Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

INTERNATIONAL JOURNAL OF SELECTION AND ASSESSMENT VOLUME 10 NUMBERS 1/2 MARCH/JUNE 2002

The Structure of Counterproductive Work


Behaviors: Dimensionality and Relationships
with Facets of Job Performance1
Paul R. Sackett*
University of Minnesota

The central themes of this article are the need for an understanding of the covariance
structure of counterproductive behaviors, and the need for an understanding of the
relationship between counterproductive behaviors and other facets of job performance.
Various literatures are reviewed to address issues of the intercorrelations among
dimensions of counterproductive behaviors and the intercorrelations between counter-
productive behavior and other facets of job performance.

Introduction sick has become widespread), and hence the behavior is


not deviant in the norm-violation sense. The behavior

C ounterproductive workplace behavior at the most


general level refers to any intentional behavior on
the part of an organization member viewed by the
may be still, however, viewed by the organization as
counterproductive. Note, though, that there are limits to
taking the organization's view: some behaviors (e.g.
organization as contrary to its legitimate interests. leaving one's job for career improvement) are counter-
`Counterproductive behavior' is distinguished from productive in the sense of being contrary to the
`counterproductivity', with the latter viewed as the organization's interests, yet do not carry the connotation
tangible outcomes of counterproductive behavior. of wrongdoing that accompanies behaviors viewed as
Counterproductive behavior is viewed here as a facet illegal, immoral, or deviant. Similarly, it may be in an
of job performance. Consistent with current concep- organization's interests to have employees willing to
tualizations of job performance (e.g. Campbell, routinely work 14-hour days without extra compen-
McCloy, Oppler and Sager 1993), performance is sation; again, an unwillingness to do so does not carry
viewed as reflecting behaviors, rather than outcomes. connotations of wrongdoing. Hence the term `legitimate'
Thus the intentional violation of safety procedures is an is included in the definition of counterproductive
example of counterproductive behavior, as such behavior as behavior contrary to the organization's
behaviors put the individual and the organization at legitimate interests.
risk. The number and cost of injuries resulting from Counterproductive behavior encompasses a broad
such counterproductive behaviors might serve as a number of domains. Gruys (1999) identified 87 separate
measure of counterproductivity. In a given time period, counterproductive behaviors appearing in the literature,
violation of safety procedures (behaviors) may not and used a rational sort and factor analytic techniques to
result in any injuries (outcomes), thus illustrating the produce 11 categories of counterproductive behaviors.
distinction between counterproductive behavior and These categories are presented here as an overview of the
counterproductivity. range of behaviors included in the counterproductive
The above definition of counterproductive behavior behavior domain. This list is presented to give a sense of
does clearly take the perspective of the organization. This the range of behaviors in this domain, rather than as an
is done in order to make clear that a behavior can be exhaustive list:
performed by many employees in a given organization
(e.g. a setting where taking sick leave when not actually 1. Theft and related behavior (theft of cash or property;
giving away of goods or services; misuse of employee
* Address for correspondence: Paul R. Sackett, Department of discount).
Psychology, University of Minnesota, Elliott Hall, 75 East River Road, 2. Destruction of property (deface, damage, or destroy
Minneapolis, MN 55455, USA. e-mail: psackett@tc.umn.edu property; sabotage production);

ß Blackwell Publishers Ltd 2002, 108 Cowley Road, Oxford OX4 1JF, UK and
350 Main Street, Malden, MA 02148, USA. 5
6 PAUL R. SACKETT

