International Journal of Nursing Studies 46 (2009) 699–707

Contents lists available at ScienceDirect

International Journal of Nursing Studies

journal homepage: www.elsevier.com/ijns

Convergence of self-reports and coworker reports of counterproductive

work behavior: A cross-sectional multi-source survey among health
care workers
Jan de Jonge a,*, Maria C.W. Peeters b
a
Eindhoven University of Technology, Department of Industrial Engineering and Innovation Sciences, Human Performance Management Group,
Eindhoven, The Netherlands
b
Utrecht University, Department of Social and Organizational Psychology, Utrecht, The Netherlands

A R T I C L E I N F O A B S T R A C T

Article history: Background: Most studies of counterproductive work behavior (CWB) are criticized for
Received 4 July 2008 overreliance on single-source self-reports. This study attempts to triangulate on behaviors
Received in revised form 16 December 2008 and perceptions of the work environment by linking job incumbent self-report with
Accepted 18 December 2008 coworker report of the job incumbent’s behaviors. Theoretical framework is the Demand-
Induced Strain Compensation (DISC) Model, which proposes in general that specific job
Keywords: resources should match specific job demands to reduce deviant behavioral outcomes such
Job demands
as CWB.
Job resources
Objectives: To test the extent to which job incumbent self-report and coworker report of
Job stress
DISC Model CWB in health care work converge, and the extent to which job incumbent-reported work-
Counterproductive work behavior related antecedents (i.e., job demands and job resources) similarly predict both self-
Residential elderly care reported and coworker-reported behaviors (in line with DISC theory).
Design: A cross-sectional survey with anonymous questionnaires was conducted, using
data from two different sources (self-reports and coworker reports).
Settings: A large organization for residential elderly care in the Northern urban area in The
Netherlands.
Participants: Self-report and coworker questionnaires were distributed to 123 health care
workers, of which 73 people returned the self-report questionnaire (59% response rate). In
addition, 66 out of 123 coworker questionnaires were returned (54% coworker response
rate). In total 54 surveys of job incumbents and coworkers could be matched.
Methods: Next to descriptive statistics, t-test, and correlations, hierarchical regression
analyses were conducted using SPSS 15.0 for Windows.
Results: Correlations and a t-test demonstrated significant convergence between job
incumbent and coworker reports of CWB. Hierarchical regression analyses showed that
both job incumbent and coworker data consistently demonstrated CWB to be related to its
work-related antecedents. Specifically, findings showed that both physical and emotional
job resources moderated the relation between physical job demands and CWB.
Conclusions: The current findings provide stronger evidence that (multi-source measured)
CWB is associated with job demands and job resources than has been provided in the past.
Moreover, the present study implies that DISC theory has the potential of making a profound
contribution to our understanding of counterproductive working behaviors in health care
work. Future longitudinal studies should investigate this kind of relations more intensely.
ß 2008 Elsevier Ltd. All rights reserved.

* Corresponding author at: Department of Industrial Engineering and Innovation Sciences, Eindhoven University of Technology, P.O. Box 513, 5600 MB
Eindhoven, The Netherlands. Tel.: +31 40 2472243; fax: +31 40 2437161.
E-mail address: j.d.jonge@tue.nl (J. de Jonge).

