华人组织中很多领导者会对部属进行差序归类，并实施偏私对待的领导行为。差序式领导虽然会激励“自己人”部属展现利领导行为，但利领导行为不一定利组织。依据Trickle-down理论，领导者的态度和行为会向下涓滴和渗透，诱使员工出现类似的偏私态度和行为，这种为人处事的风格易于导致“自己人”和“外人”的分化，引发职场“冷”暴力（职场排斥和群体隔离）。本项目以此为切入点，构建差序式领导对员工利社会组织行为（利同事、利领导和利组织行为）影响的Trickle-down模型，通过实证和多案例研究验证：差序式领导对员工利社会组织行为的直接效应，以及职场“冷”暴力的中介效应；“圈子”文化是否强化了职场“冷”暴力的中介效应（中介的调节效应）；组织自尊能否缓冲职场“冷”暴力对员工利社会组织行为的负面影响（调节的中介效应）。从而全面揭示差序式领导对员工利社会组织行为的影响机理，为员工利社会组织行为管理提供理论指导。

In Chinese organizations, many leaders like to make differential classification to their subordinates and give more partiality to their favorite subordinates. Although differential leadership may inspire in-group subordinates to show prosocial organizational behavior toward leader, it may be not beneficial to the organization. According to trickle-down theory, leaders’ attitude and behavior will trickle down and penetrate down, which will cause subordinates to show analogous attitude and behavior. The differential leadership style is easy to cause the division between in-group subordinates and out-group subordinates. So it will initiate emotional office abuse in the workplace (e.g.workplace ostracism and social groups segregation). From the perspective of emotional office abuse in the workplace, the project will build Trickle-down Model which is to show the effect of differential leadership on staff prosocial organizational behavior (e.g. prosocial organizational behavior toward individual, prosocial organizational behavior toward organization, prosocial organizational behavior toward leader). We use empirical research and multiple case study methods to verify that: whether differential leadership shows direct effect on staff prosocial organizational behavior, and emotional office abuse in the workplace (workplace ostracism, social groups segregation) plays the mediating effect between differential leadership and prosocial organizational behavior or not; whether "quanzi" culture reinforces the mediating effect of the emotional office abuse in the workplace or not (the mediated moderator model); whether organization-based self-esteem are able to decrease the negetive impact of the emotional office abuse on prosocial organizational behavior or not (the moderated mediator model). The project attempts to reveal the multi-level effcet mechanism of differential leadership on prosocial organizational behavior, so as to provide theoretical guidance for management.