Abstract
In this article distinctions are made among some of the different conceptualizations of, and formulas for, Cronbach's coefficient alpha as it is applied to dichotomous or nondi-chotomous, standardized or unstandardized, and weighted or unweighted data. Issues regarding the statistical significance of alpha and the occasionally encountered anomalies of artificially high alpha and negative values of alpha are given particular attention.