Prior studies suggest that cooperation in public good games with heterogeneous returns (benefits) is unstable and relatively low because participants disagree on the appropriate way to contribute, but the role of these normative views remains unclear because they are rarely directly measured or manipulated. We measure each participant's view on the appropriate way to contribute to the public good prior to play, and use this information to sort people into groups that either agree or disagree (two separate conditions) on how to contribute to the public good. We examine how this (dis)agreement affects contributions in a repeated public goods game with peer punishment. Furthermore, after a set of rounds, we remove one player from each group and exchange him/her for a member from another group. Each group thus receives a newcomer in place of an old-timer. The exchange is done such that the extent of normative agreement/disagreement after newcomer entry in the two conditions reverses: groups in the condition where members initially agree on normative views are due to the newcomer entry exposed to the disagreement that the groups in the other condition experienced before newcomer entry, and vice versa. We examine if, and how, newcomer entry affects cooperation in both conditions.
