I was wondering if you could help with the interpretation of a potential experimental design. We are testing two versions of a 'group' exam, score (out of 100) being the outcome variable, aggregated at the group level.

There are 10 participants in total in the experiment, so participants are rotated into groups (giving 252 unique combinations). The exam is always the same questions, and based on the same situation. The idea is to test the old exam on 126 of the 'teams' and the new one on the other 126.

We anticipate that scores will be higher for the new exam and lower for the old one, but they might converge due to the rotation of the 10 participants.

Is this a valid design and how would one go about analysing it? The issue is we would like to do an RCT type experiment but only have 10 participants.

Many thanks

  • $\begingroup$ Can you clarify whether your 'group' includes 5 students? $\endgroup$ – Isabella Ghement May 23 at 15:40
  • $\begingroup$ The idea is that for each 'round', 5 students would do the new exam (with a group score) and 5 students would do the new exam (with a group score). In the next round, the students would be randomly assigned to treatment and control groups until all combinations have been exhausted. $\endgroup$ – John5150 May 29 at 8:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.