Results of the generalized linear mixed models to model punishment investment in the public goods game.

Provided are the estimates, the standard errors in brackets and the p-values as * p<0.05,
**p<0.01,
***p<0.001.
The period, the participant’s identity and the session were added as random factors in all models (n = 864, in each round 96 participants could punish up to three subgroup members in three periods). For punishment in round 1 no previous provocation (in terms of punishment investment by the target in the previous round) is possible.
1The contribution of both, the participant (P) and her target (T), into the public good (PG) served as reference group of the categorical fixed factor participant’s and target’s PG decisions.
2The behaviour of the remaining two subgroup members was coded as 0, 1, or both contributed into the PG.