3. Misuse of information (reveal confidential infor- The Dimensionality of Counterproductive


mation; falsify records). Behaviors
4. Misuse of time and resources (waste time, alter time
card, conduct personal business during work time). A seminal body of work examining a broad range of
5. Unsafe behavior (failure to follow safety procedures; counterproductive behaviors is that of Hollinger and
failure to learn safety procedures). Clark (1983) who developed a broad list of counter-
6. Poor attendance (unexcused absence or tardiness; productive behaviors, provided a conceptual framework
misuse sick leave). for interrelating those behaviors, and collected self-report
7. Poor quality work (intentionally slow or sloppy data from large employee samples in three industries.
work). They proposed that counterproductive behaviors could
8. Alcohol use (alcohol use on the job; coming to work be grouped into two broad categories. The first is
under the influence of alcohol). `property deviance', involving misuse of employer assets.
9. Drug use (possess, use, or sell drugs at work). Examples include theft, property damage, and misuse of
10. Inappropriate verbal actions (argue with customers; discount privileges. The second is `production deviance',
verbally harass co-workers). involving violating norms about how work is to be
11. Inappropriate physical actions (physically attack co- accomplished. This includes not being on the job as
workers; physical sexual advances toward co- scheduled (absence, tardiness, long breaks) and behaviors
worker). that detract from production when on the job (drug and
alcohol use, intentional slow or sloppy work).
The central themes of this article are (1) the need for an Robinson and Bennett (1995) note that the set of
understanding of the covariance structure of counter- behaviors examined by Hollinger and Clark do not
productive behaviors; and (2) the need for an include interpersonal counterproductive behaviors, such
understanding of the relationship between counter- as sexual harassment, and set out to expand upon the
productive behaviors and other facets of job Hollinger and Clark framework. They had workers
performance. Regarding the first theme, there has been generate a large number of critical incidents of counter-
a tendency to treat each form of counterproductive productive behavior, obtained ratings of the similarity
behavior as discrete, resulting in separate literatures on between pairs of behaviors, and subjected the resulting
behavior categories such as theft, drug and alcohol use, pairwise similarity matrix to multidimensional scaling.
absenteeism, and unsafe behaviors. As a complement to They obtained a two-dimensional solution, with one
these behavior-specific literatures, it is suggested here dimension differentiating behavior toward the
that there is great value in understanding the pattern of organization (Hollinger and Clark's production and
interrelationships among various forms of counter- property deviance) from interpersonal behavior toward
productive behavior. One possibility is that these truly other organizational members (e.g. harassment, gossip,
are independent behaviors, with separate sets of verbal abuse), and the other dimension representing a
antecedents. Another is that some, or perhaps all, of continuum from minor to serious offenses. Arraying
these behaviors are substantially interrelated, with each behaviors in this two-dimensional space, Robinson and
a behavioral manifestation of a latent trait with Bennett labeled the resulting four quadrants as property
common individual difference and/or situational ante- deviance (organizational ± serious), production deviance
cedents. Understanding the pattern of interrelationships (organizational ± minor), personal aggression
would shed light on the possibility of integrating these (interpersonal ± serious, including behaviors such as
distinct literatures. There does appear to have been a harassment, and theft from co-workers), and political
recent movement toward more integrative treatments deviance (interpersonal ± minor, including behaviors
of the range of counterproductive behaviors, as such as favoritism, gossip, and blaming others for one's
reflected in Hollinger and Clark (1983), Griffin, mistakes).
O'Leary-Kelly and Collins (1998), and Robinson and It is critical to note that this categorization scheme is
Greenberg (1998). based on workers' perceptions of the similarity of pairs
With regard to the second theme, counterproductive of behaviors. When asked to judge the similarity of pairs
behavior is viewed as a facet of job performance, where of behaviors, respondents were not constrained as to the
performance is defined as all employee workplace basis for their similarity judgments. The purpose of the
behavior relevant to organizational outcomes. The multidimensional scaling technique is to identify
relationship between counterproductive behavior and dimensions underlying similarity judgements, and the
various forms of productive behavior, such as task study findings indicate that, in the aggregate, respondents
performance and organizational citizenship will be used the two dimensions of organization as target vs.
considered. other person as target and minor offense vs. serious
offense as the basis for their similarity judgments. While
it is useful to know what dimensions underlie perceptions

International Journal of Selection and Assessment ß Blackwell Publishers Ltd 2002


