RESEARCH REPORT
Blaming the Organization for Abusive Supervision: The Roles of Perceived
Organizational Support and Supervisor’s Organizational Embodiment
Mindy K. Shoss
Saint Louis University
Robert Eisenberger
University of Houston
Simon Lloyd D. Restubog
Australian National University
Thomas J. Zagenczyk
Clemson University
Why do employees who experience abusive supervision retaliate against the organization? We apply
organizational support theory to propose that employees hold the organization partly responsible for
abusive supervision. Depending on the extent to which employees identify the supervisor with the
organization (i.e., supervisor’s organizational embodiment), we expected abusive supervision to be
associated with low perceived organizational support (POS) and consequently with retribution against the
organization. Across 3 samples, we found that abusive supervision was associated with decreased POS
as moderated by supervisor’s organizational embodiment. In turn, reduced POS was related to heightened
counterproductive work behavior directed against the organization and lowered in-role and extra-role
performance. These findings suggest that employees partly attribute abusive supervision to negative
valuation by the organization and, consequently, behave negatively toward and withhold positive
contributions to it.
Keywords: abusive supervision, perceived organizational support, workplace victimization
Supervisors play an important role in the direction, evaluation,
and coaching of employees. Some supervisors are supportive,
fostering subordinates’ abilities and empowering them to achieve
their goals (House, 1996). In contrast, other supervisors humiliate,
belittle, or otherwise treat subordinates derisively (i.e., abusive
supervision; Tepper, 2000). Subordinates who perceive their su-
pervisors as abusive are more likely to engage in counterproduc-
tive work behaviors directed toward both the supervisor and the
organization (Mitchell & Ambrose, 2007; Restubog, Scott, &
Zagenczyk, 2011), as well as to reduce discretionary behaviors
carried out on behalf of the organization (Aquino & Bommer,
2003; Tepper, 2007).
Prior research has largely viewed the organization as an “inno-
cent bystander” of employees’ displaced aggression and impaired
self-regulation following abusive supervision (Mitchell & Am-
brose, 2007; Restubog et al., 2011; Thau & Mitchell, 2010).
However, recent research suggests that recipients of abusive su-
pervision hold the organization itself partly responsible. Specifi-
cally, abusive supervision was found to be negatively related to
affective organizational commitment (Tepper, Henle, Lambert,
Giacalone, & Duffy, 2008) and was more strongly related to
organization-directed counterproductive work behavior among
those who made organizational attributions for the abuse (Bowling
& Michel, 2011).
Because damaged relationships between employees and the
organization are harmful to both parties, these findings empha-
size the importance of advancing theory and research concern-
ing why and to what degree employees hold the organization
responsible for abusive supervision. Indeed, Levinson (1965)
argued that because the organization is morally and legally
responsible for the actions of supervisors in their role of direct-
ing and evaluating subordinates, employees attribute this treat-
ment partly to the organization. This assumption is embedded in
organizational support theory (Eisenberger & Stinglhamber,
2011; Rhoades & Eisenberger, 2002), which holds that favor-
able treatment by supervisors enhances employees’ perception
that the organization values their contributions and cares about
their well-being (perceived organizational support or POS). In
This article was published Online First December 3, 2012.
Mindy K. Shoss, Department of Psychology, Saint Louis University;
Robert Eisenberger, Departments of Psychology and Management, Uni-
versity of Houston; Simon Lloyd D. Restubog, Research School of Man-
agement, Australian National University, Canberra, Australia; Thomas J.
Zagenczyk, Department of Management, Clemson University.
We thank Chris Asuncion, Susie Eala, Patrick Garcia, Franco Quodala,
Jennifer Lajom, and Lemuel Toledano for their research assistance. This
research was partially supported by Australian Research Council Grant
DP1094023 and an Australian National University College of Business and
Economics Research School Grant awarded to the third author. This
research was also supported by a Summer Research Grant from the
Department of Management at Clemson University to the fourth author.
Portions of this article were presented at the annual meeting of the Society
for Industrial and Organizational Society, San Diego, California.
Correspondence concerning this article should be addressed to Mindy K.
Shoss, Department of Psychology, Saint Louis University, Saint Louis,
MO 63103. E-mail: mshoss@slu.edu
Journal of Applied Psychology © 2012 American Psychological Association
2013, Vol. 98, No. 1, 158 –168 0021-9010/13/$12.00 DOI: 10.1037/a0030687
158
the research reported here, we draw from organizational support
theory to provide insight into why and to what extent victims of
abusive supervision retaliate against the organization for treat-
ment received from their supervisors. In doing so, we offer
organizational support theory not as a replacement of previous
theoretical perspectives but rather as a complementary theory
that adds depth to researchers’ understanding of why abusive
supervision results in employee behaviors harmful to the orga-
nization.
