Paper Summary
Share...

Direct link:

Estimation of Interactions Between Covariates of Multiple Indicators Multiple Causes (MIMIC) Models With Multilevel Data

Fri, April 4, 12:25 to 1:55pm, Marriott, Floor: Fourth Level, 415

Abstract

The purpose of this simulation study is to compare the performance of estimating interactions of covariates in multiple indicators multiple causes (MIMIC) models with multilevel data structure using analyses accounting for and not accounting for the hierarchical data structure. Some of the advantages of MIMIC model include its relatively simple model specification, and flexibility in modeling the interactions between covariates. The data were simulated in multilevel structure (e.g., students nested in schools), thus analyses methods employing multilevel modeling will be appropriate. The impact of not accounting for the hierarchical data structure will be examined. The recovery of the simulated interaction effect at within, between, or both within and between levels will be also investigated.

Authors