摘要
若观测变量和潜在特质之间的关系在相比较的群体之间等同 ,就称该量表具备测量等价性。测量工具满足测量等价性的要求是进行组间差异比较的前提条件。该文首先明确测量等价性的概念及在结构方程模型中所要满足的 5个条件 ,然后以社会称许性的性别差异为例 ,阐述测量等价性检验和建立的方法。结果表明 ,男性群体和女性群体的测量等价。
Equivalent measurement is obtained when the relations between observed test scores and the latent attribute measured by the test are identical across subpopulations. Measurement equivalence serves as the precondition for conducting substantive cross group difference comparisons. First, the authors listed the 5 conditions of measurement equivalence within the framework of structural equation modeling (SEM). Then they presented how to test measurement equivalence by taking as an example the gender difference in social desirability measurement. The result indicated that equivalent measurement existed in males and females. The significance and deficiency of the study were discussed.
出处
《心理科学》
CSSCI
CSCD
北大核心
2004年第5期1083-1086,共4页
Journal of Psychological Science
基金
国家自然科学基金资助项目 (70 0 71 0 32 )
关键词
社会称许性量表
心理学研究
测量等价性
结构方程模型
组间差异比较
项目反应理论
measurement equivalence, structural equation modeling (SEM), social desirability, balanced inventory of desirable responding (BIDR)