ArticlePDF Available

Predicting the Counterproductive Employee in a Child-to-Adult Prospective Study

Authors:

Abstract and Figures

The present research tested the relations between a battery of background factors and counterproductive work behaviors in a 23-year longitudinal study of young adults (N = 930). Background information, such as diagnosed adolescent conduct disorder, criminal conviction records, intelligence, and personality traits, was assessed before participants entered the labor force. These background factors were combined with work conditions at age 26 to predict counterproductive work behaviors at age 26. The results showed that people diagnosed with childhood conduct disorder were more prone to commit counterproductive work behaviors in young adulthood and that these associations were partially mediated by personality traits measured at age 18. Contrary to expectations, criminal convictions that occurred prior to entering the workforce were unrelated to counterproductive work behaviors. Job conditions and personality traits had independent effects on counterproductive work behaviors, above and beyond background factors.
Content may be subject to copyright.
Management Department
Management Department Faculty Publications
University of Nebraska - Lincoln Year 
Predicting the Counterproductive
Employee in a Child-to-Adult
Prospective Study
Brent W. Roberts
Peter D. Harms
Avshalom Caspi
Terri E. Moffitt
∗∗
University of Illinois, Urbana–Champaign, bwrobrts@illinois.edu
University of Nebraska - Lincoln, pharms2@Unl.edu
King’s College, London, & University of Wisconsin–Madison
∗∗
King’s College, London, & University of Wisconsin–Madison
This paper is posted at DigitalCommons@University of Nebraska - Lincoln.
http://digitalcommons.unl.edu/managementfacpub/45
According to Ones (2002), “counterproductive work behav-
iors [CWBs] include but are not limited to theft, white collar
crime, absenteeism, tardiness, drug and alcohol abuse, disci-
plinary problems, accidents, sabotage, sexual harassment, and
violence” (p. 1). Clearly, employers would prefer to not hire
individuals with the propensity to engage in CWBs. One of the
primary techniques used to avoid hiring counterproductive
employees is to screen applicants using past behaviors (crim-
inal background checks) and psychological tests, such as in-
tegrity and personality tests (Ones, Viswesvaran, & Schmidt,
1993). The assumption behind screening applicants is that peo-
ple’s history, both in terms of what they have done and what
they typically do, will predict how they behave on the job. This
assumption is derived in part from the perspective that there
will be some continuity in behavior and personality, such that
past behavior will predict future behavior and personality
traits will be stable over long periods of time.
Ideally, the validity of personality tests and background
checks would be established using a prospective longitudi-
nal study in which information on employees is gathered, us-
ing valid techniques, as their lives unfold. For many reasons,
this ideal study design is elusive, as most validity research is
carried out in organizations that ask job applicants to provide
background information. Such applicants provide either self-
reports of relevant variables or retrospective accounts of their
activities using biodata collection techniques. Asking people
directly whether they have committed a crime or stolen from
their previous employer may not be the ideal technique, as
people can lie in the attempt to gain employment. Even when
a person attempts to be honest, retrospective reports of past
delinquent behaviors can be quite unreliable (Henry, Moftt,
Caspi, Langley, & Silva, 1994). Of course, background checks
only yield relevant information if the person has been arrested,
prosecuted, and convicted; has poor credit; or has led for
bankruptcy. Crime records are known to underdetect crimes
committed, as most offenses are never cleared or successfully
prosecuted. In reality, most employers must rely on the hon-
esty of applicants or their references for gathering historical
facts about their behavior.
In the present study, we present an alternative approach
to understanding the relationship of background factors to
CWBs. We took advantage of data gathered as part of the
Dunedin Multidisciplinary Health and Development Study
(Silva & Stanton, 1996). The participants in the Dunedin study
are a birth cohort of New Zealand residents who have been
tracked from birth to age 26 with good retention. Approx-
imately every 3 years, the participants were exhaustively in-
terviewed by physicians and psychologists. Along the way,
much relevant information was gathered that can be used to
predict CWBs. For the present study, we used background in-
formation collected during childhood and adolescence to pre-
dict CWBs on the job at age 26.
This approach is different than the typical validation study
in which a select group of job incumbents is surveyed and fol-
lowed in a particular occupation or set of occupations (Lau,
Au, & Ho, 2003; Sackett & DeVore, 2001). The typical valida-
tion study provides predictive validity information for con-
structs that can be assessed within a typical selection procedure
Published in Journal of Applied Psychology 92:5 (September 2007), pp. 1427–1436; doi: 10.1037/0021-9010.92.5.1427 Copyright © 2007 American
Psychological Association. Used by permission. “This article may not exactly replicate the nal version published in the APA journal. It is not the
copy of record.” http://www.apa.org/journals/jap/
Preparation of this article was supported by National Institute of Aging Grant AG19414, National Institute of Mental Health Grants MH49414 and
MH45070, and United Kingdom Medical Research Council Grant G0100527. The Dunedin Multidisciplinary Health and Development Research
Unit is supported by the Health Research Council of New Zealand. Our thanks go to the study participants and to unit director Richie Poulton for
comments on an earlier version of this article.
Submitted April 4, 2006; revised January 10, 2007; accepted: January 12, 2007.
Predicting the Counterproductive Employee in a
Child-to-Adult Prospective Study
Brent W. Roberts & Peter D. Harms
Department of Psychology, University of Illinois, Urbana–Champaign
Avshalom Caspi & Terri E. Moftt
Institute of Psychiatry, King’s College London, London, England, and Department of Psychology, University of Wisconsin–Madison
Corresponding author — Brent W. Roberts, Department of Psychology, University of Illinois, Urbana–Champaign,
603 East Daniel Street, Champaign, IL 61820; email broberts@cyrus.psych.uiuc.edu.
Abstract
The present research tested the relations between a battery of background factors and counterproductive work behaviors in a 23-year longitudi-
nal study of young adults (N = 930). Background information, such as diagnosed adolescent conduct disorder, criminal conviction records, intelli-
gence, and personality traits, was assessed before participants entered the labor force. These background factors were combined with work condi-
tions at age 26 to predict counterproductive work behaviors at age 26. The results showed that people diagnosed with childhood conduct disorder
were more prone to commit counterproductive work behaviors in young adulthood and that these associations were partially mediated by per-
sonality traits measured at age 18. Contrary to expectations, criminal convictions that occurred prior to entering the workforce were unrelated to
counterproductive work behaviors. Job conditions and personality traits had independent effects on counterproductive work behaviors, above
and beyond background factors.
Keywords: personality traits, counterproductive work behaviors, longitudinal, predictive validity
1427
1428 Ro b e R t s e t al . i n Jo u r nal o f ap p l i ed ps y c h o l o g y 92 (2007)
on a particular sample of individuals. In contrast, the present
study entailed an examination of the validity of a combination
of constructs, some that are common in selection procedures,
such as personality tests, and some that are not typically as-
sessed in a selection process, such as childhood variables. This
type of design can be used for several purposes, including test-
ing whether the typical selection constructs can serve as prox-
ies for background variables and seeing whether they are con-
founded by these constructs. The present study also focused
on a cross-section of individuals across a large number of jobs.
Therefore, the ndings are more difcult to generalize to any
particular occupation. However, akin to a generalizability
study, the ndings may then be generalizable to a larger num-
ber of occupations. In general, then, the present design pro-
vides complementary information on the long-term predictive
validity of a number of factors used in selection procedures.
The goal was to provide a more complete picture of how these
factors might affect CWBs.
In terms of previous behaviors, the Dunedin study has long
examined behavioral phenomena distinctly similar to CWBs,
such as childhood and adolescent conduct disorder, as well
as criminal activity that occurred during adolescence (Moftt,
Caspi, Rutter, & Silva, 2001). From a behavioral perspective,
there is a remarkable overlap between CWBs and the typical be-
haviors used to diagnose someone with conduct disorder. For
example, both the counterproductive worker and the person di-
agnosed with conduct disorder are described as not conform-
ing to social norms, breaking laws, lying, committing acts of vi-
olence such as physical ghts and assaults, and disregarding the
safety of others; see the Diagnostic and Statistical Manual of Men-
tal Disorders (4th ed.; DSM–IV; American Psychiatric Associa-
tion, 1994). Therefore, we hypothesized that adolescent behav-
iors indicative of conduct disorder would predict adult CWBs.
Criminal activities would appear to be highly relevant to
counterproductive behaviors, as many of the actions consid-
ered to be counterproductive are also criminal. Given the rel-
atively strong relationship between past behavior and present
behavior (Meyer et al., 2001), one would assume that criminal
activities and CWBs would be positively correlated. This is,
of course, the justication for the widespread use of criminal
background checks by employers, as businesses can be held li-
able for employees’ actions, especially if they result in harm
to other workers or clients (Connerley, Arvey, & Bernardy,
2001). Surprisingly, though, there is little or no published re-
search reporting the relationship between having a criminal
background and engaging in subsequent counterproductive
behaviors on the job. In the present study we examined both
conduct disorder and criminal conviction assessed in adoles-
cence as prospective predictors of CWBs in adulthood.
