Model comparison between hierarchical and flat families.

H: Hierarchical; F: Flat. Shown for each model: negative log model evidence −log(P(D|M)); a pseudo-r statistic (p – r2) which is a normalized measure of the degree of variance accounted for in comparison to a model with random choices; the number of subjects favoring the best fitting model based on the model evidence; The exceedance probability which represents the probability that each model (or family) is most likely among alternatives over the population. *best fitting model in each family. **best fitting model.