Improving the reliability of event-based laboratory tests of prospective memory

Laboratory tests of event-based prospective memory (ProM) require participants to perform actions in response to infrequent cues in a background task. We conducted three experiments to assess and improve the reliability of this popular procedure. In Experiment 1, we tested college students on 2 separate days and found that the alternate-forms reliability of ProM accuracy was quite low (r = .31), although general knowledge accuracy was more reliable (r = .89). In Experiment 2, a statistically significant difference in reliability emerged between conditions with a low (n = 6) versus a high (n = 30) number of ProM targets. Finally, lower ProM accuracy increased reliability in Experiment 3. Adopting these straightforward changes may enhance the search for individual differences in ProM.