The Dunedin study has also assessed cognitive ability and
personality prospectively. In the current study, we drew on
IQ scores prospectively assessed during childhood using the
Weschler tests of cognitive ability. The relationship between
cognitive ability and CWBs is unclear, as few published stud-
ies have reported on the direct relationship. Cognitive abil-
ity tends to be relatively unrelated to similar constructs such
as contextual performance (LePine & Van Dyne, 2001), yet
there is some evidence that it is negatively related to organi-
zational deviance and facets of integrity, such as theft admis-
sions (Duehr, Sackett, & Ones, 2006; Dilchert, Ones, Davis, &
Rostow, 2007). Therefore our expectation was that childhood
cognitive ability would be negatively related to CWBs.
Research to date supports the inference that personal-
ity tests are related to CWBs. Although different measures
of personality and different systems of describing personal-
ity traits have been examined, most research has found that
low agreeableness (high hostility) and low conscientiousness
are key correlates of CWBs (Berry, Ones, & Sackett, 2007; Col-
bert, Mount, Harter, Witt, & Barrick, 2004; Lau et al., 2003; Lee,
Ashton, & Shin, 2005; Ones & Viswesvaran, 2003; Sackett &
DeVore, 2001; Salgado, 2002). Some research has also found a
substantial effect for a trait variously called low emotional sta-
bility, negative affectivity, or neuroticism that predicts CWBs
such as substance abuse at work (Ones & Viswesvaran, 2001).
Potentially, the most relevant aspect of negative affectivity in
relation to CWBs is hostility, a form of low agreeableness. Lee
and Allen (2002) demonstrated that only the hostility aspect of
negative affect showed signicant relationships with a broad
range of CWBs. Further, hostility remained one of the best pre-
dictors of CWBs even when a number of other background
factors, such as age, education, and organizational character-
istics, were taken into account. In the present study we exam-
ined these traits assessed in adolescence as prospective predic-
tors of CWBs in adulthood.
To evaluate the role of personality in CWBs in young
adulthood, we drew on self-reported personality assessments
gathered at age 18 using the Multidimensional Personal-
ity Questionnaire (MPQ; Tellegen, 1982). The MPQ includes
measures that tap the facets of Agreeableness, Conscientious-
ness, and Neuroticism related to CWBs. Specically, these
scales are found in the superordinate MPQ domains of Neg-
ative Emotionality and Constraint. Negative Emotionality in
Tellegen’s system includes aggressiveness, which is typically
a marker of low Agreeableness in the Big Five framework.
Negative emotionality also includes dimensions that tap the
anxiety and alienation common to measures of Neuroticism.
Constraint is a variant of Conscientiousness that emphasizes
the impulse control and conventionality aspects of Conscien-
tiousness over the typical orderliness and achievement com-
ponents of Conscientiousness (Roberts, Chernyshenko, Stark,
& Goldberg, 2005). We also examined the relations between
CWBs and the remaining two domains tapped by the MPQ,
Agentic Positive Emotionality and Communal Positive Emo-
tionality, with no a priori expectations for how these trait do-
mains should relate to CWBs.
We drew on measures of CWBs administered when the
sample was assessed at age 26. At this age, the study mem-
bers occupied a wide range of jobs from slaughterer to gov-
ernment ofcer. In predicting CWBs, we also included organi-
zational and workplace variables assessed at age 26 that were
developed in previous research on the Dunedin study (Rob-
erts, Caspi, & Moftt, 2003). These variables include many fac-
tors that have been investigated in research on CWBs, such
as job satisfaction (Robinson & O’Leary-Kelly, 1998), occupa-
tional attainment, power, and autonomy (Hollinger & Clark,
1983; Robinson & O’Leary-Kelly, 1998).
The present design offers two advantages. First, most of
the relevant variables were assessed well before the majority
of the participants entered the labor market. Thus, we were
able to test the prospective effect of background factors such
as adolescent conduct disorder, criminal conviction, and ado-
lescent personality on subsequent CWBs in work. Second, we
included both psychological and contextual factors in the pre-
diction of CWBs under the assumption that both personality
PR e d i c t i n g th e co u n t e R PRo d u c t i v e emPl o y e e i n a ch i l d -t o -adul t PR o s P e c tiv e st u d y 1429
traits and work conditions inuence CWBs. On the basis of
previous research, we made the following hypotheses:
Hypothesis 1: Adolescent conduct disorder will be posi-
tively related to CWBs at age 26.
Hypothesis 2: Adolescent criminal convictions through
age 18 will be positively related to CWBs at age 26.
Hypothesis 3: Cognitive ability assessed in childhood will
be negatively related to CWBs at age 26.
Hypothesis 4: Personality traits from the domain of Con-
scientiousness–Constraint will be negatively related to
CWBs at age 26.
Hypothesis 5: Personality traits from the domain of Nega-
tive Emotionality (inclusive of aggression, which is also
low Agreeableness) will be positively related to CWBs
at age 26.
Hypothesis 6: Most variables representing desirable occu-
pational conditions, such as satisfaction, occupational
attainment, and autonomy, will be negatively related
to CWBs at age 26.
Method
Participants
Participants were members of the Dunedin study (see Silva
& Stanton, 1996), a longitudinal investigation of the health and
behavior of a complete cohort of consecutive births born be-
tween April 1, 1972, and March 31, 1973, in Dunedin, New
Zealand. When the children were traced for follow-up at 3
years of age, 1,037 children (91% of the eligible births, of whom
52% were boys) participated in the assessment and formed the
base sample for the longitudinal study. With regard to social
origins, the children’s families are representative of the social
class and ethnic distribution in the general population of New
Zealand’s South Island. With regard to ethnic distribution, the
Dunedin study members are of predominantly European an-
cestry. Cross-national comparisons and replication analyses
lend some condence about generalizing ndings from the
Dunedin study to other Western nations (see Moftt, Caspi,
Rutter, & Silva, 2001). Follow-ups of the sample have been
carried out at ages 5, 7, 9, 11, 13, 15, 18, 21, and 26, when 980
(96%) of the 1,019 study members still alive were assessed. The
present sample (N = 930) included all individuals who indi-
cated that they were employed currently or at some time dur-
ing the year prior to the interview (i.e., not homemakers) and
thus could report about a job experience in the past 12 months.
Study members who were in prison on their interview date re-
ported about their most recent job.
Measures
Predictors in this study were measured up to age 18, stop-
ping at that age because up to age 18 most of the cohort mem-
bers attended secondary school, lived at home with their par-
ents, and did not work full-time. Thus, predictors assessed
up to age 18 were used because it was very unlikely that the
scores would have been affected by cohort members’ experi-
ence of employment or by any CWBs committed in the jobs
held by cohort members.
Adolescent conduct disorder. Conduct disorder was measured
according to DSM–IV criteria, which identify adolescents dis-
playing a persistent pattern of behavior that violates the rights
of others, including physical harm. A diagnosis of conduct dis-
order (using a 12-month reporting period for symptoms) was
made when the research participants at each of four ages were
assessed (ages 11, 13, 15, and 18). A “lifetime” diagnosis was
arrived at by establishing whether a study member received
the diagnosis at one or more of the four ages (according to the
DSM–IV, conduct disorder is not normally diagnosed after
age 18). The internal consistency of the conduct disorder items
averaged across multiple judges and rating occasions was .83
(Moftt et al., 2001).
Adolescent criminal conviction. Computerized records of con-
victions at all courts in New Zealand and Australia for 932
study members were obtained by searching the central com-
puter system of the New Zealand Police for matches by name
and birth date. These records included convictions in Chil-
dren’s and Young Person’s Court from age 13 to age 16 in-
clusive and convictions in adult Criminal Courts from age 17
until the 18th birthday. Trafc offenses and criminal offenses
were handled by two separate law-enforcement agencies in
New Zealand at that time, and as a result trafc offenses are
not included in our criminal conviction measure, which covers
mainly property crimes, drug crimes, and violence. Informed
consent for the search was obtained during the age-18 inter-
views. Conviction data could not be obtained for study mem-
bers who did not participate in the assessment, were deceased,
did not give informed consent for the search, or lived out-
side New Zealand and Australia (the 22 study members who
did not provide consent for the search did not differ from the
whole cohort on self-reported or parent-reported delinquency
at age 15). The number of court convictions ranged from 0 to
68; 25% of male participants and 12% of female participants
had been convicted at least once.
Childhood cognitive ability. Intelligence was assessed us-
ing the Wechsler’s Intelligence Scale for Children—Revised
(Wechsler, 1974), administered at ages 7, 9, 11, and 13 years.
The IQ scores for the four age periods were averaged to form
an overall score (M = 107.7, SD = 14.3, range = 40–147).