0020-7489/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijnurstu.2008.12.010
700 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707
What is already known about this topic?
 Overreliance of self-report survey methodology in job
stress studies is of major concern. Self-report bias could
be particularly true for CWB, as employees might be
hesitative to provide truly responses.
 Health care workers could engage in deviant work
behavior due to high workload.
 Job resources such as job autonomy and workplace social
support seem to be able to moderate the negative impact
of high workload on deviant work behavior.
What this paper adds
 Job incumbent and coworker reports converged sig-
nificantly on CWB and its association with work-related
antecedents, chipping away at the assumption that self-
report of these types of work perceptions and (deviant)
behaviors necessarily results in systematic bias and
inflated relations.
 Findings draw the attention to the importance of
providing health care workers with specific, physical
and emotional, job resources to diminish the likelihood
of being engaged in deviant work behaviors.
1. Introduction
Nowadays, there is an increasing interest in voluntary
behaviors of which the intent is to harm the health care
organization, health care staff, or even clients. These
harmful behaviors can be labeled as Counterproductive
Work Behavior, or CWB (Fox and Spector, 2006). There are
many different types of CWB that can harm the organiza-
tion, as there are many motives for engaging in such
behavior. Examples of CWB are production deviance, fraud,
sabotage, theft, client or staff abuse, and withdrawal. Based
upon Robinson and Bennett (1995), we can classify CWB
broadly as those harmful behaviors targeting organiza-
tions, persons, or both. This classification is based on social
psychological theories that consider these various CWBs to
be manifestations of the same underlying motive such as
revenge or aggression (Spector et al., 2006). Furthermore,
earlier studies have found relations between CWB and
work-related and person-related antecedents. For
instance, CWB was linked to conflicts at work, workload
and lack of autonomy as well as to boredom and being
upset (e.g., Marcus and Schuler, 2004; Penney and Spector,
2005; Spector et al., 2006).
In this article the construct of CWB will be exclusively
linked to work-related antecedents. It is well-known from
the job stress literature that highly demanding aspects of
the work environment (e.g., workload) may be seen as
threats to employee well-being and health. Aversive
outcomes of job stress may be psychological (e.g., job
dissatisfaction), physical (e.g., somatic symptoms), or
behavioral (e.g., CWB). Much of the work on CWB has its
roots in the study of aggression, crime and injustice.
Spector and Fox (2002) have linked CWB to negative
emotions such as frustration, anger and anxiety in
response to work-related events. Employees may appraise
these events as stressors which induce negative emotional
reactions. This process continues with coping mechanisms
to tackle either the stressors or the negative emotions. This
may motivate behavior that will reduce negative feelings
and enhance positive feelings. One work-related outcome
of this process may be CWB (Fox et al., 2007).
Theoretical models such as Karasek (1979) job strain
model have been developed to understand the mechan-
isms of job stress, well-being and health. However, to our
knowledge, none of them have been directly linked to
CWB. In the present study, we will use a recently
developed job stress model to investigate the relation
between work-related antecedents and CWB. This model,
the so-called Demand-Induced Strain Compensation
(DISC) Model (de Jonge and Dormann, 2003), is particularly
designed for human service occupations such as health
care workers, and generally proposes that aversive out-
comes can be explained by two distinct work-related
antecedents: job demands and job resources. Job demands
refer to the degree to which the work environment
contains stimuli that require different kinds of sustained
physiological and/or psychological effort (Jones and
Fletcher, 1996). Job resources can be broadly conceptua-
lized as different kinds of energetic reservoirs that are
tapped when the individual has to cope with job demands
(Hobfoll, 1989, 2002). More specifically, the DISC Model
assumes that job demands and job resources are multi-
dimensional factors comprised of cognitive, emotional,
and physical components (cf. Hockey, 2000).
Furthermore, the DISC Model predicts that high job
demands may have adverse effects on health and well-
being that can best be counteracted through the avail-
ability and activation of functional corresponding kinds of
job resources. For example, when emotional problems
with patients arise (e.g., insolent patients), emotionally
supportive colleagues as a job resource is likely to be
helpful. If supportive colleagues are unavailable, other job
resources can be useful to some extent, for instance control
at work to handle a particular problematic patient. de
Jonge and Dormann (2006) propose that job demands are
firstly dealt with using easily available matching job
resources. If such matching job resources are not available
or when they are depleted (cf. Hobfoll, 2002), employees
will search for other job resources and will even use job
resources that do not correspond to the kind of job
demands. As a consequence, de Jonge and Dormann (2006)
state that matching job resources are most often powerful
in combating particular job demands, followed by non-
matching job resources.
With some notable exceptions (e.g., Skarlicki and
Folger, 1997; Penney and Spector, 2005; Fox et al.,
2007), empirical studies of CWB have relied primarily
upon single-source self-reports1. Its use is understandable
as it is a quite natural way of asking employees about their
