Effects of Ignoring Hierarchical Data Structure in Factor Analyses

• In Event: Roundtable Session 14
  In Roundtable Session: Issues in Multilevel Latent Variable Modeling

Abstract

Educational research often involves hierarchical or nested data such as, students nested within schools. Nested data violates the independence assumption which may lead to an incorrect error term and inflated Type I error rates. Although several studies have examined the consequences of ignoring nested data using general linear models, hardly any studies have investigated the effects when factor analysis (FA) is used. This study intends to determine if within-matrix factor structure is generally similar to total matrix factor structure using several measures with empirical data. Exploratory FA is applied to total, within, and between matrices of each dataset using Mplus. Number of factors and factor loadings are compared for the all matrices of each dataset.

Authors

• Kate Marino, The Pennsylvania State University

• Pui-Wa Lei, The Pennsylvania State University