Measurement invariance in comparative Internet use research

Abstract

Comparative studies in communication and Internet research call for equivalent measures of key constructs that are comparable across populations. This article details and applies the concept of measurement invariance within a cross-nationally comparative context. Multi-group confirmatory factor analysis is used to test configural, metric, and scalar invariance in an empirical example and structural equation modeling introduces exogenous predictors of Internet use types. Results support metric invariance for a four-factor Internet usage model in three English-speaking countries. The significance of measurement invariance testing for unbiased comparative research is discussed.