Exploring the Test of Covariate Moderation Effects in MultilevelMIMIC Models

摘要

In multilevel multiple-indicator multiple-cause (MIMIC) models, covariates caninteract at the within level, at the between level, or across levels. This studyexamines the performance of multilevel MIMIC models in estimating and detectingthe interaction effect of two covariates through a simulation and provides anempirical demonstration of modeling the interaction in multilevel MIMIC models.The design factors include the location of the interaction effect (i.e.,between, within, or across levels), cluster number, cluster size, intraclasscorrelation (ICC) level, magnitude of the interaction effect, and cross-levelmeasurement invariance status. Type I error, power, relative bias, and root meansquare of error of the interaction effects are examined. The results showed thatmultilevel MIMIC models performed well in detecting the interaction effect atthe within or across levels. However, when the interaction effect was at thebetween level, the performance of multilevel MIMIC models depended on themagnitude of the interaction effect, ICC, and sample size, especially clusternumber. Overall, cross-level measurement noninvariance did not make a notableimpact on the estimation of interaction in the structural part of multilevelMIMIC models when factor loadings were allowed to be different acrosslevels.