Abstract
(1) Background: Research examining whether activity engagement is related to cognitive functioning in older adults has been limited to using retrospective reports of activity which may be affected by biases. This study compared two measurements (estimated weekly versus reported daily), and whether these activity assessments were related to cognition in older adults; (2) Methods: Participants from US (n = 199) and Australian (n = 170) samples completed a weekly estimate of activity, followed by 7 consecutive days of daily reporting. Differences between weekly estimates and daily reports were found, such that estimations at the weekly level were lower than self-reported daily information. Multivariate multiple regression was used to determine whether total activity, activity domains and the discrepancy between assessment types (i.e., weekly/daily) predicted cognitive performance across three cognitive domains (fluid, verbal, memory); (3) Results: When activity assessments were totaled, neither predicted cognition; however, when activity was grouped by domain (cognitive, social, physical), different domains predicted different cognitive outcomes. Daily reported cognitive activity significantly predicted verbal performance (β = 1.63, p = 0.005), while weekly estimated social activity predicted memory performance (β = -1.81, p = 0.050). Further, while the magnitude of discrepancy in total activity did not significantly predict cognitive performance, domain specific differences did. Differences in physical activity reported across assessments predicted fluid performance (β = -1.16, p = 0.033); (4) Conclusions: The significant discrepancy between the measurement types shows that it is important to recognize potential biases in responding when conducting activity and cognition research.