Personality traits. As part of the age-18 assessment, partic-
ipants completed a modied version (Form NZ) of the MPQ
(Tellegen, 1982; see also Krueger, Caspi, & Moftt, 2000). The
MPQ is a self-report personality instrument designed to as-
sess a broad range of individual differences in affective and
behavioral style and yields 10 primary scales (the Absorption
scale was not included in the administration of the MPQ).
These 10 primary scales can be organized under a three-fac-
tor structure (Negative Emotionality, Positive Emotionality,
and Constraint).
Negative Emotionality is a combination of the Aggression,
Alienation, and Stress Reaction scales. Individuals high on this
dimension have a low general threshold for the experience
of negative emotions such as fear, anxiety, hostility, and an-
ger and tend to be involved in antagonistic relationships. Con-
straint is a combination of the Traditionalism, Harm Avoid-
ance, and Self-Control scales. Individuals high on this factor
tend to endorse social norms, act in a cautious and restrained
manner, and avoid thrills. Positive Emotionality is a combina-
tion of the Social Closeness, Well-Being, Social Potency, and
Achievement scales and reects positive emotional respon-
siveness, interpersonal connectedness, and dominance. We
1430 Ro b e R t s e t al . i n Jo u r nal o f ap p l i ed ps y c h o l o g y 92 (2007)
scored these superfactors by summing the relevant subscales
and used these as a supplement to the primary MPQ scales.
Measures Assessed at Age 26
Work conditions and attitudes. The participants’ workplace
environment was assessed with several variables. The rst
variable was occupational attainment, which is a composite
of job prestige, job complexity, education level, pretax hourly
wages, and whether participants got dirty on the job (reversed;
α = .75; see Roberts, Caspi, & Moftt, 2003). The second vari-
able was work autonomy, which was a composite of whether
participants had a boss (reversed), whether their boss had a
boss (reversed), whether they could set their own hours, and
whether they were responsible for a budget (α = .63). The third
workplace variable was resource power, which was a combi-
nation of items tapping whether the person had a say in hir-
ing or ring people, whether they had a say in pay raises and
pay cuts, whether they supervised others, and the number of
people they supervised = .67). The nal workplace factor,
work satisfaction, was measured with self-report interview
questions that tapped the Dunedin study members’ satisfac-
tion with work (Greenberger & O’Neil, 1993). Sample items
from the work satisfaction scale included “How satised are
you in this job?” scored on a 3-point scale with response op-
tions ranging from not satised (0) to very satised (2), “Do you
often think of quitting your job?” scored on a 2-point scale (0
= yes, 1 = no), and “Is this the right job for you?” scored on a 2-
point scale (1 = yes, 0 = no). Items were z scored before being
combined into a scale. The ve-item job satisfaction scale had
an alpha reliability of .69.
Counterproductive work behavior. Work experiences were as-
sessed by interviewing Dunedin study members about their
work as part of a life-history interview inquiring about their
transition to adulthood. Counterproductive workplace be-
haviors were assessed with 11 interview questions concern-
ing the frequency of the following behaviors over the previ-
ous year: number of times late to work, number of days absent
under pretense, number of times using prohibited work ma-
terials, number of conicts with boss, number of ghts–argu-
ments at work, number of times doing something at work that
could get them red, number of times they stole money from
work, number of times they lied on their time sheet, number of
times they stole things from work, number of times they dam-
aged work property, and number of times they were drunk or
on drugs at work. Rather than using direct frequencies, which
have problematic psychometric properties for low-base-rate
behaviors, we rst categorized scores on each behavior for the
presence (scored 1) or absence (scored 0) of the behavior. We
then subjected these 11 items to an item response theory (IRT)
analysis to determine the dimensionality of the index and to
properly score the index.
Because the items were dichotomous and ranged in their
endorsement frequency, we used IRT methods to determine
the structure and score the overall index of CWBs. For most
IRT models, it is assumed that a single underlying latent trait
or ability is sufcient to account for examinee performance.
To test the unidimensionality assumption, we implemented
a nonlinear factor analysis procedure. Specically, we used
the method of comparing the ratio of rst to second eigenval-
ues (obtained using an iterated principal-axis factor analysis)
for each within-scale matrix of tetrachoric correlations (Lord,
1980). There appeared to be support for the assumption that
one dominant dimension is being tapped by the scale’s items,
as one dimension accounted for a substantial portion of the
variance (34%).
BILOG (Mislevy & Bock, 1990) was used to estimate item
parameters, as well as participants’ latent trait. We t a 2-Pa-
rameter Logistic IRT model to the workplace deviancy items,
which involves estimating the discrimination (a) and difculty
(b) parameters. For the most part, the BILOG defaults were
used, which means that the method of marginal maximum
likelihood was used to estimate the difculty parameters, and
the discrimination parameters were estimated using marginal
maximum a posteriori, a Bayesian procedure (see Hambleton,
Swaminathan, & Rogers, 1991). We used the Bayesian proce-
dure of expected a posteriori (see Hambleton et al., 1991) to
obtain theta estimates for each participant. In terms of reliabil-
ity, the Cronbach’s alpha for the set of dichotomous items was
.60. Examination of the test information function (TIF) showed
that the scale provided reliable information for the high end of
θ, from 0 to 2.5 on the θ scale.
Although IRT analyses suggested that internal consistency
of the CWB measure is good, this study lacks evidence for test–
retest reliability or interreporter reliability for the CWB mea-
sure. However, it is relevant that the members of this birth co-
hort have been repeatedly interviewed about delicate private
topics such as illegal behavior, drug abuse, sexuality, domes-
tic violence, and suicide attempts, and by age 26 cohort mem-
bers had learned that they could trust the study team’s guar-
antee of condentiality. Thus, there is good reason to believe
that cohort members’ self reports of past-year CWBs were
trustworthy.
Results
The percentages for the items that make up the Counterpro-
ductive Work Behavior Index are shown in Table 1. The partic-
ipation rates varied markedly, with over half of the sample be-
ing late to work at least once and 1% of the sample admitting
to stealing money from work.
Correlational Analyses
The raw correlations between background factors and
CWBs also are shown in Table 1. As hypothesized, adoles-
cent conduct disorder was a statistically signicant predictor
of CWBs in young adulthood (r = .11, p < .05). Our second hy-
pothesis, concerning past ofcial conviction for criminal activ-
ity predicting CWBs, was not supported. Adolescent criminal
convictions were unrelated to committing counterproductive
activities at work. In fact, according to the item-level statistics,
people with an adolescent criminal conviction record were less
likely to get in a ght with their supervisor or to steal things
from work. Also, contradicting our third hypothesis, child-
hood cognitive ability was positively related to the overall
CWB index (r = .10, p < .05). In particular, people who were
more intelligent according to childhood measures of cogni-
tive ability were more prone to use equipment at work with-
out permission and to steal money and materials from work
(see Table 1).
PR e d i c t i n g th e co u n t e R PRo d u c t i v e emPl o y e e i n a ch i l d -t o -adul t PR o s P e c tiv e st u d y 1431
Table 2 shows the correlations between CWBs and person-
ality traits assessed at age 18. Consistent with our hypotheses,
traits from the domain of Constraint and Negative Emotional-
ity were statistically signicant predictors of CWBs at age 26.
The superfactor scale of Negative Emotionality was positively
related to CWBs (r = .21, p < .05), as were the subscales of Ag-
gression (r = .24, p < .05), Stress Reaction (r = .08, p < .05), and
Alienation (r = .15, p < .05). Conversely, the superordinate fac-
tor scale of Constraint was negatively related to CWBs (r =
.27, p < .05) as well as each constituent element of Constraint:
Self-Control (r = –.23, p < .05), Harm Avoidance (r = –.21, p <
.05), and Traditionalism (r = –.16, p < .05). The Positive Emo-
tionality superfactor scale was unrelated to CWBs, yet specic
subscales from the Positive Emotionality domain were statisti-
cally signicant predictors. Social Closeness was negatively re-
lated to CWBs (r = –.13, p < .05), and Social Potency was posi-
tively related to CWBs (r = .14, p < .05). Thus, adolescents who
were less controlled, more aggressive and alienated, more as-
sertive, and less sociable were more likely to commit counter-
productive behaviors as young adults.
Table 2 also shows that the predictive pattern of correla-
tions for the personality traits was quite consistent at the item
level of analysis. The personality traits of Self-Control and Ag-
gression had the most consistent pattern of correlates across
the CWB items. Self-Control correlated signicantly with 9
of the 11 CWB items, and Aggression correlated signicantly
with 8 of the 11 CWB items. Individuals who were less con-
trolled and more aggressive were more likely to commit the
majority of the counterproductive behaviors, with the excep-
tion of stealing money from work, which was unrelated to all
personality predictors. Of interest, Social Potency had a pat-
tern of positive correlates focused mostly on interpersonal
problems, such as having a conict with a supervisor or hav-
ing a ght at work.
