Introduction

Decision-making under uncertainty is a hallmark of daily life, yet relatively little is known about how this capacity changes across the adult lifespan. Recent epidemiological data indicate that older adults are less likely than younger adults to engage in financial risk-taking, whether through investment decisions, insurance choices, or gambling. However, the question of whether this reflects rational adaptation to changed life circumstances, risk-preference shifts, or impaired decision-making processes remains unresolved.

One influential framework distinguishes between risk (known probabilities) and ambiguity (unknown probabilities). Foundational work by Ellsberg (1961) showed that people often prefer known probabilities to unknown ones—the ambiguity effect. The present study investigated whether this preference changes with age and whether cognitive capacities such as working memory predict individual differences in risk and ambiguity attitudes across the adult lifespan.

Method

Participants

Forty-eight younger adults (mean age 24.1 years, SD = 3.4, range 19–32; 24 female) and 47 older adults (mean age 70.8 years, SD = 4.9, range 63–83; 25 female) participated. Younger adults were recruited from a university subject pool; older adults from the local community via advertisements. All participants reported no history of neurological or psychiatric illness and were on no psychoactive medications. Written informed consent was obtained; younger participants received course credit, and older participants received $30 CAD.

Procedure

Participants completed a computerized gambling task programmed in E-Prime 1.1. On each trial, they chose between a guaranteed outcome (win $2 CAD) and a gamble. Gambles varied in two dimensions: probability (p = 0.25, 0.50, or 0.75 of winning $8 CAD) and information (risk trials with known probabilities, or ambiguity trials where the winning probability was unknown but equally likely to be any value between 0 and 1). Each participant completed 96 trials (32 per condition) in randomized order. After the gambling task, participants completed the Digit Span subtest from the WAIS-III to assess working memory span. Reaction times were recorded.

Results

A 2 (Age group: younger vs. older) × 3 (Probability: 0.25, 0.50, 0.75) ANOVA revealed a significant main effect of probability on gamble acceptance (F(2,186) = 156.3, p < .001), with both groups accepting gambles more frequently as probability increased. The Age Group × Probability interaction was not significant (F(2,186) = 1.8, p = .17), suggesting comparable probability sensitivity across age groups. However, ambiguity aversion differed substantially by age: older adults chose the ambiguous gamble on 31% of ambiguity trials compared to 49% for younger adults (t(93) = 2.8, p = .006).

Working memory correlated negatively with ambiguity aversion (r = -.38, p < .001). A multiple regression predicting ambiguity-trial gamble acceptance from age, Digit Span score, and their interaction (Model R² = .26, F(3,92) = 10.8, p < .001) revealed that Digit Span was a significant independent predictor (β = .31, t = 3.1, p = .003) even after controlling for age. The Age × Digit Span interaction was not significant (β = -.12, p = .18), suggesting that working memory effects were similar across age groups. Mean reaction times were longer for ambiguity than risk trials (F(1,93) = 78.4, p < .001) and for older than younger adults (F(1,93) = 34.2, p < .001).

Discussion

These findings reveal a dissociation between age-related changes in probability processing and ambiguity preferences. While both younger and older adults integrated probability information into their decisions, older adults showed substantially greater reluctance to accept gambles with unknown probabilities. Notably, this ambiguity aversion was partly explained by working memory capacity, suggesting a cognitive mechanism underlying the effect. Individuals with lower working memory spans—whether younger or older—tended to avoid ambiguous options, possibly because maintaining multiple possible probability distributions exceeds their cognitive resources.

The preserved sensitivity to stated probability in older adults aligns with prior research suggesting that basic probabilistic reasoning remains intact in aging. The age-related increase in ambiguity aversion may reflect either a strategic adaptation to the decision environment (older adults may have experienced greater economic consequences of poor decisions) or a shift in preference toward simplicity and certainty. Future research employing process-tracing methods such as eye-tracking could clarify whether age differences reflect differences in information search strategies during ambiguous decisions. Additionally, extending this paradigm to real financial decisions and examining neural correlates via fMRI may illuminate the computational and biological mechanisms underlying age-related shifts in decision-making under uncertainty.

References

  • Camerer, C., & Weber, M. (1992). Recent developments in modeling preferences: Uncertainty and ambiguity. Journal of Risk and Uncertainty, 5(4), 325–370.
  • Ellsberg, D. (1961). Risk, ambiguity, and the Savage axioms. The Quarterly Journal of Economics, 75(4), 643–669.
  • Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–291.
  • Mather, M., & Johnson, M. K. (2000). Choice-induced preference change in memory: A selective account of enhanced liking. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(5), 1298–1313.
  • Rabbit, P., & Lau, G. (1992). Patterns of cognitive aging. Psychological Bulletin, 112(3), 427–449.
  • Wechsler, D. (1997). Wechsler Adult Intelligence Scale—Third Edition (WAIS-III). The Psychological Corporation.
  • Yechiam, E., & Ert, E. (2007). Evaluating the reliance on past choices in adaptive learning models. Journal of Mathematical Psychology, 51(2), 75–84.