Monotonic and nonmonotonic lag effects in paired-associate and recognition memory paradigms

In Experiment I the relationship between response recall and the spacing of repetitions (lag), as a function of the retention interval, was investigated in the continuous paired-associate paradigm. At the short retention intervals (2 and 8 events) the lag function was nonmonotonic. At the longer retention intervals (32 and 64 events) the lag function increased monotonically. A version of encoding variability theory was used to explain these results. The theory was then tested in Experiments II and III. In the second experiment, using the Brown-Peterson paradigm, the lag function was monotonic for uncued recall, and nonmonotonic for cued recall. In the third experiment, using the continuous recognition memory procedure, the lag function was a nonmonotonic function of the lag interval between the first two presentations when the interval between the second and third presentations was short. Increasing the latter interval produced a lag function that was again monotonic. The results of the experiments support the theory which emphasizes the nature of the cues available for retrieval.