Fig. 3
Quadratic effect of face rank on beta weight. (A) The impact of face rank on beta weight. The black line depicts the quadratic regression line and datapoints depict the mean ± 95% CI. The association is quadratic in nature whereby face ranks closer to the mean are weighted more heavily during decision-making, while face ranks further from the mean are weighted less heavily during decision-making, which can be taken as evidence of robust averaging. (B) The impact of face rank on beta weight as a function of variance (blue = low variance, red = high variance). Robust averaging is observed for high, but not low variance trials.
