The practical relevance of favouritism among students of the same study path is evident in lifelong memberships in fraternities or sororities or in high donations to faculties. In our study, we focus on the in-group favouritism of students by examining the trade-off of acting based on in-group favouritism or a performance signal when decisions are made about whom to choose as a team mate. The novel feature of your study is that the choice of a team mate is either benevolence or relevant to the own output. In the first scenario, only the payoff of the chosen subject changed, whereas in the second scenario, the decision affected the decider's own payoff as well as that of the chosen subject. The subjects ex ante knew the group type (path of study) of the pool of possible team mates and received a signal giving weak information about their ability regarding the task. Intuitively, one would expect more favouritism if the own payoff was not affected by the performance of the chosen team mate. However, we found the opposite. The subjects exerted more favouritism in the revenue sharing scenario. Possibly they expected reciprocal behaviour and less free riding if they selected a team mate belonging to their own group. Interestingly, groups formed based on favouritism did not perform significantly different from groups formed based on the performance signal.