Testing for Measurement and Structural Equivalence in Large-Scale Cross-Cultural Studies: Addressing the Issue of Nonequivalence
Testing for Measurement and Structural Equivalence in Large-Scale Cross-Cultural Studies: Addressing the Issue of Nonequivalence
复制标题
DOI：
10.1080/15305051003637306
复制
发表时间：
2010-01-01
影响因子：
1.7
通讯作者：
van de Vijver, Fons J. R.
中科院分区：
文献类型：
--
作者：
Byrne, Barbara M.;van de Vijver, Fons J. R.
跨文化比较研究中的一个关键假设是，该工具以完全相同的方式在所有群体中测量相同的结构(S)(即，该工具是测量的，并且在结构上是对等的)。结构方程模型(SEM)程序通常用于检验这些多组等价性假设。然而，当比较包括大规模的跨文化研究时，标准的结构方程模型策略在统计上和实质上都可能是非常有问题的。根据来自全球27个国家的5482名大学生对家庭价值观量表(Georgas，1999)14个条目的回答，我们描述和说明了这些困难。我们提出并报告了一项双模式双管齐下的战略，重点是国家和规模项目，以确定可能的偏见来源。对如何减少大规模跨文化研究中多组等值检验中的问题提出了建议。
A critical assumption in cross-cultural comparative research is that the instrument measures the same construct(s) in exactly the same way across all groups (i.e., the instrument is measurement and structurally equivalent). Structural equation modeling (SEM) procedures are commonly used in testing these assumptions of multigroup equivalence. However, when comparisons comprise large-scale cross-cultural studies, the standard SEM strategy can be extremely problematic both statistically and substantively. Based on responses to a 14-item version of the Family Values Scale (Georgas, 1999) by 5,482 university students from 27 nations around the globe, we describe and illustrate these difficulties. We propose and report on a dual modal two-pronged strategy that focuses on countries as well as scale items in determining the possible sources of bias. Suggestions for minimizing problems in tests for multigroup equivalence in large-scale cross-cultural studies are proffered.