Table 3 shows the correlations between work conditions
at age 26 and the propensity to commit CWBs at age 26. Con-
sistent with our hypotheses, employees who were more satis-
ed were less likely to participate in CWBs (r = –.20, p < .05).
Similarly, individuals afforded more autonomy in their work
were also less likely to commit CWBs (r = –.11, p < .05). In con-
trast, both occupational attainment and resource power had
some correlations with the items, but not the overall index of
CWB. The lack of overall relationship for occupational attain-
ment with CWBs was the result of a conicting pattern of pos-
itive and negative relationships with CWB items. People in
higher status jobs were more likely to use equipment without
permission (r = .10) and to steal from work (r = .07) but were
less likely to perform acts they felt would result in their be-
ing red (r = –.10) or to damage equipment (r = –.10). Resource
power had only two statistically signicant correlates. People
with more resource power were more likely to have ghts at
work (r = .12) and to come to work under the inuence of sub-
stances (r = .08). The correlation with ghting at work was rea-
sonable because these individuals typically supervised more
people and thus had a higher probability of conict. Work au-
tonomy was much more strongly related to factors such as not
being late to work (r = –.16) or pretending to be sick (r = –.10).
Thus, people with more autonomy were less likely to malinger
and avoid work.
Multiple Regression Analyses
It is possible that many of these variables are not unique
predictors of CWBs. For example, adolescent conduct disor-
der may account for some of the personality effects, or con-
versely, personality traits may mediate the relationship of ad-
olescent conduct disorder and adult CWBs. Also, we did not
control for other factors that may contribute to CWBs, such as
Table 1. Percentages of Participants Committing Counterproductive Work Behaviors (CWBs) and Correlations With Background Factors
Conduct Adolescent criminal Average cognitive
Counterproductive act % disorder convictions ability, ages 7–13
How many times in the past year …
have you been late to work? 51 .04 –.02 .03
have you pretended you were sick or injured, or gave another false excuse
so you could get time off work? 34 .04 .01 –.01
have you used things at work without permission (like using the telephone,
Xerox machine, computer, tools or a company car without permission)? 17 .04 .08 .15
have you had a conict with your boss or supervisor (like refusing to carry
out an assignment, told them a lie, or some other trouble with the boss)? 29 .11 .01 .01
have you lost your temper, had a ght, or got into an argument with
someone at work? 25 .15 .06 .04
have you done your job in a way that could cause you to lose it (like taking
shortcuts, missing deadlines, breaking safety rules)? 6 .14 .06 –.02
did you steal money from the place where you worked? 1 .03 –.04 .09
have you reported working hours or days (so that you could get paid) that
you really did not work? 4 .00 .01 .03
did you steal things from work, such as ofce supplies, tools, or merchandise? 17 .02 .07 .11
did you purposely damage or destroy property, equipment, tools or
merchandise where you work? 1 .12 –.01 –.03
have you been under the inuence of alcohol or drugs while you were at work? 19 .20 .08 .04
Total CWBs .11 –.01 .10
Ns = 930 for criminal activity and 922 for conduct disorder. All correlations shown in boldface are statistically signicant at p < .05.
1432 Ro b e R t s e t al . i n Jo u r nal o f ap p l i ed ps y c h o l o g y 92 (2007)
Table 2. Correlations Between Items From the Counterproductive Work Behavior (CWB) Index and Personality Traits at Age 18
Counterproductive Negative Aggres- Stress Aliena- Con- Tradition- Self- Harm Positive Social Well- Achieve- Social
act Emotionality sion Reaction tion straint alism Control Avoidance Emotionality Closeness Being ment Potency
Late to work .06 .09 .02 .02 .13 –.04 .15 .09 –.02 .00 –.03 .08 .05
Pretended to be sick .07 .08 .07 –.01 .08 .07 .09 –.02 –.05 .01 –.04 .10 .01
Used equipment
without permission .07 .11 .03 .02 .16 –.04 .13 .16 –.02 .12 –.02 –.03 .10
Conict with boss .14 .15 .03 .14 .13 .08 .08 .13 .04 –.06 .00 .04 .10
Had a ght or argument
at work .16 .14 .06 .17 .14 .08 .12 .09 .04 .08 –.02 .06 .11
Did something that
could get me sacked .17 .18 .03 .18 .17 .09 .14 .14 –.04 .11 –.02 –.04 .06
Stole money from work –.01 –.02 .01 –.02 –.01 –.06 –.00 .02 .04 –.01 .01 .02 .06
Defrauded time sheet .04 .03 .00 .06 –.07 –.07 .07 –.03 .10 .10 .08 –.03 –.03
Stole things from work .06 .04 .07 .01 .10 –.03 .07 .11 –.02 –.03 –.06 –.05 .07
Damaged work property .07 .07 .05 .04 .09 .10 –.02 .08 –.02 .10 .00 .04 .00
Was drunk or on drugs
while at work .15 .23 .02 .11 .23 .20 .18 .15 –.06 .14 –.06 –.06 .08
Total CWBs .21 .24 .08 .15 .27 .16 .23 .21 –.03 .13 –.06 –.06 .14
N = 877. All correlations shown in boldface are statistically signicant at p < .05.
PR e d i c t i n g th e co u n t e R PRo d u c t i v e emPl o y e e i n a ch i l d -t o -adul t PR o s P e c tiv e st u d y 1433
work satisfaction and autonomy. Thus, we conducted a hierar-
chical multiple regression to test the independent effects of the
background factors, personality traits, and work conditions
on CWBs. Table 4 shows the correlations among the variables
used in the regression models.
On the rst step of the regression, we entered adolescent
conduct disorder, criminal convictions that occurred prior to
entering the workforce, and childhood cognitive ability. As
can be seen in Table 5, both adolescent conduct disorder and
childhood cognitive ability were positive predictors of CWBs.
Smarter individuals diagnosed with conduct disorder in ad-
olescence were more likely to participate in CWBs in young
adulthood. Criminal activities recorded until age 18 were un-
related to CWBs at age 26. On the second step of the regres-
sion model, we entered four personality variables. We used
the lower order scales exclusively because of the complex re-
lationship of the Positive Emotionality scales with CWBs (i.e.,
the negative relationship of Social Closeness and the positive
relationship of Social Potency cancelled each other out at the
superfactor level). We also selected only one scale from each
domain of the MPQ, with the exception of Positive Emotional-
ity, in order to avoid problems with multicollinearity. We se-
lected the Self-Control scale from the Constraint domain, the
Aggression scale from the Negative Emotionality domain, and
the Social Closeness and Social Potency scales from the Pos-
itive Emotionality domain, based on their correlations with
CWBs.
When we controlled for conduct disorder, prior criminal
convictions, and childhood cognitive ability, three of the four
personality scales remained statistically signicant predic-
tors of CWBs (Self-Control, Social Closeness, and Aggression).
The inclusion of the personality variables resulted in a marked
Table 3. Correlations Between Items From the Counterproductive
Work Behavior (CWB) Index and Work Conditions at Age 26
Counterproductive Occupational Work Resource Work
act level satisfaction power autonomy
Late to work –.03 .09 .03 .16
Pretended to be sick –.03 .16 –.03 .10
Used equipment without
permission .10 –.06 –.03 .08
Conict with boss .08 .20 .06 –.05
Had a ght or argument
at work –.06 .12 .12 .04
Did something that
could get me sacked .10 .08 .00 –.06
Stole money from work –.01 –.03 .00 .00
Defrauded time sheet .02 –.06 –.01 .07
Stole things from work .07 –.05 –.01 .09
Damaged work property .10 .02 .00 –.04
Was drunk or on drugs
while at work .07 .07 .08 .07
Total CWBs .06 .20 –.04 .11
N = 927. All correlations shown in boldface are statistically signicant
at p < .05.
Table 5. Situational and Individual-Differences Variables Predicting
Counterproductive Work Behaviors
Variable Step Step Step
1 2 3
Childhood factors
Adolescent conduct disorder .19 .09 .08
Criminal convictions up to age 18 –.03 .08 –.06
Average IQ, ages 7–13 .10 .07 .08
Age 18 personality
Age 18 Self-Control .17 .16
Age 18 Aggression .14 .12
Age 18 Social Potency .05 .07
Age 18 Social Closeness .09 .08
Job characteristics at age 26
Occupational attainment .01
Work satisfaction .19
Resource power .10
Work autonomy .10
R .20 .35 .41
Note. N = 838. Standardized beta weights and Rs shown in boldface
are statistically signicant at p < .05.
