关键词: 辱虐管理, 深层调节, 互动公平, 建言

Abstract: The current research majorly explores the methods which can be used to weaken the adverse impact of abusive supervision on employee in terms of organization, team and leadership, while few studies explore this impact from the perspective of employees’ cognition regulation. Based on the theory of cognitive dissonance, the hypothesis put forward in this research is that the emotional labor strategy of deep acting may make contributions to weakening the adverse impacts of abusive supervision on employees’ interactional justice. By deep acting (including positive refocus and perspective taking), the employees could genuinely reach the private acceptance of supervisors’ abuse, reducing their self-esteem threatening. Meanwhile, as the decreased interactional justice can directly affect the voice motivation of the employees, this paper further hypotheses that the interaction between deep acting and abusive supervision has an impact on the employees' voice (including both defensive voice and prosocial voice) through the mediation of interactional justice. To verify the above hypotheses, some surveys are carried out at 3 time points in a state-owned enterprise. At time point 1, the survey conducted a test to the perception of abusive supervision and deep acting (when encountering rude treating from the supervisors) on 623 invited members in 127 teams. At time point 2, which is 2 months after time point 1, 552 members rated their perception of interactional justice. Then, 4 months after time point 1, 109 team leaders rated the defensive voice and prosocial voice of their subordinates. The final matched sample is composed of 530 members from a total of 103 teams, with a response rate of 85.07% for all the individuals. In the results, the hypotheses put forward in this study are all verified. (1)The result of moderated regression analysis suggests that deep acting can do a favor in weakening the negative connection between abusive supervision and interactional justice (B= .18，p＜.001). (2) The bootstrapping approach (2000 resamples) in Mplus 8.0 is further employed, which shows that the interaction between deep acting and abusive supervision has significant indirect effects (through the mediator, interactional justice) on both defensive voice (ind1=-.03, 95% CI = [-.06, -.01], excluding zero) and prosocial voice (ind2=.02, 95% CI = [.01, .04], excluding zero). Possible contributions of this research are as follows: firstly, it has broadened the perspective of response to abusive supervision. In the face of adverse effects caused by abuse, most existing studies only provide the correction thinking from “top-down” to respond to it, and regard employees as powerless to resist. However, from the conclusion of this study, it can be seen that employees are not totally negative in facing abuse. This paper takes “active acting to cognitive dissonance” as a breakthrough, and provides the way to turn passive into positive from the perspective of response. Secondly, it has advanced the research on the relationship between abusive supervision and voice. At present, studies generally believed that the relationship between them is negative. This study reveals that the interaction between deep acting and abusive supervision will improve the adverse effect of voice through interactional justice; and has found that this effect will have different results on different types of voice. Finally, it has verified the effect of deep acting and enriched the positive effect of emotional labor. It is known that study on emotional labor originated from scholars’ concern about its “dark side”. This paper proves the interactive effect of “deep acting strategy” and “abusive supervision”, which has a reduction effect on the negative impact of employees’ interactional justice and subsequent voice, and highlights the effectiveness of this strategy and the “positive side” of emotional labor.

Key words: abusive supervision, interactional justice, deep acting, voice