perceptions, attitudes, behaviors and feelings. However,
the overreliance on self-report survey methodology in job
stress studies is of concern (cf. Edwards, 2008). Self-reports
1
We conducted a CWB literature search in PsycINFO using the
following key words: counterproductive work(ing) behavior, counter-
productivity, and workplace deviance.
 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707
could be biased by a number of factors such as item
ambiguity, social desirability, different interpretations of
response alternatives, and the mood of the employee. In
addition, attempts to demonstrate substantive relations
among self-reported measures raise concerns such as
shared biases that distort correlations among measures
(see Spector, 2006).
Self-report bias could be particularly true for CWB, as
employees might be hesitative to provide truly responses.
Specifically, job incumbents tend to under-report ques-
tions about sensitive topics such as deviant behavior,
because that put them at risk of penalties or even job loss
(Penney and Spector, 2005). However, Fox and Spector
(1999) stated that self-reports might be the most viable
ones, given the ethical difficulties in assessing CWB, the
privileged nature of employees’ knowledge of their own
covert behaviors, and the emphasis on perceptions rather
than on objective conditions in the work environment.
Notwithstanding, more objective (i.e., non-incumbent)
measures are badly needed to increase our knowledge and
understanding of CWB.
1.1. The present study
The present study addresses the goal of going beyond
single-source self-report of counterproductive work
behavior and its antecedents. As expressed by Fox
et al. (2007), only the job incumbent is fully aware of
the deviant acts s/he actually does. Other data sources
such as peers, clients, supervisors, or management
records, are privy to only overt behaviors or the results
of behaviors. Crucial point here is that no single-source
of data is free of either contamination or deficiency (Fox
et al., 2007).
This study tries to link job incumbent self-report to
coworker report of the incumbent’s behaviors (i.e.,
triangulation on behaviors and perceptions of the work
environment). In addition, it attempts to assess the
similarity of incumbent-reported antecedents as predic-
tors of both self-report and coworker report of CWB. It can
be expected that perceptions of an individual’s behaviors
from two distinct viewpoints will demonstrate some
convergence. Convergence between self-report and cow-
orker report would likely depend on the other’s familiarity
with the incumbent’s behaviors. Overt behavior (e.g.,
verbal aggression or absence) would be more likely to be
noticed by a coworker than would covert behaviors such as
theft or pretending to work hard. Since the job incumbents
and coworkers may be focusing on variably overlapping
but non-identical behaviors, we cannot expect total
convergence.
The following two hypotheses address the extent to
which job incumbent self-report and coworker report of
CWB converge, and the extent to which job incumbent-
reported work-related antecedents similarly predict both
self-reported and coworker reported behaviors (in line
with our theoretical assumptions).
Hypothesis 1. Job incumbent’s self-reported CWB will
converge (i.e., will be positively associated) with coworker
reports of CWB.
701
Hypothesis 2a. Job resources moderate the relation
between job demands and CWB. There is a weaker associa-
tion between job demands and CWB for employees with
high job resources than for employees with low job
resources.
Hypothesis 2b. The moderating effect is found more often
for matching job resources than for non-matching job
resources.
Hypothesis 2c. Job incumbent’s reports of job demands
and job resources will be similarly associated with both job
incumbent’s self-reported CWB and coworker reports of
CWB.
2. Method
2.1. Procedure and participants
A cross-sectional survey study was conducted among
123 health care workers (i.e., head nurses, nurses, nurses’
aides, kitchen and cleaning staff) employed in a large
organization for residential elderly care, of which 73
people returned the questionnaire (59% response rate).
Employees were well-informed about the survey, and
could participate on a voluntary basis. In each case,
employees were given two questionnaires, a self-report
form and a coworker form, which could be returned in
sealed envelopes. They got instructions to label both forms
with a matching secret code, and to hand over the
coworker form to a peer familiar with the employee’s
work situation and behavior. In all, 66 out of 123 coworker
questionnaires were returned (54% coworker response
rate). At the end we had 54 matching surveys of job
incumbents and coworkers.
A breakdown of the demographics showed that nearly
all employees were female (97.1%). The mean age was
42.5 years (SD = 9.1), and ranged from 21 to 62 years. The
mean organizational tenure was 5.8 years (SD = 6.8),
whereas 16.6% worked full-time (i.e., 32 h per week).
Finally, 51.5% of the employees worked regular hours and
48.5% worked variable hours. These demographics did
not substantially differ from those of larger survey
studies in the same organization, indicating its repre-
sentativeness of the entire organization (cf. de Jonge
et al., 2008b).
2.2. Measures
Variables included in the present study are self-report
cognitive, emotional and physical job demands and job
resources on the one hand, and counterproductive work
behavior on the other. The coworker survey included the
measure of counterproductive work behavior only. Table 1
shows the psychometric properties of these measures as
well as their means, standard deviations, and zero-order
correlations.
Cognitive, emotional and physical job demands and job
resources were measured using a well-validated version of
