综述了近年来加入乘积指标的潜变量交互效应建模方法,从产生乘积指标的策略、参数约束方法、均值结构与指标中心化的关系三个方面,讨论了建模方法的简化进程。最后总结出同类方法中比较简洁又不失精确的潜变量交互效应建模方法——无需均值结构的无约束方法,并给出了建模步骤。
Recent researches on structural equation modeling with latent interaction were reviewed.The processes of simplification of the modeling were discussed from the three aspects:the strategies of creating product indicators,the approaches of constraining parameters,and the requirement of the mean structure and its relation to indicator-centering.The unconstrained model without using the mean structure was recommended,and the steps for the modeling were summarized.