Table 4. Intercorrelations Between Regression Variables
Variable 1 2 3 4 5 6 7 8 9 10 11 12
1. IQ .96
2. Adolescent conduct disorder .12 .83
3. Criminal activity up to age 18 –.06 .24
4. Age 18 Self-Control –.05 .21 .16 .79
5. Age 18 Aggression .07 .36 .19 .37 .78
6. Age 18 Social Potency .31 .09 .01 .15 .25 .76
7. Age 18 Social Closeness –.01 –.17 .12 .13 .25 .08 .75
8. Occupational level .44 .25 .15 .15 .21 .20 .10 .75
9. Work satisfaction .04 .09 –.02 .09 .12 .04 .12 .20 .63
10. Resource power .06 .08 .02 .01 .02 .11 –.01 .11 .13 .67
11. Work autonomy .10 .04 .11 .02 .00 .10 –.02 .15 .14 .32 .69
12. CWBs .08 .17 .01 .25 .25 .14 .15 –.07 .23 .06 –.09 .60
N = 838. All correlations shown in boldface are statistically signicant at p > .05. Internal consistency reliabilities are shown on the diagonal.
CWB = counterproductive work behavior.
1434 Ro b e R t s e t al . i n Jo u r nal o f ap p l i ed ps y c h o l o g y 92 (2007)
drop in the magnitude of the effect of adolescent conduct dis-
order on CWBs. In fact, according to the Sobel’s z test (Kenny,
Kashy, & Bolger, 1998; Sobel, 1982) the Social Closeness, Ag-
gression, and Self-Control scales were signicant mediators of
the relationship between conduct disorder and CWBs. Of the
three scales, Aggression was the strongest mediator (z = 5.2, p
< .05), which ts well the emphasis on “hostility” in the prior
literature on CWBs. (It should be noted that the MPQ Aggres-
sion scale does not share item content overlap with conduct
disorder; the conduct disorder scale assesses actual behaviors
committed, such as ghting or stealing, whereas the Aggres-
sion scale assesses attitudes and interests, such as enjoyment
of violent lms.)
In the third step of the regression equation, we entered the
four work variables. Consistent with much previous research,
work satisfaction was negatively related to CWBs = –.19, p
< .05), indicating that workers who are dissatised with their
work are more likely to participate in counterproductive behav-
iors. Likewise, work autonomy was negatively related to CWBs
(β = –.10, p < .05). In contrast, resource power was positively re-
lated to CWBs = .10, p < .05), indicating that employees with
more control over resources were more likely to commit CWBs.
The inclusion of the work variables did not substantially affect
any of the previous predictors. Of interest, the magnitude of
the effects of personality traits was similar to the magnitude of
the work context variables, despite the fact that the work con-
text factors were assessed concurrently with the CWBs, whereas
personality had been assessed 8 years earlier.
Discussion
The present study tested the relations between a battery
of background factors and CWBs in an ongoing longitudi-
nal study of young adults. In contrast to much of the previ-
ous research, the design of the present study was unique, in
that the background information was gathered concurrently
as the participants’ lives unfolded, rather than retrospectively.
The results showed that individuals diagnosed with child-
hood and adolescent conduct disorder were more prone to
commit CWBs in young adulthood and that these associations
were partially mediated by personality traits measured at age
18. Job conditions and evaluations had independent effects on
CWBs above and beyond personality traits. We discuss each of
these patterns and their implications in turn.
The present study conrmed that a suite of background
factors can be used to predict CWBs in young adulthood. Ado-
lescent conduct disorder proved to be the best predictor in the
rst stage of the regression model. This pattern of predictors
seems reasonable if CWBs are viewed from a theoretical stand-
point as work-context-specic features of a wider tendency to-
ward antisocial behavior across contexts. Conduct disorder ap-
pearing in childhood and adolescence is the earliest emerging
feature of this wider antisocial tendency in the life course, and
it is known to predict a variety of antisocial activities in adult-
hood (Moftt et al., 2001).
Of most interest was the fact that childhood conduct disor-
der was associated with CWBs measured in young adulthood.
This indicates that there are important continuities from child-
hood to adulthood in people’s lives that can have serious ram-
ications for their well-being and occupational success. This
is not to say that childhood character is fate, as the effect of
childhood conduct disorder was small in magnitude. More-
over, early childhood conduct disorder most likely results in
both personality and environmental issues that form a cocktail
of factors that may independently contribute to poor outcomes
in adulthood. For example, conduct disorder may contribute
to an impulsive adult personality but might also contribute
to poor educational outcomes, which may have independent,
negative effects on work outcomes (Moftt et al., 2001). The
continuity in psychological factors from childhood to adult-
hood also indicates that early interventions to ameliorate con-
duct disorder problems would have signicant, positive life
course consequences for individuals.
In contrast to the effect of adolescent conduct disorder, ad-
olescent criminal convictions did not predict CWBs in young
adulthood. Given the similarity between the two constructs
of conviction and CWBs, one would expect that past violent
convictions, for example, would be related to aggressive be-
haviors at work. In fact, criminal conviction actually had small
negative relationships with ghting or stealing at work. One
potential explanation is that being caught and convicted for
criminal activities acted as a preventative buffer against future
CWBs; these individuals learned their lesson. Another inter-
pretation is that the fear of losing one’s job is greater for those
with a criminal conviction background, which leads these peo-
ple to be more careful on the job to avoid getting red. A third
possibility arises from the established fact that conviction is an
imperfect indicator of true illicit behaviors, and most crimes
go undetected. This could imply that individuals who engage
in illicit behavior but are clever enough to evade detection and
conviction are also able to negotiate entry into jobs, whereas
in contrast, unfortunate individuals who are frequently con-
victed for their crimes experience less time employed, and as a
result have fewer opportunities to engage in CWBs. This nd-
ing is important, as there is little or no predictive validity data
on the relationship between past criminal behavior and coun-
terproductive behaviors at work. Our ndings point to the
necessity of examining these patterns more closely, as many
organizations screen potential employees for past criminal be-
havior under the assumption that they are eliminating poten-
tial problem employees, which may not be the case.
1
The pres-
ent study constituted a rigorous test of the ability of conviction
records to predict CWBs when both data sources are uncon-
taminated by each other, because conviction data were sys-
tematically obtained by a background-check search of police
computer records whereas CWBs were measured separately in
interviews carried out under conditions of strict condential-
ity that fostered frank reporting.
Unlike past criminal convictions, childhood cognitive abil-
ity had a positive relationship with CWBs. Specically, partic-
ipants with higher cognitive ability scores in childhood were
more likely to use equipment for personal uses and steal from
the organization. One interpretation of these ndings is that
high-IQ individuals have a stronger sense of entitlement,
and therefore feel justied in using equipment and taking re-
sources from an organization. The nding was surprising
given the previous research showing a negative relationship
between cognitive ability and CWBs (Dilchert et al., 2007;
Duehr et al., 2006). One reason for the discrepancy may lie in
1
It should be noted that organizations cannot use adolescent criminal
records when screening potential employees and that investigations
of adult criminal activity may show a different pattern of relations
to CWBs.
PR e d i c t i n g th e co u n t e R PRo d u c t i v e emPl o y e e i n a ch i l d -t o -adul t PR o s P e c tiv e st u d y 1435
the samples typically studied in organizational studies. Many
organizational studies use samples of relatively high-function-
ing individuals who are either college students or members of
a single organization. Cognitive ability within a single organi-
zation may play a different role than cognitive ability across
organizations, like those examined in the Dunedin study. For
example, within organizations, it may be clear to employees
that theft is a problem, and therefore employees with higher
cognitive ability may avoid stealing materials from work more
than employees with lower cognitive ability. Across organiza-
tions, however, other mechanisms and processes may drive a
different relationship between cognitive ability and CWBs. For
example, people with higher cognitive ability may pursue and
get selected into jobs that have more resources to steal. In con-
trast, people with low cognitive ability may end up dispropor-
tionately in jobs that afford fewer resources to steal (e.g., sheep
shearer, slaughterer). Thus, the positive association between
cognitive ability and theft may have occurred because higher
cognitive ability leads to jobs that afford greater opportunity
to steal, relative to low-ability jobs.
Despite controlling for these background factors, we found
that personality traits assessed at age 18 had consistent predic-
tive relations to CWBs assessed at age 26. Moreover, personal-
ity traits mediated the effect of childhood conduct disorder on
CWBs. Consistent with a wide variety of research from clinical
and industrial psychology, the two most important personal-
ity trait domains were from the domains of Agreeableness and
Conscientiousness (Berry et al., 2007). The Aggression and So-
cial Closeness scales were some of the strongest predictors of
total CWBs. Also, all four scales tapping the domain of Con-
straint, which is strongly related to Conscientiousness, pre-
dicted CWBs as well. Thus, people with a strong compulsion
to adhere to norms, control their impulses, and avoid hostil-
ity tend not to commit counterproductive activities at work.
In addition, individuals who scored higher on Social Potency
tended to commit more CWBs, largely because they were in-
volved in more interpersonal conicts. The fact that person-
ality traits mediated the effect of conduct disorder on CWBs
bodes well for the use of personality tests as selection tools.
