Average percentage of trials on which the children (n = 21) selected each of the three objects as matching each label over repeated
test trials (a) and the rate of consistent responses across the duplicate tests (b). Error bars represent standard errors of the mean.