THE STRUCTURE OF CWB 7

of similarity between counterproductive behaviors, it is organizational deviance scale and a seven-item


posited here that the key issue for understanding interpersonal deviance scale; internal consistency
interrelationships among various forms of counter- reliabilities for the two were .81 and .78 respectively.
productive behavior is the covariance of occurrence Correlations between individual behaviors are not
among these behaviors. The question is whether reported, but several useful insights can be gained about
individuals who engage in one form of counterproductive the interrelationships among behaviors. First, the
behavior are also likely to engage in others. It is possible reported internal consistency reliabilities and the
that the pattern of co-occurrence among counter- Spearman-Brown formula can be used to ascertain that
productive behaviors is quite different from the pattern the mean correlation is .26 among the organizational
of perceived similarity emerging from the Robinson and deviance items and .34 among the interpersonal deviance
Bennett perceptual similarity task. For example, theft items. Thus intercorrelations among self-reports of
from a co-worker and verbal abuse of a customer are individual behaviors are positive, but relatively modest
located close together in the multidimensional space in magnitude ± a finding not uncommon for correlations
emerging from Robinson and Bennett's analysis, in that among individual items. Second, Bennett and Robinson
both are serious offenses that target another person. report that the correlation between the organizational
However, the two behaviors may differ on a variety of deviance scale and the interpersonal deviance scale is .68.
other dimensions (e.g. a public vs. private dimension, a Correcting this for unreliability results in a corrected
planned vs. unplanned dimension), making the rate of co- correlation of .86. Thus when one moves from
occurrence of the two an open question. correlations between individual behaviors to composites
How might one get insight into the rate of co- of behaviors within a category (i.e. organizational vs.
occurrence of counterproductive behaviors? While the interpersonal), the composite correlations between the
ideal would be the objective measurement of each form two scales are very high. It is not argued here that the
of counterproductive behavior, the Achilles' heel of two forms of counterproductivity are, in fact,
counterproductivity research must be acknowledged, inseparable: to the contrary, Bennett and Robinson do
namely, that while some forms of counterproductive show differential patterns of relationships with a number
behavior are public (e.g. absence), many are acts by of other constructs (e.g. a courtesy scale is more highly
employees who do not wish to be detected (e.g. theft, related to the interpersonal deviance scale than to the
sabotage, harassment). In the face of the difficulties of organizational deviance scale).
direct observation, data on the covariance of counter- Ashton (1998) reports a similar set of findings from a
productive behaviors comes from three sources: (a) self- study of college students with work experience.
report of the rate of occurrence; (b) judgments by others Individual self-report items dealing with eight counter-
(e.g. supervisors) of the rate of occurrence; and (c) direct productive behaviors all falling into the organizational
judgments about the rate of co-occurrence of counter- deviance category used by Bennett and Robinson (e.g.
productive behaviors. Note that both self-report and absence, tardiness, alcohol use, safety violations, theft)
other-report data are, like direct observational data, were used; the mean intercorrelation among the eight
subject to interpretational difficulties. The correlations was .30, and the internal consistency reliability for a
among self-reports could conceivably be inflated as a composite of the eight was .77. These findings closely
result of social desirability in responding, and the parallel those of Bennett and Robinson.
correlations among supervisor ratings may be inflated A different approach to self-reports of counter-
as a result of halo error. As none of these strategies is productive behavior was taken by Gruys (1999). First,
clearly ideal, exemplars of each of these are reviewed in she noted that many individual counterproductive
turn, and convergence of findings across strategies is behaviors are specific to certain work contexts, and that
examined. some respondents might never have had the opportunity
The first example of the use of a self-report strategy to engage in some of the behaviors (e.g. one cannot abuse
comes from Bennett and Robinson (2000). The goal of an employee discount if one has not worked in a setting
the study was to develop a self-report instrument to asses where such a discount is available). Thus Gruys asked
the degree to which individuals engaged in counter- respondents to consider that range of circumstances
productive behavior. An extensive instrument under which one could work and respond to behaviors
development and refinement process was used, and the using a scale anchored from `no matter what the
effort was guided by the conceptual framework circumstances, I would not engage in the behavior' to
developed in the Robinson and Bennett (1995) study `in a wide variety of circumstances, I would engage in the
described above. Specifically, the goal was an instrument behavior'. Second, ratings of 87 behaviors were obtained,
with separate scales to assess behaviors aimed at the and composite scores were obtained for the 11 behavior
organization and behaviors aimed at other individuals. categories presented at the opening of this chapter (theft,
The final instrument, based on responses from a broad destruction of property, misuse of information, misuse of
spectrum of 352 working adults, had a 12-item time and resources, unsafe behavior, poor attendance,

