Goals-feedback conditions and episodic memory: Mechanisms for memory gains in older and younger adults

Abstract

Research has established that challenging memory goals always lead to score increases for younger adults, and can increase older adults’ scores under supportive conditions. This study examined beliefs and on-task effort as potential mechanisms for these self-regulatory gains, in particular to learn whether episodic memory gains across multiple trials of shopping list recall are controlled by the same factors for young and old people. Goals with feedback led to higher recall and strategic categorisation than a control condition. Strategy usage was the strongest predictor of gains over trials for both age groups. Age, goal condition, and effort also predicted scores across the entire sample. Older adults’ gains, but not younger adults’ gains, were affected significantly by the interaction of self-efficacy beliefs and goal condition, and condition interacted with locus of control to predict younger adult gains. These results emphasise the importance of self-regulatory effort and positive beliefs for facilitating goal-related memory gains.

Keywords:

• Memory
• Goal setting
• Self-efficacy
• Memory gains
• Older adults

Acknowledgements

Special thanks to On Top of the World retirement community and to Alyson Sincavage for invaluable contributions to this project. This research was supported by the Evelyn F. McKnight Brain Research Foundation.

Notes

¹For each of the strategies selected as “used most often” on the checklist, a dichotomous variable was created, indicating whether the individual had primarily used that strategy or not. We analysed the impact of each of these dichotomous variables on recall scores. None of these variables showed significant effects except for categorisation. The data for categorisation are reported in the paper (details for other strategies are available on request).

²Because we had numerous subjective and objective indicators of strategy usage, all of the strategy measures available for final trial performance were included in an initial regression analysis to predict gains over trials. To increase the power for the final analysis by reducing the number of indicators, gains were regressed on all of these strategy indicators, using backward elimination to identify the strongest subset of significant predictors from among the strategy variables. This analysis yielded three strategy measures (number of category repetitions, total number of strategies, and HCU) that were not highly correlated with each other, and together predicted the most variance in gain scores. These three variables were then included in all subsequent regressions (details for this preliminary analysis are available on request).