THE POSITIVE ECONOMICS OF METHODOLOGY

Does an observation that is consistent with a theory constitute strong er evidence for that theory if it was made after the theory was propos ed rather than before (when the observation might have influenced the theory's formation)? How and why might the research strategy of theori zing prior to observation affect the conditional probability that the theory is true? Is there a socially optimal set of research strategies and on what does it depend? In a world of asymmetric information, wha t incentive structure can induce that social optimum? This paper answe rs these questions in a simple, stylized model that provides a startin g point for a rigorous economic model of scientific method. (C) 1996 A cademic Press, Inc.