Distributional Sensitivity Results.

Distribution of items for participants in the Gaussian and Uniform conditions in Experiment 1. To ease visualization, we normalized each participant’s values. (A) We calculated the average proportion of values of each participant in each of thirty bins, so that each participant had equal weight in the resulting plot. Comparing the aggregate histograms shows that in the uniform condition participants had a flatter distribution. (B) Shape measure of each participant for the Gaussian and Uniform conditions, with the dashed line representing equal values for both conditions. Error bars are 95% confidence intervals (obtained via bootstrapping). Most participants had a more Gaussian sequence in the Gaussian condition (participants below the dashed line). Although most participants lie in the second quadrant (shaded), meaning that they had a Gaussian sequence in the Gaussian condition and a uniform sequence in the Uniform condition, several other participants lie in the first quadrant, meaning their sequences were Gaussian in both conditions. (C) Histogram of each participant’s normalized values for each condition. For most participants, there’s a clear difference between conditions, with the Gaussian values being more concentrated.