Abstract

Cronbach's Alpha is one of the most popular measure of reliability in psychological research, despite recent literature has extensively questioned its utility. In this work we summarize briefly contexts of application and critical aspects. By presenting a simple Monte Carlo experiment, we compare the performance of Alpha index with three alternatives: the Omega index and two fit indices used in the confirmatory factor analysis (CFA). The experiment suggests that Alpha shows adequate performance only in specific conditions, which are hard to find in practice. In other contexts, and especially when the underlying factors are two rather than one, Alpha tends to lead too high values. We conclude that Alpha cannot be considered the best option for reliability assessment. Rather, it is suggested the adoption a CFA-based approach that provides both more complete information on underlying factor structures, and indices that allow a more accurate reliability assessment.