Abstract
The relative contribution to learning of input and recall trials was examined under conditions in which subjects were given considerable freedom on input trials to construct their own pairs of verbal elements for learning from an array of elements and to change their pairings at any time on input trials. 25 psychology students were assigned to learn in each of three input-recall sequences. An alternating input-recall sequence produced better performance at recall than sequences in which either more input or more recall trials was present. It was concluded that the finding of relative equivalence in the contribution of input and recall trials of earlier studies resulted from the input procedures used and did not demonstrate a general principle of learning verbal materials. In conditions in which subjects are free to construct their own materials for learning, there is an optimal sequence most closely approximated in the present study by the alternating input-recall sequence.