Does an individual assign a higher value to a group decision she has explicitly agreed with? Or does she only care about the intrinsic features of the outcome? Since it is difficult to address this question in natural settings, we employ a laboratory experiment where, after the group collectively decides on an issue, each individual may propose a revision to the group decision. We find that outcomes generated by congruent mechanisms –i.e. procedures that incentivize subjects to agree with the winning alternative– are revised to a far lesser extent compared to outcomes generated by outcome-wise identical mechanisms that encourage disagreement.

JEL classification codes: D71, D72.

Keywords: implementation; mechanism design; consensus; agreement; congruence; experiment; endorsements.