Fig. 4: Linking model parameters to post- vs. pre-learning changes in intergroup impression in the Western sample.
From: Learning reduces ingroup bias more with perceived losses than gains across cultures
a Correlation between the individual weight of negative ingroup prediction errors (Wneg_in) and the post-vs. pre-learning changes in intergroup impressions in the Gain and Loss frame. These changes represent the decrease of ingroup bias in impressions after learning, that is, (ingroup-outgroup)after – (ingroup-outgroup)before. Smaller values indicate greater decreases of ingroup bias. b Framing moderated the association between the Wneg_in parameter and the learning-induced change in ingroup impression. c Framing did not moderate the association between the Wneg_in parameter and the change of outgroup impression.
