Application of a one-element model with response elimination to paired-associate learning data.

Abstract

<p>This study investigates a response elimination model of paired-associate learning. The structure of the model is identical to that of the one-element model except that the assumption of the latter of a constant probability of guessing before learning is replaced by an assumption that subjects guess from a pool of unassociated responses. It is found that the response elimination model fails to provide an exact description of performance before learning, although it does improve on the one-element model. A three-state model is also investigated and it is found that with four parameters much of the data can be accounted for.</p>