Partisan underestimation of the polarizing influence of group discussion

•In 2 experiments, groups of 4–6 like-minded participants discussed political topics.•Group discussion polarized participants' attitudes.•Participants misremembered pre-discussion attitudes as having been more extreme.•Individuals underestimated the extent to which their attitudes had polarized.•Group polarization pressures appear to be more potent than partisans realize.

Group polarization occurs when people's attitudes become more extreme following discussion with like-minded others. We hypothesized that people underestimate how much a relatively brief group discussion polarizes their own attitudes. People often perceive their own attitudes as unbiased and stable over time. Therefore, people's polarized post-discussion attitudes may cause them to misremember their pre-discussion attitudes as having been more extreme than they were. In two experiments, participants engaged in 15-minute discussions with 4–6 like-minded others regarding two political topics: whether Barack Obama or George W. Bush was the better president (Experiment 1) and whether they supported Barack Obama or Mitt Romney during the 2012 presidential election (Experiment 2). Group discussion polarized participants' attitudes, and participants misremembered their pre-discussion attitudes as having been more extreme than they actually were. Participants' polarized post-discussion attitudes significantly predicted their recalled pre-discussion attitudes, controlling for their actual pre-discussion attitudes, suggesting that their post-discussion attitudes guided reconstruction of their pre-discussion attitudes. These findings have implications for people's awareness of psychological biases and for the societal effects of partisan enclavement.