The effects on priming of word frequency, number of repetitions, and delay depend on the magnitude of priming.

Conflicting findings with respect to the effects of experimental manipulations on priming have been reported in previous studies. It is argued that, in many priming tasks, large amounts of task-relevant information are available from various sources, and that, therefore, the information available from a specific study episode will have only a small impact on overall performance level. Under such circumstances, high levels of baseline performance and small priming effects will be observed. The experiments reported here investigated the hypothesis that a high baseline performance in information-processing tasks that are used to measure priming may constrain priming effects. In a series of word-naming experiments, perceptual difficulty and, therefore, baseline performance was manipulated. Under easy conditions, priming effects were relatively small and were not affected by word frequency, spaced repetition, or delay. Under more difficult conditions, priming effects were larger, and significant effects of all of the above-mentioned experimental manipulations were observed. Under conditions that produced the largest priming effects, a significant relationship between priming and explicit-memory performance could be observed. In the last experiment, it was shown that the characteristics of the retrieval task can substantially affect the magnitude of priming. It is argued that priming effects should be considered to reflect interactions between memory traces and the information-processing components of the priming task.