The issue of measurement invariance revisited
The issue of measurement invariance revisited
复制标题
DOI：
10.1177/0022022102250225
复制
发表时间：
2003-03-01
影响因子：
3
通讯作者：
Watkins, D
中科院分区：
文献类型：
--
作者：
Byrne, BM;Watkins, D
本研究的目的有两个：（a）基于协方差结构分析，检验一个众所周知的测量工具在两个文化多样性群体中的等效性;（B）提供不等效项目的结果，以确定其不恒定性的可能决定因素;事后调查包括使用ANOVA检验项目偏倚的证据，的方法和检查项目反应和分布模式的图形显示。示例数据包括对澳大利亚（N = 497）和尼日利亚（N = 439）青少年的自我描述问卷I（SDQ-I）的四个非学术分量表的项目响应。尽管同样指定和拟合的因素结构，为两个文化群体，研究结果显示，证据的测量和结构的非不变性。结果强调了以前的警告，关于解释工具的等效性，他们也增加了越来越多的怀疑，质疑测量仪器是否可以完全等同时，用于跨文化比较。
The purposes of this study were twofold: (a) based on the analysis of covariance structures, to test for the equivalence of a well-known measuring instrument across two culturally diverse groups and (b) provided with findings of nonequivalent items, to identify possible determinants of their noninvariance; post hoc investigations included tests for evidence of item bias using an ANOVA-based approach and examination of graphical displays of item response and distributional patterns. Example data comprised item responses to the four nonacademic subscales of the Self Description Questionnaire I (SDQ-I) for Australian (N = 497) and Nigerian (N = 439) adolescents. Despite similarly specified and well-fitting factor structures for both cultural groups, findings revealed evidence of both measurement and structural noninvariance. Results underscore previous caveats regarding interpretation of instrument equivalence; they add also to the growing body of skepticism that queries whether measuring instruments can ever be totally equivalent when used in cross-cultural comparisons.