Personality tests appear to serve as a proxy for background
factors, such as childhood and adolescent conduct disorder,
which most organizations cannot assess reliably or validly.
However, some caution is warranted in making inferences
from our ndings to real personnel recruitment situations, be-
cause individuals seeking a job may be motivated to attempt
to portray their personalities in a positive light, whereas such
motivation probably did not strongly bias the cohort mem-
bers’ self-reports of personality in our research setting.
Individual differences in personality and other back-
ground factors are only part of the story behind CWB. Much
research has pointed to the importance of the work context
and job attitudes as antecedents of counterproductive activ-
ities (e.g., Robinson & O’Leary-Kelly, 1998). Consistent with
this research, we found that people in more satisfying, au-
tonomous jobs engaged in fewer counterproductive behav-
iors, whereas people with more resource power tended to
engage in more—although it should be noted that the latter
was most likely the result of experiencing more interpersonal
conict, which would be a natural consequence of supervis-
ing more people.
Despite the multiple methods and relatively long-term
prospective design of the study, the present study also has
important limitations. Most conspicuously, the ndings may
not be directly relevant to selection, as the sample studied
was quite heterogeneous in the number and type of jobs oc-
cupied. As we noted in our discussion of the counterintui-
tive ndings for childhood cognitive ability, this heterogene-
ity may result in a qualitatively different pattern of results
compared with that of more prototypical validation stud-
ies. On the other hand, the fact that personality traits and job
conditions found to be important in narrower samples rep-
licated in the Dunedin study bolsters the generalizability of
these ndings.
A second limitation was that we focused only on young
adults and therefore have little or no information concerning
whether these patterns hold for older individuals. It is quite
possible that the patterns will change with age, with differ-
ent personality traits and job conditions becoming more im-
portant for older workers. Third, we used a measure of coun-
terproductive behaviors created specically for this study. In
contrast to existing measures of CWBs, this measure empha-
sized serious transgressions, such as stealing and physical vi-
olence, rather than less extreme attitudinal features of CWBs,
such as withholding effort or work withdrawal. Limiting the
construct to more serious offenses most likely limited our
ability to predict these CWBs, as they occur at a lower base
rate than behaviors reecting a disaffected attitude. Also,
our measure was not designed to differentiate between the
two facets of CWBs—interpersonal and organizational devi-
ance (Robinson & Bennett, 1995)—which though highly cor-
related (Dalal, 2005) can have slightly different correlates
with personality variables (Berry et al., 2007). In terms of our
predictor measures, the internal consistency reliability var-
ied somewhat from measure to measure. This could, within
the present study, undermine the comparability of the mag-
nitude of the validity coefcients. Of course, a true compar-
ison of the magnitude of the relative effect sizes of person-
ality and organizational variables is best made when based
on meta-analytic data. Finally, although the predictors were
gathered using observer, interview, and self-report tech-
niques, the majority of variables were assessed via self-re-
port. Future research should incorporate observer ratings of
personality and job behaviors in order to test whether these
background factors generalize across methods.
Future research could identify and examine potential me-
diators of the relationship between personality and CWBs or
other voluntary work behaviors. Like job performance, spe-
cic types of goals may mediate this relationship (e.g., Bar-
rick, Stewart, & Piotrowski, 2002). For example, people who
are more conscientious may also have more prosocial goals.
In the work context, this would manifest itself as a propen-
sity not to engage in CWBs. Another possibility for future re-
search would be to examine the effect of type of industry or
type of vocation (Holland, 1985) on the predictors of CWBs.
For example, in Holland’s system, jobs are organized into six
broad psychological categories (i.e., Realistic, Investigative,
Artistic, Social, Enterprising, and Conventional). The ante-
cedents to CWBs may be different in Artistic jobs (e.g., artist,
architect) than in Social jobs (e.g., administrative assistant,
nurse). Moreover, these different types of jobs may afford dif-
ferent forms of CWBs, such that certain jobs engender greater
interpersonal rather than organizational deviance. Of course,
a study examining these factors would by necessity have to
be quite large.
1436 Ro b e R t s e t al . i n Jo u r nal o f ap p l i ed ps y c h o l o g y 92 (2007)
In summary, CWBs appear to be predictable from child-
hood and adolescent activities. Furthermore, the links between
childhood and adolescent conduct disorder and CWBs were
accounted for by personality traits associated with Agreeable-
ness and Conscientiousness. Finally, work conditions were
also important, independent predictors of CWBs, alluding to
the fact that an interactionist model of CWBs may be the most
productive representation of the phenomenon.
References
American Psychiatric Association (1994). Diagnostic and statistical man-
ual of mental disorders (4th ed.). Washington, DC: Author.
Barrick, M. R., Stewart, G. L., & Piotrowski, M. (2002). Personality
and job performance: Test of the mediating effects of motivation
among sales representatives. Journal of Applied Psychology, 87,
43–51.
Berry, C. M., Ones, D. S., & Sackett, P. R. (2007). Interpersonal devi-
ance, organizational deviance, and their common correlates. Jour-
nal of Applied Psychology, 92, 410–424.
Colbert, A. E., Mount, M. K., Harter, J. K., Witt, L. A., & Barrick, M.
R. (2004). Interactive effects of personality and perceptions of the
work situation on workplace deviance. Journal of Applied Psy-
chology, 89, 599–609.
Connerley, M. L., Arvey, R. D., & Bernardy, C. J. (2001). Criminal
background checks for prospective and current employees: Cur-
rent practices among municipal agencies. Public Personnel Manage-
ment, 30, 173–183.
Dalal, R. S. (2005). A meta-analysis of the relationship between organi-
zational citizenship behavior and counterproductive work behav-
ior. Journal of Applied Psychology, 90, 1241–1255.
Dilchert, S., Ones, D. S., Davis, R. D., & Rostow, C. D. (2007). Cognitive
ability predicts objectively measured counterproductive work-
place behaviors. Journal of Applied Psychology, 92, 616–627.
Duehr, E. E., Sackett, P. R., & Ones, D. S. (2006). A facet-level examina-
tion of the relationship between integrity and cognitive ability. Unpub-
lished manuscript, University of Minnesota.
Greenberger, E., & O’Neil, R. (1993). Spouse, parent, worker: Role
commitments and role-related experiences in the construction of
adults’ well being. Developmental Psychology, 29, 181–197.
Hambleton, R. K., Swaminathan, H., & Rogers, H. J. (1991). Fundamen-
tals of item response theory. Newbury Park, CA: Sage.
Henry, B., Moftt, T. E., Caspi, A., Langley, J., & Silva, P. A. (1994).
On the “remembrance of things past”: A longitudinal evaluation
of the retrospective method. Psychological Assessment, 6, 92–101.
Holland, J. L. (1985). Making vocational choices: A theory of careers. En-
glewood Cliffs, NJ: Prentice Hall.
Hollinger, R. C., & Clark, J. P. (1983). Deterrence in the workplace: Per-
ceived certainty, perceived severity, and employee theft. Social
Forces, 62, 398–418.
Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social
psychology. In D. T.Gilbert, S. T.Fiske, & G.Lindzey (Eds.), The
handbook of social psychology (pp. 233–265). Boston: McGraw-Hill.
Krueger, R. F., Caspi, A., & Moftt, T. E. (2000). Epidemiological per-
sonology: The unifying role of personality in population-based re-
search on problem behaviors. Journal of Personality, 68, 967–998.
Lau, V. C. S., Au, W. T., & Ho, J. M. C. (2003). A qualitative and quan-
titative review of antecedents of counterproductive behavior in or-
ganizations. Journal of Business and Psychology, 18, 73–99.
Lee, K., & Allen, N. J. (2002). Organizational citizenship behavior and
workplace deviance: The role of affect and cognitions. Journal of
Applied Psychology, 87, 131–142.
Lee, K., Ashton, M. C., & Shin, K. H. (2005). Personality correlates of
workplace anti-social behavior. Applied Psychology: An Interna-
tional Review, 54, 81–98.
LePine, J. A., & Van Dyne, L. (2001). Voice and cooperative behavior as
contrasting forms of contextual performance: Evidence of differen-
tial relationships with Big Five personality characteristics and cog-
nitive ability. Journal of Applied Psychology, 86, 326–336.
Lord, F. M. (1980). Applications of item response theory to practical testing
problems. Hillsdale, NJ: Erlbaum.
Meyer, G. J., Finn, S. E., Eyde, L. D., Kay, G. G., Moreland, K. L., Dies,
R. R., et al. (2001). Psychological testing and psychological assess-
ment. American Psychologist, 56, 128–165.
Mislevy, R. J., & Bock, R. D. (1990). BILOG 3 Item Analysis and Test
Scoring With Binary Logistic Models[Computer program].
Mooresville, IN: Scientic Software.
