A Sample Application on Multilevel Structural Equation Models

Abstract

The data collected in order to test the research questions in social sciences usually have multilevel data structures. Because of the limitations of single level analysis, the violation of independence of observations assumption in empirical studies was ignored; either groups or individuals were used as units in the analysis for a long period of time. Such kinds of problems in the studies with hierarchical/clustered data brought about the development of multilevel modeling techniques. Multilevel Structural Equation Modeling allows developing different models which can be tested for each level of the hierarchical data (within group and between groups) simultaneously. In current study, it was aimed to investigate two-level structural model by using data generated from the Monte Carlo data generation facility of Mplus program and to present a demonstration for researchers. Within this context, the model was introduced using the analysis performed by Mplus 5.1, the Mplus syntax used in the analysis was explained and the interpretation of the model parameters was discussed.