Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We study the role of performance differences in a task requiring cognitive effort on in-group bias. We show that the in-group bias is strong in groups consisting of high-performing members, and it is weak in low-performing groups. This holds although high-performing subjects exhibit no in-group bias as members of minimal groups, whereas low-performing subjects strongly do. We also observe instances of low-performing subjects punishing the in-group favoritism of low-performing peers. The same does not occur in high-performing or minimal groups where subjects generally accept that decisions are in-group biased.