ABSTRACT
The testing effect is often considered a recollection-related phenomenon. However, recent work has observed a benefit of testing to both recollection and familiarity on immediate and delayed final tests. Further, although aging populations show marked declines in recollection, older and younger adults often benefit similarly from testing. This finding suggests that the testing effect in older adults may function via relatively preserved familiarity and lends further support to the hypothesis that the testing effect does not function solely via recollection-related processes. The current study builds on this work to better understand the mechanisms from the dual-process perspective that underlie the testing effect in both younger and older adults. To this end, younger (18–22 year old) and older (65–82 year old) adults studied words, took cued-recall tests on half of the words, and took a final Remember-Know recognition test on all words immediately or after a 1-day delay. At both delays, older and younger adults exhibited a testing effect in both recollection and familiarity, although the magnitude of the testing effect in recollection was reduced for older relative to younger adults. Implications for theories of the testing effect and its application in older adult populations are explored.
Acknowledgements
We are grateful to Dave Balota, Roddy Roediger, Chris Zerr, Nate Anderson, and Thomas Spaventa for helpful feedback and discussion on this project. This research was completed in partial fulfilment of the requirements for the degree of Master of Arts at Washington University in St. Louis to Ruth Shaffer.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Author contributions
Ruth Shaffer and Kathleen McDermott conceptualised and designed the study. Ruth Shaffer collected data, performed data analysis, and wrote the original draft of the manuscript. Both authors contributed to manuscript revisions.
Data availability
All data can be accessed via the Open Science Framework (https://osf.io/4hbgx/).
Notes
1 An alternative measure of accuracy, d’, was calculated for comparison and to adjust for potential differences in response bias across subjects. However, with one exception (see Endnote 2), all patterns of results remained the same when using d’. Thus, only accuracy in terms of hits minus false alarms is reported.
2 The corresponding d’ analysis produced the same patterns of results. However, when a single outlier was removed from the d’ analysis (younger adult, no delay group), the marginally significant interaction between age group and initial learning condition no longer approached significance (p = .139).