Abstract
The five experiments reported in this paper examined the effect of two theoretically motivated, explicit mental procedures on performance in different versions of the four-card, abstract selection task. There was no overall benefit as a result of carrying out these procedures; instead, performance depended upon the extent to which subjects carried out the procedures successfully. Most crucially, the experiments demonstrated that correct selection depends not simply upon envisaging the right counter-example, but upon identifying those cards that could contain potential counter-examples. Such findings are broadly consistent with the theory of mental models. However, contrary to the theory, identification is not sufficient for correct selection. The externalization technique used in these studies provides a way to examine the loci of performance differences in the selection task and therefore provides a way to test rival theoretical claims.