Confirmatory versus comparative approaches to theory testing

This article examines the widely held view that manipulation checks, measures of process, and repeated operationalizations in different settings are frequently essential for rigorous tests of theory. This Confirmatory Approach to theory testing is contrasted with the Comparative Approach, which asserts that any procedures are adequate if they serve in demonstrating the superiority of one explanation to its rivals. Our analysis favors the Comparative Approach. It is shown that manipulation checks, measures of process, and repeated operationalizations are not necessary nor always sufficient for rigorous tests. They have no special status in relation to other convergence procedures that are accepted by the Comparative Approach for producing rigorous theory tests.