ß Blackwell Publishers Ltd 2002 Volume 10 Numbers 1/2 March/June 2002


8 PAUL R. SACKETT

poor quality work, alcohol use, drug use, inappropriate thus not permitting a direct comparison of these results
verbal actions, inappropriate physical actions). Thus with those obtained using self-report or other-report.
correlations are reported not between individual However, because Gruys had the same respondents
behaviors, but between behavior categories. Highly perform both the paired comparison likelihood of co-
similar findings emerged from a sample of 115 students occurrence task and the self-rating task described earlier,
and a sample of 343 college alumni; the focus here is on it is possible to compare the matrix of co-occurrence
the alumni sample. The mean correlation among the 11 ratings with the correlation matrix from the self-report
behavior category composites was .50; the higher values ratings, as both tasks involved the same 11 behavior
than those found in studies of individual behaviors are categories previously outlined. The correlation between
interpreted here as reflecting the higher reliability of the elements of the co-occurrence matrix and the
composites. Combining the 11 category composite into elements of the self-report correlation matrix was ÿ.17,
an overall grand composite results in an internal which was not significantly different from zero. Thus
consistency reliability of .92. perceptual data about likelihood of co-occurrence does
A second strategy for insight into the covariance of not produce a similar pattern of interrelationships among
counterproductive behaviors comes from ratings made by counterproductive behaviors from that resulting from
others (e.g. supervisors) of the degree to which self-report. At a more general level, though, the results do
individuals engage in various counterproductive behav- match one aspect of the data emerging from self-report
iors. Hunt (1996) reported a large-scale study using this and other-report research, namely, a pattern of positive
strategy, involving ratings of over 18,000 employees in 36 relationship among counterproductive behaviors. On a
companies. Individual ratings items were combined into scale where 1 means that two behaviors are very unlikely
composites; Hunt reports correlations among the to co-occur and 7 means that two behaviors are very
composites. Five of the dimensions rated fall into the likely to co-occur, the average rating was 4.2.
counterproductivity domain: attendance, off-task Thus self-report, other-report, and direct judgments of
behavior (e.g. unauthorized breaks, personal business likelihood of co-occurrence support the notion of
on work time), unruliness, theft, and drug misuse. The positive interrelationships among counterproductive
mean correlation among the composite measures for behaviors. Self-report data indicates positive correlations
these five dimensions was .50 ± a figure that corresponds in the range of .30 between individual counterproductive
precisely to the mean correlation reported in Gruys's behaviors, but higher correlations of about .50 between
study using composites obtained from self-reports of composites of related behaviors, a finding replicated with
counterproductivity. data using supervisor ratings. It appears reasonable to
A final strategy for insight into the covariance of think in terms of an overall counterproductivity
counterproductive behaviors involves obtaining direct construct, as the true score correlation between Bennett
judgments about the likelihood of co-occurrence of and Robinson's two domains of organizational and
various counterproductive behaviors. Gruys (1999) interpersonal deviance is .86, the reliability of a grand
employed this strategy in addition to the self-report overall composite across Gruys's 11 behavioral domains
research described above. Like Robinson and Bennett is .92, and the reliability of a grand composite across five
(1995), she obtained direct judgments of the similarity of behavioral domains in Hunt's work is .83.
counterproductive behaviors. However, while Robinson This proposal of an overall counterproductivity
and Bennett permitted respondents to construe construct does not argue against research focusing on
`similarity' in whatever fashion they chose, Gruys more specific forms of counterproductivity. What is
explicitly presented respondents with the task of making suggested here is a hierarchical model, with a general
judgments of the likelihood that a person who engaged in counterproductivity factor at the top, a series of group
one form of counterproductivity on the job would also factors, such as the organizational deviance and
engage in another. Gruys reports two dimensions interpersonal deviance factors identified by Bennett and
underlying these likelihood judgements. Like Robinson Robinson (2000) below this general factor, and specific
and Bennett, a dimension emerged distinguishing acts behavior domains, such as theft, absence, safety, and
that primarily harm the organization (e.g. theft, absence) drug and alcohol use below these group factors.
from acts that primarily harm other individuals (e.g. Researchers and practitioners may focus at difference
verbal and physical acts toward others). In contrast to levels of this hierarchy for different applications. For
Robinson and Bennett's minor vs. serious dimension, example, in many personnel selection settings
Gruys's second dimension differentiated acts that detract organizations are interested in identifying prospective
from job performance (e.g. absence, intentionally doing employees who will not engage in the broad range of
poor quality work, safety violations) from harmful acts counterproductive behaviors, and thus may focus on the
in the workplace not directly related to job performance broad counterproductivity construct. In contrast, an
(e.g. theft). The direct judgements of likelihood of co- intervention may be sought that will deal effectively
occurrence do not translate into a correlational metric, with a single specific problem behavior (e.g. widespread