the DISC Questionnaire (DISQ 1.1), which was particularly
developed for testing this theoretical model (e.g., van den
Tooren and de Jonge, 2008; van de Ven et al., 2008).
702 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707
Cognitive job demands primarily impinge on brain pro-
cesses involved in information processing, e.g., ‘‘Employee
X will need to display high levels of concentration and
precision at work’’. Emotional job demands can be defined
as the effort needed to deal with job inherent emotions
and/or organizationally desired emotions during inter-
personal transactions, e.g., ‘‘Employee X will have to
display emotions (e.g., towards clients, colleagues or
supervisors) that are inconsistent with his/her current
feelings’’. Physical job demands refer to static and dynamic
physical exertion at work, e.g., ‘‘Employee X will have to lift
or move heavy persons or objects (more than 10 kg)’’.
Cognitive job resources refer to the opportunity to
determine a variety of task aspects and to use problem
solving skills, e.g., ‘‘Employee X would have the opportu-
nity to take a break when tasks require a lot of
concentration’’. Emotional job resources refer to emotional
support from colleagues or supervisors, e.g., ‘‘Other people
(e.g., clients, colleagues or supervisors) would be a
listening ear for employee X when he/she has faced a
threatening situation’’. Finally, physical job resources refer
to instrumental support from colleagues and supervisors,
or ergonomic aids at work, e.g., ‘‘Employee X would receive
help from others (e.g., clients, colleagues or supervisors) in
lifting or moving heavy persons or objects’’. All but one
scale consists of five items (except for emotional demands:
six items) that can be scored on a 5-point frequency scale,
ranging from 1 (never or very rarely) to 5 (very often or
always). Cronbach’s alphas in this study vary from .66 to
.89 (see Table 1).
2.2.1. Counterproductive work behaviors
Ten items modified by Kelloway et al. (2002) from the
Robinson and Bennett (1995) list of deviant workplace
behaviors were used to measure interpersonal and
organizational counterproductive behaviors. These ten
CWBs were considered to be manifestations of the same
underlying motive, and were filled out by both the job
incumbents and their coworkers. Respondents were asked
to report how often they had engaged in each of the ten
listed behaviors in the recent past, with a 5-point
frequency scale ranging from 1 (never) to 5 (very often).
For instance, ‘intentionally worked slow’, ‘stayed out of
sight to avoid work’, ‘taken company equipment or
merchandise’, ‘blamed your coworkers for your mistakes’,
and ‘gossiped about your supervisor’. The scale’s uni-
dimensionality was empirically justified by principal
factor analysis. This factor explained 55% of the variance
Table 1
Descriptive statistics and Pearson correlations of the measures (Cronbach’s Alphas on diagonal) (N = 54).
Measure Items M SD
(1) Cognitive demands 5 3.32 .73
(2) Emotional demands 6 2.47 .62
(3) Physical demands 5 2.76 .91
(4) Cognitive resources 5 3.52 .61
(5) Emotional resources 5 3.85 .60
(6) Physical resources 5 3.75 .64
(7) CWB (self-reports) 10 1.39 .35
(8) CWB (coworkers) 10 1.29 .37
*
p < .05 (two-tailed).
**
p < .01 (two-tailed).
in the job incumbent’s group, and 69% of the variance in the
coworker group. Cronbach’s alphas were .81 for self-report
CWB and .84 for coworker report of CWB (Table 1).
2.3. Analytical strategy
SPSS 15.0 for Windows (SPSS, 2006) was used for
analysing the data. Mean scores for CWB of job incumbents
and coworkers were compared using a t-test for indepen-
dent samples. Relations between CWB and its antecedents
were analysed by hierarchical regression analyses. No
significant violations of linear regression assumptions were
detected. In the first step the standardized main terms of
the job characteristics were included (i.e., cognitive,
emotional as well as physical job demands and job
resources). Postulated moderating effects were tested
by adding multiplicative interaction terms (demands 
resources) of standardized job demands and job resources
into the second step of the regression analyses (cf. Aiken
and West, 1991). However, due to the large number of
possible interaction effects in one single analysis we
decided to split the analysis. According to the theoretical
assumptions of the DISC Model used, we split the analyses
into matching and non-matching demand–resource inter-
action testing (cf. de Jonge and Dormann, 2006; van den
Tooren and de Jonge, 2008). Specifically, two hierarchical
regression analyses were conducted for each CWB outcome,
including either three moderating terms that were com-
puted by multiplying corresponding job demands and job
resources (e.g., cognitive demands  cognitive resources),
or six moderating terms that were computed by multi-
plying non-corresponding job demands and job resources
(e.g., cognitive demands  physical resources). Finally, we
should note that our initial analyses included demographic
characteristics such as age and education as a further
control (cf. Penney and Spector, 2005). Since including these
covariates did not affect the pattern of significant findings,
we omitted them from the analyses presented here for
reasons of parsimony (i.e., ratio of cases to predictor
variables). Due to the limited number of male respondents,
controlling for gender was not necessary at all.
3. Results
Descriptive statistics for the measures are presented in
Table 1. Mean scores for CWB of job incumbents and
coworkers were compared, and a t-test showed no mean
differences (t(104) = 1.37, p = n.s.). Prevalence of CWB
1 2 3 4 5 6 7 8
.84
.57** .82
.26 .29 .89
.56** .08 .17 .66
.04 .04 .39* .26 .72
.07 .06 .04 .26 .45** .75
.03 .31* .10 .04 .26 .18 .81
.07 .03 .15 .04 .31* .15 .46** .84
 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707 703

