Necessary Conditions for the Predictive Validity of Strategic Groups: Analysis Without Reliance on Clustering Techniques

Abstract

To investigate necessary conditions for the predictive validity of strategic groups, we grouped firms from five industries, previously studied in strategic groups research, longitudinally by performance using a nonparametric technique; the groupings were confirmed with discriminant function analysis. Stability of performance group membership was examined by ordinal time series analysis and found to be lacking, casting doubt on the efficacy of mobility barriers and hence on the predictive validity of strategic groups.