Although many studies have demonstrated a positive relation-
ship between favorable treatment by the supervisor and POS
(Rhoades & Eisenberger, 2002), the impact of negative treatment
by supervisors on POS has received little attention. This is signif-
icant because negative experiences should not be regarded merely
as a deviation from the positive, but rather viewed as a separate
phenomenon (Eby, Butts, Lockwood, & Simon, 2004). Actions by
the organization and its representatives that are viewed as discre-
tionary are much more strongly associated with POS than actions
over which the organization has limited control (e.g., the favor-
ableness of union contracts or health and safety regulations; Eisen-
berger, Cummings, Armeli, & Lynch, 1997). Abusive supervision
similarly involves willful mistreatment of subordinates (Tepper,
2007). From the employees’ viewpoint, if the organization assigns
supervisory responsibilities to an individual who humiliates and
insults them, the organization as well as the supervisor acts will-
fully and holds them in contempt, which should relate negatively
to POS. Likewise, the organization’s failure to take corrective
action to lessen or prevent the abuse may suggest to employees
that the organization does not value their well-being (i.e., low
POS).
Hypothesis 1. Abusive supervision is negatively related to
POS.
Organizational support theory also suggests that employees may
differ in the extent to which they ascribe responsibility to their
organization for abusive supervision. Employees recognize that
supervisors act in part toward them on the basis of distinctive
motives and values as well as their common interests with the
organization. Thus, employees vary in the extent to which they
identify their supervisor with the organization (supervisor’s orga-
nizational embodiment, SOE; Eisenberger et al., 2010). Eisen-
berger and colleagues found that the relationship between leader–
member exchange and affective organizational commitment was
stronger among employees who believed that their supervisor
embodied the organization to a greater degree. Adopting similar
logic here, we suggest that supervisor’s organizational embodi-
ment strengthens the negative relationship between abusive super-
vision and POS.
Hypothesis 2. SOE moderates the abusive supervision–POS
relationship such that high SOE strengthens the negative re-
lationship between abusive supervision and POS.
Intentional devaluation by others, including one’s employer,
is demeaning and challenges one’s status as a worthwhile
individual (Restubog, Hornsey, Bordia, & Esposo, 2008). Just
as there is a positive reciprocity norm that calls for the return of
favorable treatment, there is a negative reciprocity norm that
validates and invites the return of mistreatment (Eisenberger,
Lynch, Aselage, & Rohdieck, 2004; Gouldner, 1960). Where
high POS conveys the organization’s positive valuation of the
employee’s contributions and concern about the employee’s
well-being, low POS tends to evoke revenge by (a) disregarding
the organization’s social responsibility to act supportively to
those dependent on it (cf. Berkowitz & Daniels, 1964), (b)
failing to maintain commonly accepted standards of humane
treatment of employees, and (c) providing an external threat to
self-esteem (Baumeister & Leary, 1995). A longitudinal inves-
tigation by Ferris, Brown, and Heller (2009) revealed that
employees with low POS had low organization-based self-
esteem, which, in turn, was related to increased counterproduc-
tive work behavior.
Thus, we expected POS to mediate the relationship between
abusive supervision and employees’ counterproductive work
behaviors directed against the organization. We also expected
POS to mediate the relationship between abusive supervision
and employees’ in-role and extra-role work performance. De-
creased POS resulting from abusive supervision likely lessens
employees’ motivation to help the organization achieve its
objectives. As previously noted, we expected SOE to moderate
the relationship between abusive supervision and POS. That is,
when employees identify the supervisor with the organization,
they are likely to view the organization as treating them poorly
and retaliate accordingly.
Hypotheses 3–5. The conditional indirect effect of abusive
supervision on (Hypothesis 3) organization-directed coun-
terproductive work behavior, (Hypothesis 4) in-role perfor-
mance, and (Hypothesis 5) extra-role performance via POS
will be stronger when SOE is high than when SOE
is low.
