Extended Data Fig. 4: Individual parameter estimates for the winning model in each experiment (cf. Figure 3).
From: Asymmetric reinforcement learning facilitates human inference of transitive relations
a, Experiment 1 (model Q1). b, Experiments 2-4 (model Q2*+P). Coloured dots show individual participant estimates, where blue, orange and pink colours refer to Experiment 2, 3 and 4, respectively. The diamond shape denotes the mean across experiments. While the estimates of η showed large variability with many values close to the upper bound, we observed no improvement in model fit when increasing the upper bound further.
