Fig. 6: Group influence on children’s evaluation of group-opposing evidence in Study 3.
From: Group membership biases children’s evaluation of evidence

A Unit reduction in confidence in children’s belief after seeing counterevidence by condition: The distribution of responses for children in the No Group Condition is in blue, the distribution of responses for children in the Group Condition is in red. Individual data points (open circles) along with the means (closed circles) by condition are in white. Note that three children who increased in confidence after seeing counterevidence were removed from this plot (but not from our analyses) for visual simplicity. Statistical test: model comparison testing the effect of condition as described in the main text (two-sided test; χ²(1) = 60.00, p = 0.019, R2 = 0.09; n = 60 biologically independent replicates). To account for multiple comparisons across panels, we first conducted a full-null model comparison before testing individual effects. B Children’s initial confidence ratings and mean unit change towards the belief supported by counterevidence: Blue and red circles represent children’s initial ratings of confidence by condition; arrows represent means for the unit-difference between children’s initial and secondary ratings of confidence by condition. Note that means include both when a child’s initial belief was elephants or lions (represented on only the left side of the scale for visual simplicity). Overlaid are children’s distribution of responses in (A). The far-left and far-right squares indicated children were really sure there were more elephants or lions, respectively. The middle-left and middle-right squares indicated children were kind of sure there were more elephants or lions, respectively. The middle square indicated children did not know. C Proportion of children revising their belief about the boxes after seeing two boxes of counterevidence: Data are presented as mean values +/- SEM, n = 51 biologically independent replicates. Statistical test: model comparison testing the effect of condition as described in the main text (two-sided test: χ²(1) = 7.46, p = 0.006, R2 = 0.18; n = 51).