Measurement Equivalence Using Generalizability Theory: An Examination Of Manufacturing Flexibility Dimensions

摘要

As the field of decision sciences in general and operations management in particular has matured from theory building to theory testing over the past two decades, it has witnessed an explosion in empirical research. Much of this work is anchored in survey-based methodologies in which data are collected from the field in the form of scale items that are then analyzed to measure latent unobservable constructs. It is important to assess the invariance of scales across groups in order to reach valid, scientifically sound conclusions. Because studies have often been conducted in the field of decision sciences with small sample sizes, it further exacerbates the problem of reaching incorrect conclusions. Generalizability theory can more effectively test for measurement equivalence in the presence of small sample sizes than the confirmatory factor analysis (CFA) tests that have been conventionally used for assessing measurement equivalency across groups. Consequently, we introduce and explain the generalizability theory (G-theory) in this article to examine measurement equivalence of 24 manufacturing flexibility dimension scales that have been published in prior literature and also compare and contrast G-theory with CFA. We show that all the manufacturing flexibility scales tested in this study were invariant across the three industry SIC groups from which data were collected. We strongly recommend that G-theory should always be used for determining measurement equivalence in empirical survey-based studies. In addition, because using G-theory alone does not always reveal the complete picture, CFA techniques for establishing measurement equivalence should also be invoked when sample sizes are large enough to do so. Implications of G-theory for practice and its future use in operations management and decision sciences research are also presented.