笔者将实验设计为一个主体之间的实验，其中（1）目标人的团体成员和（2）他或她的犯罪历史被独立操纵。然后，三个因变量被指定为2（小组成员：内群体与外类群）×2（犯罪记录：第一次与重复）方差分析。严格按照假设检验法，对各变量进行科学量化。In order to test the second hypothesis, the author use dummy-code method to take both independent variables into account (Group membership: –1 =outgroup offender; +1 = ingroup offender; Criminal history:
The article coauthored by Mario Gollwitzer and Livia Keller (2009) aims to test whether criminal history has a stronger effect on the punitive responses to the offender when he or she is a member of one’s own social category than when he or she belongs to a different social group. Although some previous literature has been devoted to study how the offender’s criminal history and membership are related to the punishment severity respectively, few has probed into how the offender’s membership and his or her criminal history interact with each other. Specifically, this study focuses on an intergroup context where an important social norm is violated, without taking political, historical and other factors into account.
There are two hypothesis (1) punishment severity is higher for repeat ingroup offenders should than first-time ingroup offenders, while criminal history should have no effect on punitive response in the case of outgroup members; (2) such effect should only works in ingroup condition via societal concerns and anger/outrage as mediators. In this way, membership and criminal history are the independent variables; anger/ outrage, societal concerns, and punishment severity is the dependent variables.
97 students from psychology classes at the University of Koblenz-Landau were asked to finish a questionnaire on the case about a fellow student, a student of either psychology or biology and either a first-time or a repeat offender, who hid important books on psychology from the university library. 86 questionnaires were effective.
The authors designed the experiment as a between-subjects one, in which (1) the target person’s group membership and (2) his or her criminal history were manipulated independently. Then, the three dependent variables were designated to a 2 (Group membership: ingroup vs. outgroup) × 2 (Criminal history: first-time vs. repeat) variance analysis. Strictly according to the hypothesis testing method, all the variables were scientifically quantified. In order to test the second hypothesis, the author use dummy-code method to take both independent variables into account (Group membership: –1 =outgroup offender; +1 = ingroup offender; Criminal history:
–1 = first-time offender; +1 = repeat offender). And meanwhile, they standardized the two mediators (societal concerns, anger/outrage) and the dependent variable (punishment severity).
The results are supportive to both of the hypotheses. As to the first hypothesis, in the ingroup condition, a repeat offender should face more anger/outrage, higher societal concerns, and more severe punishment than a first-time one. While in the outgroup condition, criminal history had little impact on the three variables.
As to the second hypothesis, in the case of an ingroup offender, when taking societal concerns as the mediator, the indirect effect of criminal history on punishment severity was considerable. This result indicates that repeat ingroup offenders tend to raise more societal concerns than first-time ingroup offenders, and so that they are punished more severely. Furthermore, when taking anger/outrage as the mediator, the indirect effect of criminal history on punishment severity was also considerable. This shows that repeat ingroup offenders tend to evoke more anger and moral outrage than first-time ingroup offenders, and thus are punished more severely. In the case of an outgroup offender, however, criminal history has no significant indirect effect on punishment severity via neither societal concerns nor anger/outrage as the mediator.