Differential effects of delay on time-based prospective memory in younger and older adults

ABSTRACT

The current study measured forgetting in a time-based, naturalistic prospective memory (PM) task. In Experiment 1, younger and older participants were asked to mail a stamped postcard on a date that was delayed 1, 2, 5, 14, or 28 days in the future. In Experiment 2, a different sample of older participants completed the same task with similar delays to replicate results for the older sample in Experiment 1. Overall, older participants were more likely than younger participants to mail the postcard on time. In addition, delay affected on-time return rates more for the younger participants than the older participants. Younger participants’ return rates illustrated the typical forgetting curve seen in numerous retrospective memory studies (i.e., rapid decline at shorter delays and slower decline for longer delays). However, older participants’ return rates only declined at the longest delays. These results indicate that time-based PM performance declines with an increase in delay, but the form of the decline may differ across age groups.

Keywords:

• Prospective memory
• Forgetting
• Aging

Notes

¹As noted by Kvavilashvili and Fisher (2007), spontaneous retrieval was highly likely to occur in naturalistic PM tasks. We acknowledge that this is possibly reflected in our data; however, our questionnaire did not ask participants to discriminate between intentional and spontaneous retrieval.

²It should be noted that the majority of participants who returned the questionnaire also mailed the postcard on time, thus resulting in a restriction of range for some analyses. Thus, we report primarily qualitative analyses for some of the data.

³We used an ANOVA to examine the age group × delay interaction (see Rosenthal & Rosnow, 1991, for use of ANOVA with dichotomous dependent variables), which was significant, F(4,190) = 2.53, p = .042, indicating that the pattern of performance across delays differed for the two age groups.

⁴To examine whether individual differences as assessed by the measures of cognitive abilities modulated PM accuracy, we performed a logistic regression in which PM accuracy was the outcome variable and delay, age, education, and the scores from the O-Span, DSST, Trails B, and Shipley were entered as predictors. Only two of the predictors approached conventional significance levels: Years of education was negatively related to PM performance (B = −0.30, p = .053) and working memory, as assessed by the O-Span task, was positively related to PM accuracy (B = 0.12, p = .068).