A note on theta reliability and metric invariance

A basic problem in empirical research using internal consistency reliability coefficients such as Cronbach's alpha is that the coefficients are sensitive to scale changes in the data. This leads to problems in cross-study comparison where the same variables are perhaps being measured, but in different units, and in substantive theoretical generalization, which must always be qualified in terms of a specified metric. In this note we show that coefficient theta, recently put forth by David J. Armor, does not have this defect—a property which should commend its application in sociological research.