International Journal of Selection and Assessment ß Blackwell Publishers Ltd 2002


THE STRUCTURE OF CWB 9

violation of safety procedures). The perspective taken Hough, Toquam, Hanson and Ashworth 1990). Project A
here is that the decision as to where to focus one's investigated a broad array of predictor and criterion
intervention and/or measurement efforts on this variables across a set of military jobs. Of great interest is
continuum from general factor to specific behaviors is the focus on the careful identification of constructs
best made by recognizing the interrelationships among underlying measured variables. Thirty-two different
counterproductive behaviors. criterion measures were obtained; factor analytic work
was done to identify a set of five criterion constructs
underlying these measures. These constructs include core
Counterproductive Behaviors and Job technical proficiency, general soldiering proficiency,
Performance effort and leadership, physical fitness and military
bearing, and maintaining personal discipline. The two
The previous section focused on interrelationships soldiering proficiency measures reflect the task perfor-
among different forms of counterproductive behavior. mance domain; effort and leadership the citizenship
The relationship between counterproductive behaviors domain, and maintaining personal discipline the counter-
and other behaviors making up the broad domain of job productive behavior domain. Construct-level scores were
performance is now considered. The most prominent obtained by combining different measures, including
contemporary framework for viewing job performance is work samples, supervisor ratings, and indicators from
that of Campbell (Campbell et al. 1993), who offers eight administrative records, including the number of
performance components: job-specific task proficiency, disciplinary infractions. Criterion scores were obtained
non-job specific task proficiency, written and oral for 4,039 soldiers in nine military enlisted jobs.
communication, demonstrating effort, maintaining The key findings are observed uncorrected mean
personal discipline, facilitating peer and team perfor- correlations of ÿ.19 and ÿ.17 between counter-
mance, supervision/leadership, and management/ productivity and the general and specific task perfor-
administration. The maintaining personal discipline mance dimensions, and a mean correlation of ÿ.59
dimension reflects the counterproductivity domain as between counterproductive behavior and the effort/
discussed here. leadership construct. Thus the relationship between
A series of interrelated frameworks have been offered counterproductive behavior and quantity and quality of
that focus on a set of behaviors variously labeled as task performance is quite low, in contrast to the
citizenship behaviors (Smith, Organ and Near 1983) relationship between counterproductive behavior and
prosocial behaviors, (Brief and Motowidlo 1986), and effort/leadership, which is quite high.
contextual performance (Borman and Motowidlo 1993). The second large data set is the work of Hunt (1996),
What these have in common is a focus on positive referenced earlier in the context on the interrelationship
behaviors that contribute to organizational effectiveness, among various forms of counterproductive behavior.
but that do not reflect core job tasks. These include Hunt's focus was on what he termed `generic work
helping others, persistence and extra effort, and behavior', namely, behaviors common across jobs and
supporting the organization. While there are differences not specific to the tasks of any given jobs. The earlier
in emphasis in these different frameworks (e.g. some discussion focused on five dimensions of counter-
frameworks require that a behavior be discretionary, i.e. productive behavior; also relevant here are two dimen-
not formally rewarded by the organization, while others sions labeled industriousness and persistence, which
do not include this restriction), the behavioral domains correspond conceptually to the effort/leadership
covered by these frameworks are largely overlapping. An dimension in Project A, and to the broad construct we
emerging literature differentiates and contrasts the task are labeling citizenship. Hunt derived composite
performance domain and the citizenship/prosocial/ measures of these dimensions using a sample of over
contextual performance domain (referred to as 18,000 supervisory ratings across 36 organizations.
citizenship, as a term to reflect the broad domain, e.g. For this article, the psychometric theory of composites
Conway 1999; Motowidlo and Van Scotter 1994). was used to estimate the correlation between a
Adding the counterproductive behavior domain that is citizenship composite made up of Hunt's industriousness
the focus of this article leads to a broad conception of and persistence dimensions and a counterproductive
three primary performance domains: task performance, behavior composite, made up of Hunt's five counter-
citizenship performance, and counterproductive behav- productive behavior dimensions. The resulting
ior, and prompts questions as to relationships between correlation is ÿ.67, which is quite similar to the value
counterproductive behavior and the other two domains. of ÿ.59 obtained in Project A.
The focus here in examining this issue is primarily on The final source is a meta-analysis by Viswesvaran,
three large multi-sample data sets. The first is the US Schmidt and Ones (1999) of the interrelationship among
Army Selection and Classification Project, commonly supervisor ratings. They sorted ratings from the
referred to as Project A (Campbell 1990; McHenry, published literature into eight dimensions. Sample sizes

