+1 (208) 254-6996 [email protected]


TABLE 3 Results of Moderated Regression Analysis for Autonomy as Moderator

Don't use plagiarized sources. Get Your Custom Essay on
Just from $13/Page
Order Essay

CWB organizational CWB personal

Step Independent variable Total R2 mR2 Total R2 mR2

1 Org. Constraints .10¤¤¤ .10¤¤¤ .06¤¤¤ .06¤¤¤ 2 Autonomy .14 .03¤¤¤ .06 .06 3 Org. Con. £ Autonomy .14 .00 .07 .01

1 Confict .11¤¤¤ .11¤¤¤ .16¤¤¤ 2 Autonomy .14 .03¤¤¤ .16 .00 3 Confict £ Autonomy .14 .00 .18 .02¤¤

1 Procedural Justice .07¤¤¤ .07¤¤¤ .02¤ .02¤ 2 Autonomy .11 .04¤¤¤ .02 .00 3 P. Just. £ Autonomy .11 .00 .03 .01

1 Distributive justice .03¤¤ .03¤¤ .01 .01 2 Autonomy .08 .05¤¤¤ .01 .00 3 D. Just. £ Autonomy .08 .00 .03 .02¤

Note. N D 292. ¤ p < .05. ¤¤ p < .01. ¤¤¤ p < .001.

TABLE 4 Results of Moderated Regression Analysis for Trait Anxiety as Moderator

CWB organizational CWB personal

Step Independent variable Total R2 mR2 Total R2 mR2

1 Org. Constraints .10¤¤¤ .10¤¤¤ .06¤¤¤ .06¤¤¤ 2 Trait Anxiety .15 .05¤¤¤ .10 .04¤¤¤ 3 Org. Con. £ T. Anxiety .15 .00 .13 .03¤¤

1 Confict .11¤¤¤ .11¤¤¤ .16¤¤¤ .16¤¤¤ 2 Trait Anxiety .15 .04¤¤¤ .18 .02¤ 3 Confict £ T. Anxiety .15 .00 .19 .01¤

1 Procedural Justice .07¤¤¤ .07¤¤¤ .02¤ .02¤ 2 Trait Anxiety .12 .05¤¤¤ .07 .05¤¤¤ 3 P. Just. £ T. Anxiety .12 .00 .07 .00

1 Distributive justice .03¤¤ .03¤¤ .01 .01 2 Trait anxiety .10 .07¤¤¤ .07 .05¤¤¤ 3 D. Just. £ T. Anxiety .10 .00 .07 .00

Note. N D 292. ¤ p < .05. ¤¤ p < .01. ¤¤¤ p < .001.




TABLE 5 Results of Moderated Regression Analysis for Trait Anger as Moderator

CWB organizational CWB personal

Step Independent variable Total R2 mR2 Total R2 mR2

1 Org. Constraints .10¤¤¤ .10¤¤¤ .06¤¤¤ .06¤¤¤ 2 Trait Anger .17 .07¤¤¤ .15 .09¤¤¤ 3 Org. Con. £ T. Anger .17 .00 .16 .01

1 Confict .11¤¤¤ .11¤¤¤ .16¤¤¤ .16¤¤¤ 2 Trait Anger .18 .07¤¤¤ .23 .07¤¤¤ 3 Confict £ T. Anger .18 .00 .26 .03¤¤¤

1 Procedural Justice .07¤¤¤ .07¤¤¤ .02¤ .02¤ 2 Trait Anger .16 .09¤¤¤ .14 .12¤¤¤ 3 P. Just. £ T. Anger .17 .01 .14 .00

1 Distributive justice .03¤¤ .03¤¤ .01 .01 2 Trait anger .14 .11¤¤¤ .14 .13¤¤¤ 3 D. Just. £ T. Anger .14 .00 .14 .00

Note. N D 292. ¤ p < .05. ¤¤ p < .01. ¤¤¤ p < .001.

anxiety did not moderate relations between stressors and organizational CWB. Also, a signifcant trait anger–stressor interaction was found for personal CWB, but not organizational CWB. In each of these last two cases, the pattern of signifcant moderation revealed that when the trait was high, there was a steeper slope between stressor and CWB than when the trait was low. That is, for individuals high in trait anxiety, but not for individuals low in trait anxiety, higher levels of constraints were associated with higher levels of personal CWB. Similarly, for individuals high in trait anger, but not for those low in trait anger, higher levels of confict were associated with higher levels of personal CWB. Thus, only very limited support was found for Hypothesis 4. Personality characteristics (trait anger and anxiety) interacted as expected with stressors for personal, but not organizational, CWB, and weak evidence suggested that autonomy moderated the confict–personal CWB relation in the opposite direction as hypothesized.