Moftt, T. E., Caspi, A., Rutter, M., & Silva, P. A. (2001). Sex differences
in antisocial behavior: Conduct disorder, delinquency, and violence in
the Dunedin longitudinal study. Cambridge, England: Cambridge
University Press.
Ones, D. S. (2002). Introduction to the special issue on counterproduc-
tive behaviors at work. International Journal of Selection and Assess-
ment, 10, 1–4.
Ones, D. S., & Viswesvaran, C. (2001). Personality at work: Criterion-
focused occupational personality scales used in personnel selec-
tion. In B. W.Roberts & R.Hogan (Eds.), Personality psychology in
the workplace (pp. 63–92). Washington, DC: American Psycholog-
ical Association.
Ones, D. S., & Viswesvaran, C. (2003). Personality and counterproduc-
tive behaviors. In A.Sagie, S.Stashevsky, & M.Koslowsky (Eds.),
Misbehavior and dysfunctional attitudes in organizations (pp. 211–
249). Houndmills, Hampshire, England: Palgrave Macmillan.
Ones, D. S., Viswesvaran, C., & Schmidt, F. L. (1993). Comprehensive
meta-analysis of integrity test validities: Findings and implications
for personnel selection and theories of job performance. Journal of
Applied Psychology, 78, 679–703.
Roberts, B. W., Caspi, A., & Moftt, T. E. (2003). Work experiences and
personality development in young adulthood. Journal of Personal-
ity and Social Psychology, 84, 582–593.
Roberts, B. W., Chernyshenko, O., Stark, S., & Goldberg, L. (2005). The
structure of Conscientiousness: An empirical investigation based
on seven major personality questionnaires. Personnel Psychology,
58, 103–139.
Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant work-
place behaviors: A multidimensional scaling study. Academy of
Management Journal, 38, 555–572.
Robinson, S. L., & O’Leary-Kelly, A. M. (1998). Monkey see, monkey
do: The inuence of work groups on the antisocial behavior of em-
ployees. Academy of Management, 41, 658–672.
Sackett, P. R., & DeVore, C. J. (2001). Counterproductive behaviors at
work. In N.Anderson, D. S.Ones, H.Sinangil, & C.Viswesvaran
(Eds.), Handbook of industrial, work, and organizational psychology
(Vol. 1). London: Sage.
Salgado, J. (2002). The Big Five personality dimensions and counter-
productive behaviors. International Journal of Selection and Assess-
ment, 10, 117–125.
Silva, P. A., & Stanton, W. (Eds.). (1996). From child to adult: The Dune-
din study. Auckland, New Zealand: Oxford University Press.
Sobel, M. E. (1982). Asymptotic condence intervals for indirect ef-
fects in structural equation models. In S.Leinhardt (Ed.), Sociologi-
cal methodology 1982 (pp. 290–312). Washington, DC: American So-
ciological Association.
Tellegen, A. (1982). Brief manual of the Multidimensional Personality
Questionnaire. Unpublished manuscript. University of Minnesota.
Wechsler, D. (1974). Manual for the Wechsler Intelligence Scale for Chil-
dren—Revised. New York: Psychological Corporation.
... 36 Furthermore, Roberts et al, showed that employees diagnosed with conduct disorder in childhood or adolescence more often reported counterproductive work behaviours, including disciplinary problems, theft, drug and alcohol abuse. 37 These findings demonstrate the difficulties of being and staying at work for those who experience(d) mental health problems and highlight the necessity for this group to receive support from employers and occupational healthcare providers. ...
Article
Full-text available
Objectives Mental health problems (MHPs) during childhood and adolescence are negatively associated with having a paid job in young adulthood. Yet, little is known about how young adults function at work, that is, do they experience difficulties in meeting their job demands given their health state. This longitudinal study aims to examine the impact of MHPs from childhood to young adulthood on young adults’ work functioning (WF). Methods Data were used from 1004 participants in the TRacking Adolescents’ Individual Lives Survey, a Dutch prospective cohort study with 18-year follow-up. MHP trajectories, including 11, 13.5, 16, 19, 22 and 26 age points, were identified using growth mixture models. WF was assessed at age 29 with the Work Role Functioning Questionnaire 2.0 (WRFQ). Regression analyses were conducted to examine the association between MHP trajectories and WF. Results Young adults with high-stable trajectories of internalising and externalising problems reported lower WF (mean WRFQ scores of 70.5 and 70.7, respectively) than those with low-stable trajectories (78.4 and 77.2), that is, they experience difficulties in meeting the work demands for more than one work day per full-time work week. Young adults with moderate-stable or decreasing MHP trajectories reported lower WF scores compared with those with low-stable trajectories. Conclusions Both persistent high and elevated levels of MHPs from childhood to young adulthood are associated with lower WF scores in young adulthood compared with low-level MHPs. Occupational healthcare professionals should support young workers with a history of MHPs to optimise their work functioning.
... Unsurprisingly, highly conscientious individuals follow the rules of society, maintain social decorum, and think before acting (Jackson et al., 2010). Higher conscientiousness is associated with less drug abuse (Bogg & Roberts, 2004), less counterproductive behavior at work (Roberts et al., 2007), better management of relationships that are prone to conflict (Finkel & Campbell, 2001), and greater exhibition of integrity and moral character (Cohen et al., 2014). ...
Article
Sharing misinformation can be catastrophic, especially during times of national importance. Typically studied in political contexts, the sharing of fake news has been positively linked with conservative political ideology. However, such sweeping generalizations run the risk of increasing already rampant political polarization. We offer a more nuanced account by proposing that the sharing of fake news is largely driven by low conscientiousness conservatives. At high levels of conscientiousness there is no difference between liberals and conservatives. We find support for our hypotheses in the contexts of COVID-19, political, and neutral news across eight studies (six preregistered; two conceptual replications) with 4,642 participants and 91,144 unique participant-news observations. A general desire for chaos explains the interactive effect of political ideology and conscientiousness on the sharing of fake news. Furthermore, our findings indicate the inadequacy of fact-checker interventions to deter the spread of fake news. This underscores the challenges associated with tackling fake news, especially during a crisis like COVID-19 where misinformation impairs the ability of governments to curtail the pandemic. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
... This notation is important because inside sales agents endure considerable schedule surveillance, which counters the need for individuals to have greater control over their time. When employees have a say or some level of choice over their work hours, it enhances their feeling and sense of autonomy (Bureau et al., 2018;Roberts et al., 2007). The ability to control one's time is present in this company's attempt to maximize autonomous motivation by allowing its agents to choose their schedule one of two ways. ...
Article
Purpose Previous sales research remains limited to analyzing the influence of sales activities with sales agent tenure. To date, research on this subject has focused on the downstream direct or indirect impact of sales activities to sales performance, failing to consider whether sales activities impact a sales agent’s tenure. This paper aims to assess the effect of sales activities on sales performance and sales agent engagement on sales agent tenure through the lens of autonomous motivation to better understand sales activities as an overall sales process antecedent Design/methodology/approach Through the utilization of secondary sales operational data, this research demonstrates the influence of sales activities on multiple sales agent outcomes, while depicting the importance of sales managers creating an autonomous motivational climate. Findings This research demonstrates the direct relationship between sales activities to job engagement and sales performance. However, sales activities have a negative relationship to sales agent tenure, which require a sales manager to create an autonomous motivation to mediate the relationship between sales activities and sales agent tenure. Practical implications Organizations are provided with sample methodology and analysis to better determine how a culture grounded in autonomous motivation mediates sales activities and can be a catalyst for improving sales agent tenure. Then, provide a better understanding of the effect of actual sales activities on important sales department work outcomes. Originality/value The model is the first to test holistically the influence of sales activities on sales performance, sales agent engagement and tenure jointly by using actual secondary operational data. This study provides a glimpse of the real world balance a sales manager must consider between climate and activities. Plus, this study takes initial steps to study sales agent engagement, an under-researched construct in sales research.
... Regarding the organizational field, conscientious individuals tend to avoid stealing and drinking during the working hours, to be late or absent at work and they are less likely to quit their jobs (Bowling & Burns, 2010;Roberts et al., 2007). ...
Article
Full-text available
This paper tries to determine whether there is an association between conscientiousness and cooperation and whether this relationship is mediated by sociability. Conscientiousness, cooperation and sociability are psychological constructs that have an important role in an individual’s life, therefore it is important to determine in which way the relationship between these variables can improve the quality of life and the expression of adaptive behaviours. Regardless of age, conscientiousness is a trait that can be found in both children and adults and it has many benefits for the quality of life because it has an important role in many life outcomes. The findings of this study shown that although sociability doesn’t have a strong mediating role on the relationship between conscientiousness and cooperation, conscientiousness predicts cooperation.
... In answer to this movement, many states have laws that restrict the use of criminal background checks during hiring procedures (Zeitler & Luisi, 2016). In support of this legislation, research with young adults has shown no correlation between criminal activity and future poor behaviors at work (Roberts et al., 2007). ...
