Why a Confirmation Strategy Dominates Psychological Science.

Abstract

Our research explored the incidence and appropriateness of the much-maligned confirmatory approach to testing scientific hypotheses. Psychological scientists completed a survey about their research goals and strategies. The most frequently reported goal is to test the non-absolute hypothesis that a particular relation exists in some conditions. As expected, few scientists reported testing universal hypotheses. Most indicated an inclination to use a confirmation strategy to test the non-absolute hypotheses that a particular relation sometimes occurs or sometimes does not occur, and a disconfirmation strategy to test the absolute hypotheses that a particular relation always occurs or never occurs. The confirmatory search that dominates the field was found to be associated with the testing of non-absolute hypotheses. Our analysis indicates that a confirmatory approach is the normatively correct test of the non-absolute hypotheses that are the starting point of most studies. It also suggests that the strategy of falsification that was once proposed by Popper is generally incorrect given the infrequency of tests of universal hypotheses.