Visualizing⋅matching⋅generalizing: Case identification hypotheses and case-level data analysis

Abstract

The traditional and still dominant logic among nearly all empirical positivist researchers in schools of management is to write symmetric (two-directional) variable hypotheses (SVH) even though the same researchers formulate their behavioral theories at the case (typology) identification level. The behavioral theory of the firm, theories of buyer behavior, and Miles and Snow's typology of organization's strategy configurations (e.g., "prospectors, analyzers, and defenders") are iconic examples of formulating theory at the case identification level. When testing such theories, most researchers automatically, unconsciously, switch from building theory of beliefs, attitudes, and behavior at the case identification level to empirically testing of two-directional relationships and additive net-effect influences of variables. Formulating theory focusing on creating case identification hypotheses (CIH) to describe, explain, and predict behavior and then empirically testing at SVH is a mismatch and results in shallow data analysis and frequently inaccurate contributions to theory. This paper describes the mismatch and resulting unattractive outcomes as well as the pervasive practice of examining only fit validity in empirical studies using symmetric tests. The paper reviews studies in the literature showing how matching both case-based theory and empirical positivist research of CIH is possible and produces findings that advance useful theory and critical thinking by executives and researchers.