Fig. 2.

In (a) choice accuracy for each test trial gamble pair is shown collapsed across test block. In (b) the change in choice accuracy from actor to observer learning sessions (observer accuracy–actor accuracy) is plotted separately for each of the nine test blocks. Pairs are labeled according to the probability of a win for each stimulus. Choice accuracy is measured as the probability that participants chose the stimulus with the highest probability of a win. Actor and observer learning differed only for the 40/20 p{win} pair, with observers showing significantly lower accuracy compared to actors. Error bars show the standard error of the mean.