ß Blackwell Publishers Ltd 2002 Volume 10 Numbers 1/2 March/June 2002


10 PAUL R. SACKETT

varied by dimensions, and ranged from roughly 2,000 to produces stronger cross-category correlations, it is also
11,000. Here ratings of job knowledge, quantity of the case that aggregation across categories yields broad
output and quality of output are combined into the broad overall measures of counterproductive behavior with
dimension of task performance, ratings of interpersonal internal consistency reliability in the .8ÿ.9 range. Similar
competence, effort, and leadership into the broad findings emerge from research using self-report measures
dimension of citizenship, and ratings of compliance/ and from research using evaluations by others (e.g.
acceptance of authority corresponded to the domain of supervisor ratings).
counterproductive behavior. The mean unobserved While behavior categories are substantially
correlation between counterproductive behavior and interrelated, the relationship is not so strong as to
ratings in the citizenship and task domains was ÿ.57 question the value of measuring or attempting to modify
and ÿ.54 respectively. behavior in a specific category. Thus for some purposes it
There is clear convergence with respect to the may be useful to focus on a single category, as in the case
relationship between counterproductive behavior and of an organizational intervention in a setting where a
the citizenship domain: mean observed r is ÿ.59 in particular behavior or behavior category is of great
Project A, ÿ.67 in Hunt, and ÿ.57 in Viswesvaran et al. concern (e.g. an intervention aimed at curbing increasing
In contrast, findings regarding the relationship between absenteeism). At the same time, the sizable inter-
counterproductive behavior and task performance are relationships between categories of counterproductive
quite discrepant: a mean r of ÿ.18 in Project A and a behavior indicate that it will often be useful to view
mean r of ÿ.54 in the Viswesvaran et al. meta-analysis. It counterproductive behavior broadly. The fact of these
is posited here that these differences reflect differences in interrelationships indicates that shared antecedents of
the conceptualization and measurement of task different types of counterproductive behaviors are likely,
performance in the two studies. The task performance and the acquisition of knowledge about the counter-
measures in Project A were work sample and job productivity domain would be leveraged by the inclusion
knowledge measures. They reflect measures of maximum of multiple categories of counterproductive behavior in
performance: what the individual `can do' when individual studies.
performance is closely monitored. In contrast, the Once counterproductive behavior is considered in the
Viswesvaran et al. study was restricted to supervisor aggregate, rather than as a large number of individual
ratings, which can generally be seen as reflecting behaviors, the question of the interrelationships between
measures of typical performance: what the individual counterproductive behaviors and other facets of job
`will do' over an extended period of time (Sackett, performance arises. Current conceptualizations of job
Zedeck and Fogli 1988). The `can do' measures are performance identify a task performance domain and a
primarily a function of knowledge and skill, while the set of closely interrelated domains labeled as prosocial
`will do' measures are also influenced by the full range of behavior, citizenship behavior and contextual
individually and situationally driven motivational factors performance by various scholars; the label `citizenship
± factors that also affect counterproductive behavior. behavior' is adopted here. This article examined three
Thus counterproductive behavior would be expected to large data sets that shed light on the interrelationships
correlate more highly with typical task performance than among task performance, citizenship behavior, and
with maximum task performance. counterproductive behavior. Relationships between task
performance and counterproductive behavior vary quite
widely across studies, with very low relationships found
Conclusion when task performance is operationalized as task
proficiency: what the employee can do. Much stronger
The aim of this article was to examine issues and key relationships are found when both facets of performance
literatures related to the structure of counterproductivity, are obtained by the same measurement method (e.g.
including both the interrelationships among various supervisor ratings) and when task performance is
counterproductive behaviors, and the relationship of operationalized as typical task performance: what the
counterproductive behavior to other facets of job employee will do.
performance. The most general conclusion is the Of considerable interest is the convergence across
consistent finding of positive interrelationships among disparate data sets as to the relationship between
the full range of counterproductive behaviors. The citizenship behaviors and counterproductive behaviors.
strength of the relationship increases as one aggregates Correlations of about ÿ.60 emerge consistently, leading
from individual behaviors to sets of behaviors within a to the consideration of whether the citizenship and
broader conceptual category (e.g. the theft-attendance counterproductive domains should be viewed as opposite
relationship is weaker with single items measuring each poles of a single dimension, with citizenship reflecting the
one than when multi-item scales are used for each positive pole and counterproductive behavior the
behavior category). While aggregation within a category negative pole (Puffer 1987). The perspective taken here

International Journal of Selection and Assessment ß Blackwell Publishers Ltd 2002


