Figure 10.
Individual accuracy against the number of shared estimates τ, before (filled dots) and after (empty dots) social information sharing, in the Sorted (black), Unsorted (blue), and Aggregated (red) treatments. The data were separated into instances where subjects received social information of which the average (M) was lower than their personal estimate Xp (D = M − Xp < 0) and instances of which the average was higher than their personal estimate (D > 0). Solid and dashed lines are model simulations before and after social information sharing, respectively. Individual accuracy hardly changes after social information sharing when D < 0, while it improves when D > 0. Error bars were computed using a bootstrap procedure described in the Statistical methods, and roughly represent 1 s.e.