Fig. 5: Age trends in the social influence model (n = 166 participants).

a Susceptibility to social influence. Difference between the beta distribution mean for p and p’ as a measure of social impact (y-axis) by age (x-axis). Black lines and shapes represent decisions from description, and blue lines and shapes represent decisions from experience. Round shapes and solid lines show trials with safe, triangles and dotted lines show trials with risky social information. Positive values indicate that participants were more likely to choose a risky option, and negative values indicate that participants were more likely to choose a safe option after seeing social information. Shapes show the average in every age bin, the error bars denote the bootstrapped 95% confidence interval, and the lines show a linear fit. b–e Model parameters. Panels depict the population statistics of posterior means of model parameter estimates for each individual (y-axis) by age (x-axis). Boxplots show the interquartile range, whiskers the 95% confidence interval, dots outliers, and the solid line in the box denotes the median of parameter estimates. b Internal uncertainty, which takes separate values for decisions from description (black) and experience (blue). c Social sensitivity, a weight that people put on social information irrespective of their uncertainty, for risky (red) and safe (dark blue) advice. d Reward sensitivity e temperature or randomness of choices.