Long-Term Storage Losses: A Dilemma for Multistore Models.

The Chechile and Meyer storage-retrieval analysis was used to study the retention of 18 paired associates. Prior to testing on each trial, Ss digit-shadowed for 20 sec to assure that testing was from the presumed long-term store of multistore models. Ss (N = 80) learned the first A-B list to either a partial learning or an overlearning criterion and then had six trials on a second list. The overlearning Ss showed only retrieval difficulties, whereas the partial-learning Ss showed difficulties with both storage and retrieval. The results are interpreted as contrary to multistore memory models. An alternative framework is suggested to account for the results of the study.