A theory of turnover, social capital losses, and store performance was developed and tested in 38 locations of a restaurant chain. We assessed the ability of social capital losses to predict variance in store-level performance above and beyond that predicted by overall turnover rate and in-role performance losses from turnover. We further predicted that turnover rate and network density would moderate the curvilinear relationship between social capital losses and performance. Results strongly supported turnover rate's moderation of social capital losses: performance declines from network disruptions were more apparent when overall turnover was low. The network density interaction prediction was not supported.

This study reanalyzes data from Tepper's (2000) two-wave study regarding the effects of subordinates' perceptions of supervisory abuse to assess previously unexamined relationships. As predicted, we found that subordinates who more rather than less strongly perceived that they had been abused by supervisors tended to use regulative maintenance tactics with higher frequency. Further, the positive relationship between abusive supervision and subordinates' psychological distress was exacerbated by subordinates' use of regulative maintenance communications, and that relationship was reduced by subordinates' use of direct maintenance communication. Theoretical and practical implications are discussed. In recent years, management researchers have investigated abusive supervision, subordinates' perceptions of supervisors' sustained displays of hostile verbal and nonverbal behaviors (Tepper, 2000: 178).1 Abusive supervision in the form of ridiculing, undermining, and yelling at subordi nates is a source of chronic stress that produces serious negative consequences (Tepper, 2007). Like victims of domestic abuse (Emery & Laumann Billings, 1998), victims of abusive supervision ex perience heightened psychological distress (Duffy, Ganster, & Pagon, 2002), indications of strain that involve dysfunctional thoughts and emotions (e.g., anxiety, depression, and emotional exhaustion). Investigations of how employees respond to abu sive supervision suggest that subordinates perceiv ing more rather than less of it engage in more retal iation and revenge behavior (

A limitation of the organizational citizenship behavior (OCB) literature is that theory and empirical evidence suggest that some employees define OCBs as part of their job. A theoretical framework that addresses this problem is tested in this article. The framework focuses on 2 effects: a role enlargement effect (i.e., employees with more favorable attitudes define OCB as inrole behavior, which, in turn, results in greater citizenship) and a role discretion effect (i.e., the relationship between employees' attitudes and their citizenship will be stronger among employees who define OCB as extrarole behavior). In tests of this framework with 2 independent samples of supervisor-subordinate dyads, role definitions were found to moderate several relationships between procedural justice and OCB, providing support for the role discretion effect. Implications for OCB theory and research are discussed.

The authors thank Ed Miles for comments and suggestions on earlier versions of this research. This article was accepted under the editorship of Daniel Feldman.

The relationships among merit pay raises, trait positive affectivity (PA), and reactions to merit pay increases (pay attitudes and behavioral intentions) were explored in a longitudinal study of hospital employees. Drawing on signal sensitivity theory, the authors expected that PA would moderate the relationship between merit pay raise size and reactions to the increase such that pay raise size would be more strongly related to pay attitudes and behavioral intentions among those low in PA. Results strongly supported the predictions in the case of reactions to the raise amount (happiness and effort intentions) but not for pay level satisfaction. Implications of the results and directions for future research are identified.