THE STRUCTURE OF CWB 11

is that while two negatively correlated variables may behaviors. Academy of Management Review, 11, 710±725.
usefully be combined into a composite for some Campbell, J.P. (1990) An overview of the Army selection and
classification project (Project A). Personnel Psychology, 43,
purposes, the treatment of the two as reflecting a single 231±239.
continuum has implications of mutual exclusivity, Campbell, J.P., McCloy, R.A., Oppler, S.H. and Sager, C.E.
namely, that the fact that an individual with high (1993) A theory of performance. In N. Schmitt and W.
standing on citizenship cannot also be high in counter- Borman (eds) Personnel Selection in Organizations. San
productive behavior, and vice versa. There are ready Francisco: Jossey-Bass.
Conway, J.M. (1999) Distinguishing contextual performance
examples of the highly productive employee who is also from task performance for managerial jobs. Journal of
engaging in extensive counterproductive behavior (as Applied Psychology, 84, 3±13.
when a high performer viewed as beyond suspicion is Griffin, R.W., O'Leary-Kelly, A. and Collins, J. (1998)
caught embezzling). The fact of a high correlation Dysfunctional work behaviors in organizations. In C.L.
between the two domains can be usefully used in Cooper and D.M. Rousseau (eds) Trends in Organizational
Behavior, vol. 5. Chichester: John Wiley and Sons Ltd.
research on common antecedents, and a composite of Gruys, M.L. (1999) The dimensionality of deviant employee
the two reflecting an individual's contribution to the performance in the workplace. Unpublished doctoral
organization can be created without adopting a bipolar dissertation, University of Minnesota.
single dimensional view. Guastello, S.J. and Rieke, M.L. (1991) A review and critique of
In sum, this article argues for the value of a broad and honesty test research. Behavioral Sciences and the Law, 9
(Fall), 501±523.
integrative focus on counterproductive behavior, for Hollinger, R.C. and Clark, J.P. (1983) Theft By Employees.
searching for common antecedents in light of the Lexington, MA: D.C. Heath & Company/Lexington Books.
interrelationships among various individual counter- Hunt, S.T. (1996) Generic work behavior: An investigation into
productive behaviors, and for viewing counterproductive the dimensions of entry-level, hourly job performance.
behavior as an important facet of job performance. Personnel Psychology, 49, 51±83.
