Applied Hierarchical Linear Modeling for Cross-Cultural Comparison

Abstract

We explain and demonstrate the application of Hierarchical Linear Modeling (HLM) in cross-cultural research. This method of analysis has not been sufficiently explored in social work research, and it can be a highly useful and appropriate statistical approach for making cross-cultural comparisons. We explain the rationale for HLM or multilevel modeling for cross-cultural data analysis, and we provide an example in which we use Stata to test for neighborhood effects across race groups using survey data. We provide Stata commands and examples of testing for invariance of effects across groups while controlling for heteroscedasticity due to neighborhood level effects. Finally, we included geomaps based on the data to provide visualization of neighborhood effects.