In sum, the present research draws from organizational support
theory to provide insight into why and to what extent abusive
supervision is associated with employees’ less favorable orienta-
tion toward the organization. In particular, support for our hypoth-
eses would suggest that, to the extent that employees view their
supervisor as representative of the organization, abusive supervi-
sion plays a more central role in employees’ view of the organi-
zation than previously considered. Specifically, when a supervi-
sor’s organizational embodiment is high, abusive supervision is
likely to be strongly associated with employees’ perceptions that
they are devalued by the organization. We expected to find that
this harm to the employee–organization relationship leads to im-
portant negative consequences for the organization. While our
hypotheses propose full mediation, in future research, other pos-
sible mediators of the relationship between the Abusive Supervi-
sion 3 SOE interaction and performance harmful to the organiza-
tion (including such emotional reactions as anger and depression)
should be considered.
We tested our hypotheses in three samples, using constructive
replication (Lykken, 1968) to obtain a wider diversity of measure-
ments of the key work outcomes and to reduce possible concerns
regarding method variance (see Table 1). As shown in Table 1,
Sample 1 was used to test Hypotheses 1–3; Samples 2 and 3 were
used to test all hypotheses. In Sample 1, we assessed the outcome
variable (supervisor-rated counterproductive work behavior) 3
months after the assessment of the other variables. In Sample 2, we
159
ABUSIVE SUPERVISION AND POS
used a cross-sectional design. Finally, in Sample 3, we assessed
employees’ perceptions of abusive supervision and SOE at Time 1
and assessed perceptions of POS a year later at Time 2. With
Sample 3, we also obtained peer reports of counterproductive work
behavior and extra-role performance at Time 2 because peers may
be able to more readily observe these behaviors than supervisors.
Thus, the use of peer reports constructively replicates findings
from Samples 1 and 2. To further strengthen our confidence in our
findings, we also obtained archival performance records that cap-
ture specific dimensions of work performance over a 12-month
period in Sample 3, which may be more accurate than a one-shot
assessment of performance obtained in supervisor ratings. Across
the three samples, we obtained constructive replications as we used
self-reports (because employees have more accurate knowledge of
their counterproductive work behavior; Berry, Carpenter, & Bar-
ratt, 2012) as well as supervisor and peer ratings of work behav-
iors, which are likely to yield more comprehensive evidence than
using self-reports alone. Thus, we are able to corroborate the
moderated mediation effects across three sample groupings, three
different designs, and various construct operationalizations (self-
ratings, peer ratings, supervisor ratings, and archival performance
data).
Method
Participants and Procedure
Sample 1. English language surveys were distributed to 326
full-time employees enrolled in part-time master’s of business
administration (MBA) programs in a large university in the Phil-
ippines. The vast majority of Filipinos, especially in business
organizations, speak English (Bernardo, 2004). We assessed abu-
sive supervision, POS, and SOE at Time 1. We obtained supervisor
ratings of employees’ counterproductive work behavior 3 months
later at Time 2.
Of 326 part-time MBA students who received questionnaires,
239 returned the surveys (73.3% response rate). Three months later
(Time 2), each of the 239 participants received a behavioral rating
form to be completed by their supervisor. One hundred fifty-nine
supervisors completed surveys (66.5% response rate). Eleven sur-
veys were disregarded because of incorrect or missing identity
codes. Thus, the two waves of data collection resulted in 148
matched employee–supervisor dyads. A series of t tests established
that participants who had supervisor reports did not differ signif-
icantly from those without supervisor reports in terms of gender,
age, tenure, and perceptions of abusive supervision. Two research
assistants randomly contacted 10% of the participating supervi-
sors; all provided accurate information supporting the integrity of
the data.
Approximately half of the participants were male (51.4%). The
average age of employees was 28.3 years (SD 5 4.14, range 5
23– 41). Participants’ tenure was reported as follows: 1–5 years:
51%; 6 –10 years: 44.8%; and 11–15 years: 4.2%. Participants
worked in a variety of occupations, such as human resources and
administration, marketing and sales, engineering and operations,
and information technology. Among the supervisor participants,
53.4% were men. Average age and tenure were 34.96 years (SD 5
5.26) and 5.43 years (SD 5 4.00), respectively.
Sample 2. Surveys were distributed to 565 full-time working
professional members of a large professional organization in the
Philippines. Surveys were completed by 372 employees
(65.84% response rate) as well as 273 supervisors, who pro-
vided behavioral ratings of their employee’s performance
(48.3% response rate), yielding a total of 254 employee–
supervisor dyads after those with missing data were removed
(no supervisor rated more than one employee). Participants with
and without supervisor reports did not differ in terms of gender,
age, tenure, and perceptions of abusive supervision. To ensure
the integrity of the data, two research assistants randomly
contacted 10% of the participating supervisors; all supervisors
provided information that matched the information provided in
the surveys. Employees were 63% female, averaged 30.8 years
old, and worked in a wide variety of occupations. Employees’
organizational tenure were as follows: less than 1 year: 12%;
1–5 years: 55%: 6 –10 years: 15%; 11–15 years: 4%; 16 –20
years: 6%; 21–25 years: 4%; more than 30 years: 0.4%; and
0.8% failed to report their tenure. The majority of the employ-
ees (87%) held college/university degrees.
