Department of Psychology.
Department of Psychology, The University of Tulsa.
J Appl Psychol. 2014 Nov;99(6):1204-21. doi: 10.1037/a0037965. Epub 2014 Sep 29.
Sexual harassment has received a substantial amount of empirical attention over the past few decades, and this research has consistently shown that experiencing these behaviors has a detrimental effect on employees' well-being, job attitudes, and behaviors at work. However, these findings, and the conclusions that are drawn from them, make the implicit assumption that the empirical models used to examine sexual harassment are properly specified. This article presents evidence that properly specified aggregate construct models are more consistent with theoretical structures and definitions of sexual harassment and can result in different conclusions about the nomological network of harassment. Results from 3 large samples, 2 military and 1 from a civilian population, are used to illustrate the differences between aggregate construct and reflective indicator models of sexual harassment. These analyses suggested that the factor structure and the nomological network of sexual harassment differ when modeling harassment as an aggregate construct. The implications of these results for the continued study of sexual harassment are discussed. (PsycINFO Database Record (c) 2014 APA, all rights reserved).
在过去几十年里，性骚扰受到了大量的实证关注，这些研究一致表明，经历这些行为对员工的幸福感、工作态度和工作行为都有不利影响。然而，这些发现以及从中得出的结论，都隐含着一个假设，即用来检验性骚扰的实证模型是正确指定的。本文提供的证据表明，适当指定的聚合结构模型更符合性骚扰的理论结构和定义，并可能对骚扰的规范网络得出不同的结论。本文使用来自两个军事样本和一个民用样本的 3 个大样本的结果来说明性骚扰的聚合结构模型和反射指标模型之间的差异。这些分析表明，当将性骚扰建模为聚合结构时，其因子结构和规范网络会有所不同。这些结果对性骚扰的持续研究的影响进行了讨论。（APA 版权所有 2014）。
