Consistent and low is the only way to go: A polynomial regression approach to the effect of abusive supervision inconsistency.

The literature on abusive supervision largely presumes that employees respond to abuse in a relatively straightforward way: When abuse is present, outcomes are unfavorable, and when abuse is absent, outcomes are favorable (or, at least less unfavorable). Yet despite the recognition that abusive supervision can vary over time, little consideration has been given to how past experiences of abuse may impact the ways employees react to it (or, its absence) in the present. This is a notable oversight, as it is widely acknowledged that past experiences create a context against which experiences in the present are compared. By applying a temporal lens to the experience of abusive supervision, we identify abusive supervision inconsistency as a phenomenon that may have different outcomes than would otherwise be predicted by the current consensus in this literature. We draw from theories on time and stress appraisal to develop a model that explains when, why, and for which employees, inconsistent abusive supervision may have negative outcomes (specifically, identifying anxiety as a proximal outcome of abusive supervision inconsistency that has downstream effects on turnover intentions). Moreover, the aforementioned theoretical perspectives dovetail in identifying employee workplace status as a moderator that may buffer employees from the stressful consequences of inconsistent abusive supervision. We test our model using two experience sampling studies with polynomial regression and response surface analyses. Our research makes important theoretical and practical contributions to the abusive supervision literature, as well as the literature on time. (PsycInfo Database Record (c) 2023 APA, all rights reserved).