Sample 3. At Time 1, surveys were administered to 1,310
full-time employees from a large financial organization in the
Philippines via the interoffice mailing system and were completed
during lunch breaks. We received 428 employee surveys, repre-
senting a response rate of 32.7%, which is acceptable for mail
Table 1
Summary of Research Design for Each Sample
Variable
Sample 1
Hypotheses 1–3 tested
Sample 2
Hypotheses 1– 4 tested
Sample 3
Hypotheses 1– 4 tested
Outcome assessed CWB CWB, in-role performance, and
extra-role performance
CWB, in-role performance, and extra-
role performance
Outcome rating source
CWB Supervisor report Self-report Peer report
In-role performance Supervisor report Supervisor report; archival
performance data
Extra-role performance Supervisor report Peer report
Research design 3-month lag for supervisor
report of CWB
Cross-sectional 1 year between assessment of abusive
supervision/SOE and remaining
variables
Note. CWB 5 counterproductive work behavior (specifically, we assessed organization-directed counterproductive work behavior); SOE 5 supervisor’s
organizational embodiment.
160
SHOSS, EISENBERGER, RESTUBOG, AND ZAGENCZYK
questionnaires (Cavana, Delahaye, & Sekaran, 2001). The Time 1
survey assessed demographic variables, abusive supervision, and
SOE.
A second wave of data (Time 2) was collected approximately
12 months after the first survey was disseminated. At this point,
we collected data on mediator (i.e., POS) and outcome variables
(i.e., in-role performance, organizational citizenship behaviors
[OCBs], and counterproductive work behaviors). The Time 2
survey was administered to 428 employees who participated in
the Time 1 survey. In addition, the personnel division of the
organization identified one peer or coworker who worked
closely with the focal employee on a regular basis to provide
ratings on the extent to which the employee had engaged in
extra-role and organization-directed counterproductive work
behaviors. A total of 266 employee surveys were retrieved for
a response rate of 62.15%. In addition, we received 195 super-
visor surveys and 235 peer surveys. After removing target
employees, peers, and supervisors who inattentively filled out
the survey, we had 187 employees with complete Time 1 and
Time 2 data. One month after Time 2 data collection, we
obtained archival performance records from the organization
for these employees. With the exception of tenure, t(423) 5
22.89, p , .01, there were no significant differences between
those who participated in Time 1 data collection only and those
who participated in both time points in terms of gender,
t(426) 5 1.87, p 5 .06, and age, t(426) 5 21.38, p 5 .17. Of
the final sample, 55.1% were female; the average age was 30.95
years. Most (94.1%) of the participants were permanent em-
ployees and had worked for their organization between 1 and 5
years (77.3%).
Measures
For all of the scales except counterproductive work behavior,
respondents rated their agreement with each statement using a
7-point Likert-type scale (1, strongly disagree; 7, strongly agree).
We used a 5-point Likert-type scale to assess counterproductive
work behavior in Sample 1 (1, never; 5, every day) and a 7-point
scale in Samples 2 and 3 (1, never; 7, daily).
Abusive supervision (Sample 1, a 5 .91; Sample 2, a 5 .91;
Sample 3, a 5 .87). We measured abusive supervision using
eight items from Tepper’s (2000) Abusive Supervision Scale.
Bivariate correlations in an independent sample of 97 call center
employees in the Philippines indicated that the shortened and full
versions of this scale are highly related (r 5 .97, p , .01).
POS (Sample 1, a 5 .80; Sample 2, a 5 .69; Sample 3, a 5
.74). In line with prior work in POS, we assessed employees’
POS using the eight highest loading items of the Survey of Per-
ceived Organizational Support (Eisenberger, Huntington, Hutchi-
son, & Sowa, 1986; Rhoades & Eisenberger, 2002).
SOE (Sample 1, a 5 .95; Sample 2, a 5 .94; Sample 3, a 5
.92). We used five items to assess the extent to which employees
identified their supervisor with the organization (Eisenberger et al.,
2010). We took the four items from the original scale that assesses
the core construct of the employees’ perception that the supervisor
shares the identity of the organization and added a fifth, similar
item to increase potential scale reliability (Hellman, Fuqua, &
Worley, 2006).