Table 2
Hierarchical regression models of CWB (both self-report and coworker report) with job demands, job resources, and their matching interactions (N = 54).

Outcomes

CWB (self-report) CWB (coworker report)

B (SE) B (SE)

Demands and resources

Cognitive job demands .15 (.08) .09 (.09)
Cognitive job resources .15* (.07) .11 (.07)
Emotional job demands .18** (.06) .01 (.06)
Emotional job resources .20** (.07) .26*** (.07)
Physical job demands .19** (.07) .27*** (.07)
Physical job resources .04 (.06) .00 (.05)

Matching interactions
Cognitive demands  cognitive resources .03 (.06) .01 (.06)
Emotional demands  emotional resources .07 (.06) .07 (.07)
Physical demands  physical resources .17** (.06) .22** (.07)

Full model R2 = .44, F(9,44) = 3.10, R2 = .44, F(9,44) = 2.84,

p = .008, adjusted R2 = .30 p = .014, adjusted R2 = .28

B = unstandardized coefficient; SE = standard error.

*
p < .05 (two-tailed).
**
p < .01 (two-tailed).
***
p < .001 (two-tailed).

seems to be not very high, given scores of M = 1.39 and
M = 1.29, respectively. Convergence (i.e., a significant
positive correlation) was found between self- and cow-
orker ratings of CWB (r = .46, p < .001). An inspection of the
correlations in Table 1 indicates that in all but one cases,
job demands and job resources showed identical associa-
tions with self-report CWB and coworker reports of CWB,
respectively. Specifically, emotional demands were only
positively associated with self-report CWB, and not with
coworker reports of CWB.
Results of the first set of hierarchical regression
analyses with matching demand–resource interactions
are depicted in Table 2. As a result of the a priori
standardization of all variables, unstandardized regression
coefficients are presented accordingly (cf. Aiken and West,
1991; Jaccard et al., 1990). As can be seen, the work-related
antecedents added substantially to the prediction of both
self-report and coworker reports of CWB (adjusted R2s are
.30 and .28, respectively). A significant matching interac-
tion in which physical resources moderated the relation
between physical demands and self-report CWB is
presented in Fig. 1 according to Aiken and West (1991)
graphical method. Values of the predictor variables were
chosen one standard deviation below and above the mean.
Two simple regression lines were then generated by
entering the values in the equation. Finally, a precise test of
slope significance of the respective simple regression lines
was carried out to allow inferences as regards the
significance of an individual slope (cf. Dawson and Richter,
2006). Fig. 1 shows that an increase in physical demands
was related to more CWB when physical resources are low
( 1SD; simple slope test: t = 3.16, p < .01). Physical
demands were not associated with CWB when physical
resources are high (+1SD; simple slope test: t = .28,
p = n.s.). Main effects may be interpreted in terms of the
average effect of a predictor across values of a moderator.
This average effect usually will be a meaningful piece of
information as well (cf. Jaccard et al., 1990). As far as our
main effects are concerned, findings showed that emo-
tional demands were positively associated with self-report
CWB, emotional job resources were negatively associated
with self-report CWB, and physical job demands as well as
cognitive job resources were positively associated with
self-reports of CWB.
With respect to coworker reports of CWB, regression
analyses showed a similar significant matching interaction
effect. As shown in Fig. 2, an increase in physical demands
was associated with more CWB when physical resources

Fig. 1. Interaction between physical job demands and physical job Fig. 2. Interaction between physical job demands and physical job
resources for self-report CWB. Phys. = physical. resources for coworker report of CWB. Phys. = physical.
704 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707

Table 3
Hierarchical regression models of CWB (both self-report and coworker report) with job demands, job resources, and their non-matching interactions
(N = 54).

Outcomes

CWB (self-report) CWB (coworker report)

B (SE) B (SE)

Demands and resources

Cognitive job demands .07 (.09) .05 (.09)
Cognitive job resources .08 (.08) .14 (.08)
Emotional job demands .13* (.06) .06 (.06)
Emotional job resources .16* (.07) .24** (.07)
Physical job demands .10 (.07) .12 (.07)
Physical job resources .03 (.06) .02 (.06)

Non-matching interactions
Cognitive demands  emotional resources .04 (.06) .24** (.09)
Cognitive demands  physical resources .15 (.08) .11 (.08)
Emotional demands  cognitive resources .02 (.08) .01 (.08)
Emotional demands  physical resources .15 (.08) .06 (.09)
Physical demands  cognitive resources .03 (.05) .02 (.05)
Physical demands  emotional resources .14* (.06) .19** (.07)

Full model R2 = .55, F(12,41) = 3.24, R2 = .50, F(12,41) = 2.53,

p = .004, adjusted R2 = .38 p = .019, adjusted R2 = .30

B = unstandardized coefficient; SE = standard error.

*
p < .05 (two-tailed).
**
p < .01 (two-tailed).

are low ( 1SD; simple slope test: t = 4.30, p < .001). Again,
physical demands were not related to CWB when physical
resources are high (+1SD; simple slope test: t = .71,
p = n.s.). Main effect results indicated that emotional job
resources were negatively associated with coworker
reports of CWB, whereas physical job demands were
positively associated with coworker report of CWB.
Table 3 reports the findings of the second set of hie-
rarchical regression analyses with non-matching demand–
resource interactions in the prediction of CWB. First of all,
the work-related antecedents added substantially to the
prediction of self-report CWB as well as coworker report
of CWB (adjusted R2s are .38 and .30, respectively). Second,
Figs. 3 and 4 showed a significant non-matching interaction
in which emotional resources moderated the relation
between physical demands and CWB. This interaction
effect was nearly identical for both self-reports (Fig. 3)
and coworker reports (Fig. 4).
More specifically, both figures showed that an increase
in physical demands was associated with more CWB when
emotional resources are low ( 1SD) The simple slope tests
are t = 2.87 (p < .01) for Fig. 3 and t = 3.70 (p < .001) for
Fig. 4. In contrast, physical demands were not significantly
related to CWB when emotional resources are high (+1SD;
simple slope tests: t = .38 (p = n.s.) and t = .67 (p = n.s.)
for the respective figures). In addition, Table 3 shows a
third non-matching interaction between cognitive
demands and emotional resources in the prediction of
coworker reports of CWB. As depicted in Fig. 5, the first
regression line indicated a trend: an increase in cognitive
demands was related to more CWB when emotional
resources are high (+1SD; simple slope test: t = 1.74,
p < .10). However, cognitive demands were negatively
associated with CWB in case of low emotional resources
( 1SD; simple slope test: t = 2.05, p < .05). Finally, main
effect findings showed that emotional resources were
negatively associated with both self-reports and coworker
reports of CWB. Emotional demands were positively
associated with self-report CWB.

