A Positivity Effect in Older Adults’ Memorability Judgments of Pictures

Abstract

Background/Study Context: A growing literature suggests that older adults are biased to preferentially cognitively process positively valenced information. The authors investigated whether this bias extended to preferential selection of information to remember, and also examined whether the arousal invoked by stimuli biased item selection and memory.

Methods: Thirty older (63–88 years of age) and 30 younger (18–25 years of age) adults viewed emotional (positive, negative) and neutral pictures that varied in arousal (low, high), and were asked to select a subset they deemed memorable (memorability judgments), before recalling pictures. Repeated-measures analyses of variance were conducted to examine aging-related differences in selection and recall of positive, negative, and neutral pictures, and of low- and high-arousal pictures.

Results: Older adults selected more positive pictures as memorable, whereas in younger adults selection did not differ by valence. In both age groups, recall of positive pictures was highest. Older adults selected more low- than high-arousal pictures as memorable, although recall was greater for high- than low-arousal pictures in both age groups.

Conclusion: Findings are consistent with the view that the aging-related positivity bias is under cognitive control, and suggest an awareness of this in older adults. Future investigations should seek to disentangle the influence of positive valence from other factors (e.g., perceptual, semantic, arousal level) on older adults’ memorability judgments.

Acknowledgments

Jennifer C. Tomaszczyk is now at Toronto Rehab, Research Department.

Notes

Note. Data are mean values with standard deviation in parentheses.

¹Older adults took longer to select pictures (M = 107 s, SD = 44.44) than younger adults (M = 85 s, SD = 28.04), F(1, 58) = 4.96, p < .05. Because older adults took more time to select pictures than young, this may have allowed them additional time to encode pictures, leading to the lack of age difference in recall performance.

²Correlation analyses conducted to address the concern that mood may explain our significant findings did not reveal any statistically significant relations between participant mood (positive or negative affect) and picture selection, and did not reveal any systematic relation between mood and recall performance or accuracy of picture memorability predictions, for pictures of any valence type.

³Additionally, we repeated this analysis using proportions of selected and of nonselected pictures that were recalled as the dependent variable, which indicates recall of specific selected and nonselected items. Only the main effect of selection status reached significance, F(1, 56) = 137.91, : Selected items (M = 0.74, SD = 0.14) were more likely to be recalled than nonselected items (M = .46, SD = 0.16). There was also a marginal Selection Status × Age Group interaction, F(1, 56) = 3.81, : Younger adults recalled more nonselected pictures (M = 0.51, SD = 0.16) than older adults (M = 0.41, SD = 0.15), t(58) = 2.45, p < .05, whereas there were no age differences in recall of selected pictures (M _younger = 0.75, SD _younger = 0.15; M _older = 0.74, SD _older = 0.13), p > .1.

⁴Participant ratings of picture valence, arousal, and personal relevance were analyzed using mixed ANOVAs with the between-subjects factor of age group (younger, older), and the within-subjects factors of picture valence category (negative, neutral, positive) and picture arousal category (low, high). Analyses confirmed that participants rated pictures of each valence type, and arousal type, consistently with the normative IAPS ratings given above. In terms of personal relevance ratings, an Age Group × Valence × Arousal interaction, F(2, 116) = 4.10, , indicated that older adults rated negative low-arousal, neutral high-arousal, neutral low-arousal, and positive low-arousal pictures as more personally relevant than younger adults (all ps < .05). However, correlations between rated personal relevance, number of pictures selected, and number of pictures recalled revealed only that, for younger adults, higher ratings of personal relevance of positive pictures was related to greater selection of positive pictures.

Note. *p < .05; **p < .01.

⁵Correlational analyses conducted to address the concern that age group differences on the neuropsychological measures (FSIQ, Trails A and B, Digit Span tasks) may have contributed to the observed pattern in picture selection and recall did not reveal any statistically significant relations between these measures and picture selection, recall, or accuracy of picture memorability predictions, for pictures of any valence type.