Counterproductive work behavior. In Sample 1, supervi-
sors were asked to rate their employees’ counterproductive
behaviors using eight behavioral checklist items from the Coun-
terproductive Work Behavior Checklist (CWB–C; Spector et
al., 2006) at Time 2. Seven of the items came from the produc-
tion deviance (e.g., “Purposely failed to follow instructions”)
and withdrawal (e.g., “Came to work late without permission”)
subscales. The final item reflected speaking negatively about
the organization (a 5 .85). We selected these behaviors be-
cause they are less strongly associated with concerns about
being punished than more serious forms of organization-
directed counterproductive work behavior (Fox & Spector,
1999); therefore, supervisors may more readily observe or
otherwise be informed about them. Bivariate correlations in an
independent sample of 158 full-time employees in the United
States indicate that the shortened and full versions of this scale
are highly correlated when self-reported (r 5 .88, p , .01). In
Sample 2, employees were asked to respond to Aquino, Lewis,
and Bradfield’s (1999) eight-item measure of organization-
directed counterproductive work behavior (a 5 .84). In Sample
3, peers provided reports of employees’ organization-directed
counterproductive work behavior by responding to the five
highest factor-loading items in the organizational deviance
measure developed by Aquino et al. (1999; a 5 .75). Bivari-
ate correlations in an independent sample of 158 employees in
the Philippines indicate that the shortened and full versions of
this scale are highly correlated (r 5 .93, p , .01).
In-role performance. We assessed in-role performance in
two ways. First, supervisors in Samples 2 (a 5 .90) and 3 (a 5
.94) assessed their subordinates’ in-role performance with the
four highest loading items from Williams and Anderson’s
(1991) in-role performance scale. We used this four-item mea-
sure due to survey length restrictions. Prior research has dem-
onstrated that the shortened version has good reliability (e.g.,
Podsakoff, Bommer, Podsakoff, & MacKenzie, 2006; Restu-
bog, Bordia, & Tang, 2006). Bivariate correlations in an inde-
pendent sample of 180 employees in the Philippines revealed
that the shortened scale and the full scale were highly correlated
(r 5 .85, p , .01). Second, we obtained archival performance
data in Sample 3 one month after Time 2 data collection. The
archival performance records provide an overall performance
measure based on an assessment of critical work behaviors
(e.g., planning and organization, communication skills) and
performance of key result areas using a 5-point behaviorally
anchored rating scale (1, needs improvement; 5, excellent). This
overall rating encompasses ratings made over a 12-month work
period (beginning shortly after the Time 1 data collection).
Extra-role performance. In Sample 2, supervisors assessed
their subordinate’s extra-role performance using a four-item
scale designed by Podsakoff, MacKenzie, Moorman, and Fetter
(1990; a 5 .94). We chose to assess the civic virtue dimension
of OCBs because of length restrictions imposed by the partic-
ipating organization and because civic virtue has been argued to
assess behavior most clearly directed to aid the organization
(Organ, 1988; Robinson, 1996; Robinson & Morrison, 1995),
which should be influenced by POS. Consistent with the meta-
analytic finding that the OCB dimensions are closely related
(LePine, Erez, & Johnson, 2002), we found with an independent
sample of 138 bank employees in the Philippines that the
161
ABUSIVE SUPERVISION AND POS
four-item Civic Virtue Behavior subscale and the 20-item OCB
full scale were highly correlated (r 5 .80, p , .01). In Sample
3, peers provided ratings on this measure (a 5 .90).
Control variables. In order to rule out alternative explana-
tions for our findings, we examined employee age, gender, and
tenure as potential control variables in all three samples because
they have been linked to counterproductive work behavior
(Berry, Ones, & Sackett, 2007). In Sample 1, because tenure
was significantly related to POS (r 5 2.22, p 5 .01), we
included it in the analyses (Becker, 2005). In Sample 2, because
gender was significantly related to counterproductive work
behavior (r 5 2.21, p , .01), we controlled for gender. In
Sample 3, no controls were significantly related to the out-
comes; thus we did not include them in the analyses.