Fig. 4. Interaction between physical job demands and emotional

Fig. 3. Interaction between physical job demands and emotional job job resources for coworker report of CWB. Phys. = physical and
resources for self-report CWB. Phys. = physical and Emot. = emotional. Emot. = emotional.
 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707
Fig. 5. Interaction between cognitive job demands and emotional
job resources for coworker report of CWB. Cogn. = cognitive and
Emot. = emotional.
4. Discussion
In the present study, we investigated a methodological
issue that has emerged in much job stress research, but
particularly in empirical studies of counterproductive
work behavior (CWB). The overreliance of single-source
self-report survey methodology is of concern because of (1)
under-report bias due to sensitive topics like CWB, and (2)
artificially inflated associations due to shared biases across
different measures within a single questionnaire. To
demonstrate the robustness of our findings across different
reporting groups, we collected CWB data from both job
incumbents and their coworkers, along with additional
work-related antecedents from the job incumbents. Our
aims were two-fold: (1) we tested convergence of these
two sets of CWB reports, and (2) we tested convergence in
associations among job incumbent’s work-related ante-
cedents and CWB outcomes from both job incumbent and
coworker reports. For the second aim, we used a recently
developed job stress model as a heuristic framework
to investigate the relation between work-related ante-
cedents and CWB.
In general, the results were supportive of the hypoth-
eses. More specifically, our first hypothesis was confirmed
in the sense that job incumbent and coworker reports
converged significantly on CWB. We did not find any mean
differences or low, non-significant, associations between
both sets of CWB reports. Apparently, people who work
closely together may be good judges of one another’s
behaviors, even deviant ones. Several other studies have
found convergence for different sources of CWB reports as
well (e.g., Penney and Spector, 2005; Fox et al., 2007). For
instance, Penney and Spector (2005) found convergence
(i.e., a significant positive correlation) between self-reports
and coworker report of CWB of r = .24 (p < .01) in their
sample of nearly 300 employees. As mentioned before,
since the job incumbents and coworkers may be focusing
on variably overlapping but non-identical behaviors, we
can never expect total convergence.
As shown in Tables 2 and 3, both job incumbent and
coworker data consistently demonstrated CWB to be related
to job demands, job resources, as well as their matching
interactions (support for Hypothesis 2c). The total amount
of variance explained (ranging from adjusted R2 = .28 to
adjusted R2 = .38) was satisfying, given methodological
705
considerations as well as the multi-causal etiology of
employee health and well-being (cf. Semmer et al., 1996).
More specifically, we detected five moderating effects of job
resources in the relation between job demands and CWB,
and four out of five interaction effects were in the predicted
direction (some support for Hypothesis 2a). Among the
theoretically valid interactions, two interactions were of
matching kind (i.e., physical job demands in combination
with physical job resources), and two interactions were of
non-matching kind (i.e., physical job demands in combina-
tion with emotional job resources). Strictly speaking,
Hypothesis 2b regarding the likelihood of finding match-
ing interaction effects was not confirmed. However, after
dividing the number of theoretically valid moderating
effects by the number of tested moderating terms (100%),
we might conclude that moderating effects of job resources
are more likely in case of match (one out of three: 33.3% for
both self-report and coworker report) than non-match (one
out of six: 16.7% for both types of reports). These findings are
in line with other empirical studies into the DISC Model (e.g.,
de Jonge and Dormann, 2006; de Jonge et al., 2008a,b; van
den Tooren and de Jonge, 2008).
One, albeit non-matching, interaction (i.e., cognitive
demands  emotional resources) did not converge bet-
ween the two data sources. Moreover, this interaction
effect was not according theoretical predictions, too (i.e.,
the sign of the interaction effect was positive rather than
negative). Next to the fact that this happens more often
with non-matching than with matching interactions (cf. de
Jonge et al., 2008a), this discrepant result across data
sources might have several theoretical and methodological
explanations which are not necessarily mutually exclusive.
First, it is possible that job incumbent reports are more
accurate, which accounts for this kind of associations in
case of coworker reports (Frese and Zapf, 1988). Second,
employee’s perceptions of demands and resources may be
affected by cognitive and/or emotional experience. For
instance, individuals who experience frequent negative
emotions might perceive or report more job demands. In
