This document describes Study 2 of a research project on social influence and optimism. In the study, participant pairs will play an estimation game and receive bogus feedback about their partner's confidence levels. Researchers hypothesize that participants will update their estimates more when their partner agrees and is confident, and that brain regions tracking a partner's rating will be more active for confident versus unconfident ratings. The study uses a 2x2 design to test the effects of partner agreement (agree vs. disagree) and confidence (confident vs. unconfident) on participants' willingness to change their bets.