Results
Measurement
We conducted confirmatory factor analyses (CFAs) in order
to assess the measurement model involving the self-rated vari-
ables for Samples 1 and 2 as the key study variables were
assessed at the same time. As recommended by Byrne (2012),
we used MLM estimation as it is robust against nonnormality in
the data and allowed the errors for the negatively worded POS
items to covary to take into account the fact that the similarity
in wording causes additional covariance to that of the focal
factor (Reeve et al., 2007). The hypothesized model fit the data
reasonably well: Sample 1, x
2
(180) 5 313.69, p , .01, confir-
matory fit index (CFI) 5 .92, root mean square error of ap-
proximation (RMSEA) 5 .07; and Sample 2, x
2
(365) 5 692.86,
p , .01, CFI 5 .90, RMSEA 5 .06. For both samples, the
models fit better than the alternative nested models at p , .01,
supporting the discriminant validity of these constructs. In line
with the procedures proposed by Podsakoff, MacKenzie, Lee,
and Podsakoff (2003) and by Williams, Cote, and Buckley
(1989), we tested for the influence of common method variance
with a separate CFA model wherein the self-reported items
loaded both on their respective factors and on a method factor.
The average variance explained in the items by the method
factor in Sample 1 was 17% and in Sample 2 was 14.5%, which
is below the 25% median reported by Williams et al. (1989) for
studies using self-reported variables.
Zero-Order Correlations and Hypotheses Testing
We present descriptive statistics, variable intercorrelations, and
scale reliabilities (a) for all three samples in Table 2. Consistent
Table 2
Descriptive Statistics and Variable Intercorrelations in Samples 1, 2, and 3
Variable M SD 1 2 3 4 5 6 7 8 9 10
Sample 1
1. Gender 1.49 0.50 —
2. Age 28.30 4.14 2.13 —
3. Tenure
a
1.49 0.58 2.07 .75
pp
—
4. Abusive supervision 2.18 1.20 2.11 2.15 2.15 (.91)
5. SOE 5.13 1.27 2.08 .10 2.01 2.48
pp
(.95)
6. POS 5.01 0.95 2.11 2.07 2.22
p
2.34
pp
.34
pp
(.80)
7. CWB 1.63 0.65 .08 .06 .10 .38
pp
2.12 2.32
pp
(.85)
Sample 2
1. Gender 1.63 0.48 —
2. Age 30.84 8.28 .14
p
—
3. Tenure
a
2.58 1.39 .12 .61
pp
—
4. Abusive supervision 2.26 1.12 .03 .15
p
.11 (.91)
5. SOE 4.43 1.46 2.03 .09 2.06 2.25
pp
(.94)
6. POS 4.76 0.85 .12 .10 2.02 2.23
pp
.42
pp
(.69)
7. CWB 1.78 0.82 2.21
pp
2.09 2.07 .17
pp
2.07 2.19
pp
(.84)
8. Extra-role performance 5.00 1.20 .04 2.03 .01 2.16
pp
.23
pp
.31
pp
2.04 (.94)
9. In-role performance 5.24 1.21 .02 2.04 2.09 2.20
pp
.24
pp
.30
pp
2.02 .70
pp
(.90)
Sample 3
1. Gender 0.45 0.50 —
2. Age 30.95 4.89 .19
p
—
3. Tenure
a
1.98 0.51 .02 .47
pp
—
4. Abusive supervision 2.36 1.02 .00 .08 2.02 (.87)
5. SOE 4.45 1.27 .12 2.01 2.07 2.28
pp
(.92)
6. POS 4.78 0.78 .06 .06 .00 2.18
p
.39
pp
(.74)
7. CWB 1.52 0.75 2.07 2.03 .00 .22
pp
2.24
pp
2.30
pp
(.75)
8. Extra-role performance 5.08 1.17 2.04 .00 2.06 2.17
p
.14 .32
pp
2.10 (.90)
9. In-role performance 5.25 1.23 2.10 2.02 2.03 2.17
p
.14 .34
pp
2.12 .68
pp
(.94)
10. Archival performance rating 3.72 0.78 2.02 2.05 2.05 2.12 .13 .35
pp
2.08 .54
pp
.64
pp
—
Note. Reliability coefficients are displayed in the diagonal. SOE 5 supervisor’s organizational embodiment; POS 5 perceived organizational support;
CWB 5 counterproductive work behavior.
a
Tenure was coded as follows: 1 5 1–5 years, 2 5 6 –10 years, 3 5 11–15 years, 4 5 16 –20 years, 5 5 21–25 years, 6 5 26 –30 years, 7 5 more than
30 years.
p
p , .05.
pp
p , .01.
162
SHOSS, EISENBERGER, RESTUBOG, AND ZAGENCZYK