addition, mood congruence theory (e.g., Isen, 1984)
suggests individuals may recall more negative memories
(such as high cognitive workload or incidents of CWB).
Whether this is due to attentional bias, reporting bias, or
accurate reporting of perceptions/memory is not clear and
would be an outstanding topic for further research. Third,
coworkers will likely observe only a subset of behaviors,
making the coworker reports based on more limited
information.
While the current study tested one possible alternative
to single-source self-report methodology, it has several
limitations. A first limitation concerns the cross-sectional
design of the study, which precludes causation. The
present findings should be replicated in studies with
longitudinal designs. Second, the fact that nearly all
participants were women and also were employees of one
organization for residential elderly care in relation to the
relatively small sample size, raises questions about
the generalizability of our results to men and to other
organizational settings. An advantage of the study
population was that a heterogeneous sample of health
care jobs was investigated (i.e., head nurses, nurses,
706 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707
nurses’ aides, kitchen and cleaning staff), increasing
variance in terms of the focal constructs. Third, it could
be that job incumbents tended to choose friends who
would report on the job incumbent’s behavior more
benignly than other coworkers. It is also possible that the
selection of a friend increased job incumbent-coworker
convergence, as they were more likely to have ongoing
discussions of their working conditions and their own
behaviors. In future research, this limitation could be
addressed by obtaining coworker reports from the entire
workgroup, or even supervisor reports and management
records. In addition, random selection of a well-informed
coworker participant is recommended. Finally, as in most
other studies of CWB, we assessed CWB as a single
dimension. Spector et al. (2006) showed evidence for
differential relations with antecedents which suggest the
use of more specific dimensions to assess CWB in future
research such as abuse towards others, sabotage, theft,
and withdrawal.
From a practical point of view, the results suggest that
health care workers who are confronted with high physical
job demands and have low physical job resources are at
risk for CWB. Enhancing specific job resources such as the
availability of hoists and/or instrumental support from
colleagues, enables health care workers to deal with their
high physical job demands, and decreases the likelihood of
being engaged in deviant behaviors such as intentionally
working slow and/or avoiding work. Further, our findings
suggest that employees who are confronted with high
physical demands and have low emotional resources are at
risk for CWB, too. In this case, it would be good to enhance
their emotional resources (e.g., emotional support from
colleagues and supervisors) to make them capable to deal
with high physical demands and, consequently, counteract
CWB. To conclude, work-related interventions on job
demands in health care work should primarily focus on
specific, physical and emotional, job resources to diminish
deviant behaviors.
4.1. Conclusions
Our results have contributed to empirical evidence
chipping away at the assumption that self-report of work
perceptions measured and (deviant) behaviors necessarily
results in systematic bias and inflated relations (cf. Spector,
2006). The present study implies that DISC theory has the
potential of making a profound contribution to our
understanding of counterproductive working behaviors.
The current findings provide stronger evidence that CWB is
associated with job demands and job resources than has
been provided in the past. The objective of multi-source
data is to overcome the limitations of each form of single-
source data. Further empirical research is badly needed to
use additional sources of data together with sophisticated
research methodologies to further triangulate on these
substantial relations between work characteristics and
deviant behaviors at work.
Conflict of interest
None.
Funding
Research funded by Eindhoven University of Technol-
ogy, The Netherlands.
Ethical approval
Approval was given by all employees and their
management. They were well-informed about the survey,
and could participate on voluntary basis.
References
Aiken, L.S., West, S.G., 1991. Multiple Regression: Testing and Interpreting
Interactions. Sage Publications, Newbury Park, CA.
Dawson, J.F., Richter, A.W., 2006. Probing three-way interactions in
moderated multiple regression: Development and application of a
slope difference test. Journal of Applied Psychology 91 (4), 917–
926.
Edwards, J.R., 2008. To prosper, organizational psychology should. . .
overcome methodological barriers to progress. Journal of Organiza-
tional Behavior 29, 469–491.
Fox, S., Spector, P.E., 1999. A model of work frustration-aggression.
