Sample size in multilevel structural equation modeling – the Monte Carlo approach

Abstract

In the process of sample selection, an important issue is the relationship between sample size and the type and complexity of the statistical model, which is the basis for testing research hypotheses. The paper presents methodological aspects of sample size determination in multilevel structural equation modelling (SEM) in the analysis of satisfaction with the banking products in Poland. The multilevel SEM results from the necessity to take into account both the sample size at the level of individual respondents, as well as at the higher level of analysis and the intraclass correlation coefficient. A comparison of factor loading bias based on the Monte Carlo simulation is made for different cluster sizes and the number of clusters.