.Boss unjustifiably firing employee.Employee sexually harassing co-worker.Employee accepting kickbacks.Employee endangering him- or herself by not following safety procedures.Boss leaving early and leaving his/her work for employees to do.Employee hiding in hack room to read the newspaper.Employee stealing company equipment/merchandise.Employee acting foolish in front of customers.Employee verbally abusing customers.Employee working unnecessary overtime.Employee calling in sick when not.Boss showing favoritism to certain employees.Boss gossiping about employees.Employee talking with co-worker instead of working.Employee stealing money from cash drawer.Employee misusing discount privilege.Employee wasting company resources by turning up the heat and opening the windows.Employee blaming co-worker for mistakes.Employee misusing expense account.Employee going against boss's decision.

^^2 Academy of Management Journal April

Employees competing with co-workers in a nonbeneficial way.Boss blaming employees for his/her mistakes.Boss refusing to give employee his/her earned benefits or pay.Employee making personal long distance calls or mailing personal packages from work.Employee endangering co-workers by reckless behavior.Employee stealing co-worker's possessions.Boss asking employee to work beyond job description.

Sandra L. Robinson is an assistant professor of management at the Stern School ofBusiness, New York University. She received a Ph.D. degree in organizational behaviorfrom Northwestern University. Her current research interests include workplace devi-ance, psychological contracts, employee voice, and employees' behavioral responses todissatisfaction.

Rebecca J. Bennett is an assistant professor of management at the University of Toledo.She received her Ph.D. degree in organizational behavior from Northwestern Univer-sity. Her current research interests include workplace deviance, employees' behavioralreactions to unfair punishment, and the relationship between organizational citizen-ship behavior and workplace deviance.

Copyright of Academy of Management Journal is the property of Academy of Management and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.