Assessing group functioning can be part of group-based learning, referred to here as collaborative learning. The aim of assessment is to encourage collaboration and ensure individual accountability, thereby reducing social loafing and increasing fairness for members. There are many rules proposed by experts, behavioral criteria that promote effective group work, against which group members can be assessed (category-based assessment). In our group-assessment protocol, students define their own rules; a minimum of 5 and a maximum of 7. The main reason for this approach is to give students maximum autonomy. A possible objection to such an approach is that students may overlook some essential criteria and thus (learn to) collaborate incorrectly. To see to what extent expert- and student rules match, we compared expert rules (124 from 12 publications) with student rules (810 from 147 secondary school and university groups). We find that they create the same rules, but that there are some differences in emphasis. We conclude that students have sufficient understanding of collaboration and that when they participate in multiple projects in different groups, they are likely to be exposed to all of the experts’ rules.