Results of the current study provide some support for predictions derived from the stress/emotion/CWB model (Spector et al., 1998). Specifcally, it was found that job stressors, including perceived injustice related to both negative emotions and CWB (Hypothesis 1); that negative emotions related to CWB (Hypothesis 2); and that in most cases there was at least partial mediation of emotions in the relations between job stressors and CWB (Hypothesis 3). This last test is critical to the model and suggests emotions play a central role in the process from stressor to strain in general and CWB in particular.




Indeed, inspection of the mediator analyses in Table 2 shows patterns of relations that are generally supportive. In seven of the eight cases, the stressor variable was correlated signifcantly with CWB. Negative emotion was correlated signifcantly with all stressor and CWB variables. The regression results for all seven cases in which the stressor was related signifcantly to CWB showed a reduction in the regression coeffcient when emotion was added to the equation containing only the stressor as a predictor. In three cases the coeffcient for the stressor lost signifcance, and in fve of seven cases, the magnitude of the coeffcient was reduced by more than half.

In general, organizational stressors (such as constraints and injustice) were more closely associated with organizational than personal types of CWB, and interper- sonal confict was more closely associated with personal than organizational CWB. Yet, the results of post hoc z tests showed these differences to be signifcant only in the cases of autonomy and procedural justice being more strongly associated with organizational CWB and confict being more strongly associated with personal CWB. Similarly, support for mediation was somewhat stronger for organization- targeted than person-targeted CWB. Altogether, the tighter linkage between con- straints inhibiting job performance and CWB directed at the organization rather than other people, suggests the two types of CWB may follow somewhat separate paths.

The failure to fnd more evidence for complete mediation is also not unexpected. Spector and Fox (in press) discussed how causality is likely multidirectional, with background mood or emotional state predisposing people to perceive or not per- ceive job stressors. Thus, people experiencing high levels of negative mood may be hyperreactive to job stressors. This can result in a cycle in which background mood predisposes a person to perceive the work situation as a job stressor, which further induces negative mood and heightened strain.

In general, our results for Hypotheses 1 and 2 were consistent with prior re- search in showing a relation of justice with negative emotions (Skarlicki & Folger, 1997; Zohar, 1995), and with CWB (Skarlicki & Folger, 1997). Justice also was correlated signifcantly with both of the other job stressors (interpersonal confict and organizational constraints). These fndings lend further support to the idea that perceptions of injustice can be conceptualized as a form of perceived job stress. Situations seen by people as unfair are stressors that may lead to negative emo- tions and presumably to subsequent strains beyond CWB. For example, justice has been shown to relate signifcantly with job (dis-)satisfaction (e.g., Moorman, 1991). Subsequent work should expand this area of inquiry to look at other types of strains (e.g., physical symptoms and ill health) that have been linked to stressors at work (Jex & Beehr, 1991).

Our proposed model suggested a moderating role for trait anger and anxiety. In general, the data were nonsupportive of moderation. Using a conservative (.01) ®

(corrected for multiple signifcance tests), only the interactive effects of autonomy and confict, trait anxiety and constraints, and trait anger and confict for per- sonal CWB remained signifcant. Using a more liberal ® of .05, however, patterns




emerged which suggest the usefulness of further investigation of moderating roles for trait anger and anxiety in stressor-personal CWB relations. In these cases, as expected, individuals higher in the affective trait reported being more sensitive to stressors in their personal CWB reactions. There was no evidence of modera- tion for organizational CWB or with the justice variables. Perhaps organizational CWB is more affected by environmental variables than personality, whereas per- sonal CWB is affected by both. Indirect support for this idea comes from work with organizational citizenship behavior (OCB), which in some ways is the op- posite of CWB. McNeely and Meglino (1994) found that the personality trait of empathy predicted OCB directed toward other people at work but not toward the organization. Of course, it should be kept in mind that we looked at only two personality variables and four stressors, and it is conceivable that results would differ for other variables. Furthermore, moderated regression is known to suffer from low statistical power, so future tests should have larger samples.

One of the most surprising fndings was the failure of the data to support the predicted moderating role of job control (autonomy), particularly in the relations between task-related stressors (constraints and injustice) and task-related (organi- zational) CWB. It may be that the category of organizational CWBs is too broad, and contains many behaviors that are not task-specifc, such as destroying com- pany property. To explore this possibility, the moderator analysis was performed using a subset of the organizational CWB items, which appear to directly in- volve work sabotage. This post hoc test revealed a signifcant interaction effect of distributive justice by autonomy on work sabotage behaviors, in the expected direction. That is, for individuals who perceived low autonomy, but not those who perceived high autonomy, low distributive justice was associated with high lev- els of work sabotage behaviors. Future studies might systematically divide the CWB items into subscales according to their relevance to task performance, inter- personal relations, withdrawal, and so forth in order to determine more specifc, differential patterns of relations to organizational antecedents such as job charac- teristics.

Order your essay today and save 10% with the discount code ESSAYHELP