Trial Frequency Outweighs Trial Duration in Associative Learning: Generality and Boundary Conditions

Perceived contingency of a single cue and outcome is based on the relative exposure to four types of events: Cue-outcome pairings (A events), cue-alone presentations (B events), outcome-alone presentations (C events), and events in which neither the cue nor the outcome is presented (D events). Previous experiments found increases in the frequency of event-affected ratings of the perceived contingency between the cue and outcome, even compared to conditions with proportional decreases in the duration of trials (i.e., adjusted frequency conditions). The present experiments tested the generality and boundaries of this adjusted frequency effect by examining whether it generalizes to ratings of multiple cue-outcome dyads, to a cued-recall test, and to both sequential and simultaneous cue-outcome presentations. Experiment 1 revealed a strong effect of frequency but no effect of duration after training with a single cue-outcome dyad; however, a duration effect emerged when training consisted of five cue-outcome dyads. Experiment 2 showed an effect of duration as well as an adjusted frequency effect in contingency ratings after training with five dyads. Experiment 3 extended these observations to a cued-recall test after training with 10 cue-outcome dyads. Experiment 4 used five dyads and found a within-experiment effect of duration on both contingency ratings and cued-recall scores. Whereas Experiments 1 to 4 varied the A events, Experiment 5 varied frequency and duration of the D events with 10 cue-outcome dyads and revealed effects of duration as well as frequency on both cued-recall and cue-outcome contingency ratings. In summary, these experiments detected an increase in the importance of event duration with increases in the number of dyads. Moreover, subject ratings of contingency closely tracked results in a cued-recall test, suggesting that a common mechanism underlies these two measures.