Journal of Organizational Behavior 20, 915–931.
Fox, S., Spector, P.E., 2006. The many roles of control in a stressor-emotion
theory of counterproductive work behavior. In: Perrewe, P.L., Gan-
ster, D.C. (Eds.), Research in Occupational Stress and Well Being, vol.
5. Elsevier JAI, Amsterdam, pp. 171–201.
Fox, S., Spector, P.E., Goh, A., Bruursema, K., 2007. Does your coworker
know what you’re doing? Convergence of self- and peer-reports of
counterproductive work behavior. International Journal of Stress
Management 14 (1), 41–60.
Frese, M., Zapf, D., 1988. Methodological issues in the study of work
stress: objective vs. subjective measurement of work stress and the
question of longitudinal studies. In: Cooper, C.L., Payne, R. (Eds.),
Causes, Coping and Consequences of Stress at work. Wiley, Chiche-
ster, England, pp. 375–410.
Hobfoll, S.E., 1989. Conservation of resources: a new attempt at concep-
tualizing stress. American Psychologist 44, 513–524.
Hobfoll, S.E., 2002. Social and psychological resources and adaptation.
Review of General Psychology 6 (4), 307–324.
Hockey, G.R.J., 2000. Work environments and performance. In: Chmiel, N.
(Ed.), Introduction to Work and Organizational Psychology: A
European Perspective. Blackwell Publishers, Oxford, pp. 206–
230.
Isen, A.M., 1984. Toward understanding the role of affect in cognition. In:
Wyer, Jr., R.S., Srull, T.K. (Eds.), Handbook of Social Cognition, vol. 3.
Erlbaum, Hillsdale, NJ, pp. 179–236.
Jaccard, J., Turrisi, R., Wan, C.K., 1990. Interaction Effects in multiple
Regression. Sage, Newbury Park.
Jones, F., Fletcher, B., 1996. Job control and health. In: Schabracq, M.J.,
Winnubst, J.A.M., Cooper, C.L. (Eds.), Handbook of Work and Health
Psychology. Wiley, Chichester, England, pp. 33–50.
de Jonge, J., Dormann, C., 2003. The DISC Model: Demand-Induced Strain
Compensation Mechanisms in Job Stress. In: Dollard, M.F., Wine-
field, A.H., Winefield, H.R. (Eds.), Occupational Stress in the Service
Professions. Taylor & Francis, London, pp. 43–74.
de Jonge, J., Dormann, C., 2006. Stressors, resources, and strain at work: a
longitudinal test of the triple match principle. Journal of Applied,
Psychology 91 (6), 1359–1374.
de Jonge, J., Dormann, C., van den Tooren, M., 2008a. The demand-induced
strain compensation model: renewed theoretical considerations and
empirical evidence. In: Näswall, K., Hellgren, J., Sverke, M. (Eds.), The
Individual in the Changing Working Life. Cambridge University Press,
Oxford, pp. 67–87.
de Jonge, J., Le Blanc, P.M., Peeters, M.C.W., Noordam, H., 2008b.
Emotional job demands and the role of matching job resources:
a cross-sectional survey study among health care workers.
International Journal of Nursing Studies 45 (10), 1460–
1469.
Karasek, R.A., 1979. Job demands, job decision latitude, and mental strain:
Implications for job redesign. Administrative Science Quarterly 24,
285–307.
Kelloway, E.K., Loughlin, C., Barling, J., Nault, A., 2002. Self-reported
counterproductive behaviors and organizational citizenship beha-
 J. de Jonge, M.C.W. Peeters / International Journal of Nursing Studies 46 (2009) 699–707
viors: separate but related constructs. International Journal of Selec-
tion and Assessment 10, 143–151.
Marcus, B., Schuler, H., 2004. Antecedents of counterproductive behavior
at work: a general perspective. Journal of Applied Psychology 89 (4),
647–660.
Penney, L.M., Spector, P.E., 2005. Job stress, incivility, and counterpro-
ductive work behavior (CWB): the moderating role of negative affec-
tivity. Journal of Organizational Behavior 26, 777–796.
Robinson, S.L., Bennett, R.J., 1995. A typology of deviant workplace
behaviors: a multidimensional scaling study. Academy of Manage-
ment Journal 38, 555–572.
Semmer, N., Zapf, D., Greif, S., 1996. ‘Shared job strain’: a new approach for
assessing the validity of job stress measurements. Journal of Occupa-
tional and Organizational Psychology 69, 293–310.
Skarlicki, D.P., Folger, R., 1997. Retaliation in the workplace: the roles of
distributive, procedural, and interactional justice. Journal of Applied
Psychology 82, 434–443.
707
Spector, P.E., 2006. Method variance in organizational research: truth
or urban legend? Organizational Research Methods 9 (2), 221–
232.
Spector, P.E., Fox, S., Penney, L.M., Bruursema, K., Goh, A., Kessler, S., 2006.
The dimensionality of counterproductivity: are all counterproductive
behaviors created equal? Journal of Vocational Behavior 68, 446–460.
Spector, P.E., Fox, S., 2002. An emotion-centered model of voluntary work
behavior. Some parallels between counterproductive work behavior
and organizational citizenship behavior. Human Resource Manage-
ment Review 12, 269–292.
SPSS, 2006. SPSS for Windows, Release 15.0.0. SPSS Inc., Chicago.
van den Tooren, M., de Jonge, J., 2008. Managing job stress in nursing:
what kind of resources do we need? Journal of Advanced Nursing 63
(1), 75–84.
van de Ven, B., Vlerick, P., de Jonge, J., 2008. The interplay of job demands,
job resources and cognitive outcomes in informatics. Stress and
Health 24, 375–382.