Raykov Tenko
Michigan State University, East Lansing, USA.
Educ Psychol Meas. 2024 Jun 24:00131644241256626. doi: 10.1177/00131644241256626.
A latent variable modeling procedure for studying factorial invariance and differential item functioning for multi-component measuring instruments with nominal items is discussed. The method is based on a multiple testing approach utilizing the false discovery rate concept and likelihood ratio tests. The procedure complements the Revuelta, Franco-Martinez, and Ximenez approach to factorial invariance examination, and permits localization of individual invariance violations. The outlined method does not require the selection of a reference observed variable and is illustrated with empirical data.
讨论了一种用于研究具有名义项目的多组件测量工具的因子不变性和项目功能差异的潜在变量建模程序。该方法基于一种利用错误发现率概念和似然比检验的多重检验方法。该程序补充了Revuelta、Franco-Martinez和Ximenez的因子不变性检验方法，并允许定位个体不变性违反情况。所概述的方法不需要选择参考观测变量，并通过实证数据进行了说明。
