Biconditional Rule-Difficulty

Abstract

Three different groups of college age Ss received biconditional rule-learning problems that were altered in such a way as to permit a direct test of an unpublished model of Ss' behavior proposed by C. K. Sawyer and P. Johnson and substantially extended by Salatas and Bourne (1972). The present experiment obtained strong support for the model and evidence to support the widely reported suggestion that the principal difficulty with a biconditional rule is that S must learn to classify together two groups of elements that share no elements in common. The major outcome of the experiment was that strong empirical support was obtained for a useful but arbitrary assumption by Salatas and Bourne concerning a metric for evaluating biconditional rule-difficulty.