... Unsurprisingly, highly conscientious individuals follow the rules of society, maintain social decorum, and think before acting (Jackson et al., 2010). Higher conscientiousness is associated with less drug abuse (Bogg & Roberts, 2004), less counterproductive behavior at work (Roberts et al., 2007), better management of relationships that are prone to conflict (Finkel & Campbell, 2001), and greater exhibition of integrity and moral character (Cohen et al., 2014). ...
Preprint
Sharing of misinformation can be catastrophic, especially during times of national importance. Typically studied in political context, sharing of fake news has been positively linked with conservative political ideology. However, such sweeping generalizations run the risk of increasing already rampant political polarization. We offer a more nuanced account by proposing that sharing of fake news is largely driven by low conscientious conservatives. At high levels of conscientiousness there is no difference between liberals and conservatives. Using Covid-19 as a backdrop, we find support for our hypotheses across six studies (five pre-registered; one conceptual replication), with 3,195 participants and 73,108 unique participant-news observations. We find desire for chaos as the psychological mechanism driving the effect. Furthermore, fact-checker interventions were inadequate to deter the spread of fake news. This underscores the challenges associated with tackling fake news, especially during a crisis like Covid-19 where misinformation threatens to exacerbate the pandemic even further.
Article
Full-text available
Nepoželjna organizacijska ponašanja (NOP) štetna su ponašanja koja donose milijunske gubitke organizacijama, a ujedno škode psihološkoj dobrobiti i produktivnosti zaposle-nika. Osobine ličnosti su jedna od glavnih determinanti NOP-a, no istraživanja odnosa ličnosti i NOP-a uglavnom su se foku-sirala na svjesne/eksplicitne aspekte ličnosti i tek su u manjem broju razmatrala psihološke mehanizme putem kojih osobine ličnosti djeluju na NOP. Ovaj rad uključuje pregled istraživanja odnosa ličnosti i NOP-a pri čemu, u odnosu na ranije preglede, razmatra i kako nesvjesni/implicitni aspekti ličnosti mogu pomoći u objašnjavanju NOP-a. U radu se fokusiram se na metodu za mjerenje implicitnih aspekata ličnosti koja je poka-zala najveći potencijal za predviđanje radnog ponašanja-Test uvjetovanog rezoniranja. Nakon toga, razmatram psihološke mehanizme putem kojih eksplicitni i implicitni aspekti ličnosti djeluju na NOP, koristeći stavove prema poslu i organizaciji i emocije na radnom mjestu kao varijable koje objašnjavaju odnos osobina ličnosti i NOP-a. Konačno, na temelju pregleda istraživanja odnosa osobina ličnosti, stavova, emocija i NOP-a, razvijam teoretski model odnosa ove četiri varijable, koji olakšava razumijevanje složenih psiholoških mehanizama putem kojih implicitni i eksplicitni aspekti ličnosti dovode do NOP-a. Ključne riječi: ličnost, implicitni aspekti ličnosti, stavovi, emo-cije, nepoželjno organizacijsko ponašanje primenjena psihologija, str. 189-210
Article
Full-text available
Background: Unemployment can diminish physical, psychological and social health. In this context, research shows that people with mental illness have even more difficulties finding occupation. Thus, some countries, such as Sweden, strive after creating job opportunities for this specific group. We investigated the effect of having an occupation on life satisfaction among individuals with mental illness and whether self-reported physical and psychological health mediated the relationship between being (un)employed and life satisfaction. Method: Two-hundred eighty-seven individuals (148 males, 134 females, and 5 missing information) with mental illness, who received support and services from Swedish Municipalities in Blekinge, self-reported occupation, life satisfaction, and physical and psychological health. Results: Participants who reported having an occupation reported also significantly higher levels of life satisfaction, physical health, and psychological health compared to those without occupation. Nevertheless, these differences were rather small (Eta2 < 0.06). Moreover, the indirect effect of having an occupation on life satisfaction through physical and psychological health was significant. Finally, the total indirect effect of physical and psychological health (i.e., psychophysiological health) accounted for 53% of the total effect of having an occupation on life satisfaction. Conclusion: For individuals with mental illness there seems to be an almost equal importance of indirect and direct effects of having an occupation on their levels of life satisfaction. More specifically, while there are differences in life satisfaction within this population in relation to having an occupation, having an occupation leads to the sense of good psychophysiological health, which in turn helps individuals with mental illness to feel satisfied with their lives.
Article
Purpose The purpose of this paper is to explore variation in the responses of human resource practitioners and managers to criminal offenses. Design/methodology/approach This paper considers background checks as a personnel selection test. In the first study, 280 professionals with hiring experience indicate how various criminal offenses, described as having occurred either within the past year or several years ago, would affect their evaluation of an applicant for a call center position. In the second study, a separate sample of 109 practitioners evaluates criminal as well as non-criminal transgressions that might appear on a background report. Findings In Study 1, both the apparent seriousness of an offense and its recency influence modal responses. Even non-violent misdemeanors from several years ago, however, are judged as automatically disqualifying by some participants. Study 2 shows that a practitioner’s attitude toward criminal offenses is distinct from their attitude to related forms of stigma. Results from both studies find associations between demographic variables and general willingness to accept applicants with criminal records. Originality/value This work provides quantitative data on practitioner reactions to several specific criminal offenses for a specific job context. By considering differences among offenses and among gatekeepers, rather than among applicants, it identifies challenges to fair implementation of background checks during the hiring process.
Article
Past criminal and financial transgressions, even when of questionable job relevance, can harm job‐seekers' prospects. In three hiring simulation experiments, manipulations of situational factors hypothesized to alter decision focus moderate the effect of background report information on applicant evaluations. In Study 1, emphasizing in‐role performance as a hiring goal decreases the weight placed on transgressions relative to standard qualifications. Study 2 demonstrates this context effect with a different design and experienced hiring professionals, and finds that delaying background checks heightens a transgression's impact. In Study 3, presenting novel positive information in conjunction with a criminal offense mitigates the delay effect. By adopting a behavioral decision perspective, this research illuminates how situational factors shape the impact of a widespread form of stigma.
Article
Full-text available
This article summarizes evidence and issues associated with psychological assessment. Data from more than 125 meta-analyses on test validity and 800 samples examining multimethod assessment suggest 4 general conclusions: (a) Psychological test validity is strong and compelling, (b) psychological test validity, is comparable to medical test validity, (c) distinct assessment methods provide unique sources of information, and (d) clinicians who rely exclusively on interviews are prone to incomplete understand-ings. Following principles for optimal nomothetic research , the authors suggest that a multimethod assessment battery provides a structured means for skilled clinicians to maximize the validity of individualized assessments. Future investigations should move beyond an examination of test scales to focus more on the role of psychologists who use tests as helpful tools to furnish patients and referral sources with professional consultation.
Chapter
Employees can engage in a wide spectrum of misbehaviour in organizations. Such counterproductivity costs employers billions of dollars annually worldwide (Ones, 2002). The extent of actual, psychological and societal costs to organizations can be better understood when one considers the multitude of different ways employees can misbehave. Geddes and Baron (1997) found that 69 per cent of managers reported having experienced verbal aggression. Wimbush and Dalton (1997) used multiple methods and samples to estimate the base rate for employee theft, and found ‘Depending on the level one ascribes to nontrivial employee theft, .. . [the different techniques of estimating employee theft] . . . converge on theft rates over 50 per cent’ (p. 756). It is estimated that substance abuse costs the United States alone more than $135 billion each year (DeCresce et al., 1990). Harwood, Fountain and Livermore (1992) estimated that in the United States economy $82 billion in lost potential productivity could be attributed to alcohol and drug abuse in 1992 ($67.7 billion and $14.2 billion, respectively).
Article
This cross-level field study, involving 187 employees from 35 groups in 20 organizations, examined how individuals' antisocial behaviors at work are shaped by the antisocial behavior of their coworkers. We found a positive relationship between the level of antisocial behavior exhibited by an individual and that exhibited by his or her coworkers. We also found that a number of factors moderated this relationship. Finally, we found that dissatisfaction with coworkers was higher when individuals engaged in less antisocial behavior than their coworkers.
Book
Ronald K. Hambleton; H. Swaminathan; H. Jane Rogers., The following values have no corresponding Zotero field: Label: B496 ID - 337
Article
This paper presents a qualitative and quantitative review of the antecedents of counterproductive behaviors (CPB). The qualitative review organizes antecedents identified in past research into four broad categories: personal, organizational, work, and contextual factors. The quantitative review includes metaanalyses of 40 published studies with a combined sample size of 42,359. The results indicate that employees who are young or dissatisfied engage in more CPBs. In addition, absenteeism is more prevalent among employees who are young, female, have lower income, have lower job satisfaction, and who perceive a stronger absence norm, or a stronger ability to be on time.