McHenry, J.J., Hough, L.M., Toquam, J.L., Hanson, M.A. and
Ashworth, S. (1990) Project A validity results: The
relationship between predictor and criterion domains.
Note Personnel Psychology, 43, 335±354.
Motowidlo, S.J. and Van Scotter, J.R. (1994) Evidence that task
1. This article contains material that is part of a larger performance can be distinguished from contextual
performance. Journal of Applied Psychology, 79, 475±480.
chapter entitled `Counterproductive behaviors at Puffer, S.M. (1987) Prosocial behavior, noncompliant behavior,
work', authored by Paul R. Sackett and Cynthia J. and work performance among commission salespeople.
DeVore, to be published in N. Anderson, D. Ones, H. Journal of Applied Psychology, 72, 615±621.
Sinangil and C. Viswesvaran (eds) International Robinson, S.L. and Bennett, R.J. (1995) A typology of deviant
Handbook of Work Psychology. Sage Publications. workplace behaviors: A multidimensional scaling study.
Academy of Management Journal, 38, 555±572.
Robinson, S.L and Greenberg, J. (1998) Employees behaving
badly: Dimensions, determinants, and dilemmas in the study
References of workplace deviance. In C.L. Cooper and D.M. Rousseau
(eds) Trends in Organizational Behavior. New York: John
Ashton, M.C. (1998) Personality and job performance: The Wiley & Sons.
importance of narrow traits. Journal of Organizational Sackett, P.R., Zedeck, S. and Fogli, L. (1988) Relationships
Behavior, 19, 289±303. between measures of typical and maximum performance.
Bennett, R.J. and Robinson, S.L. (2000) The development of a Journal of Applied Psychology, 73, 482±486.
measure of workplace deviance. Journal of Applied Smith, C.A., Organ, D.W. and Near, J.P. (1983) Organizational
Psychology, 85, 349±360. citizenship behavior: Its nature and antecedents. Journal of
Borman, W.C. and Motowidlo, S.J. (1993) Expanding the Applied Psychology, 68, 653±663.
criterion domain to include elements of contextual Viswesvaran, C., Schmidt, F.L. and Ones, D.S. (1999) The role
performance. In N. Schmitt and W. Borman (eds) Personnel of halo error in interdimensional ratings: The case of job
Selection in Organizations. San Francisco: Jossey-Bass. performance ratings examined via meta-analysis.
Brief, A.P. and Motowidlo S.J. (1986) Prosocial organizational Manuscript submitted for publication.

ß Blackwell Publishers Ltd 2002 Volume 10 Numbers 1/2 March/June 2002

You might also like