I didn't know what tags I should use for sure; so apologies if not appropriate or enough ones.
Suppose we have 5 people answering 20 yes/no questions. One of them, call her a, is fixed as the one whose answers are assumed to be the correct ones, the answers of the other 4 being evaluated in comparison to a's: if their answer to a certain question corresponds to a's answer, they will get a pass for that question, otherwise fail. Now, repeat the whole scenario for each and every one of the other members of the group: each time, fix someone (among the 5) as a reference, and evaluate others' responses compared to her's. Finally, suppose we repeat the entire scenario so far for 4 other groups of 5 people.
I'm interested to know what kind of mathematical modeling method you'd recommend to express this scenario. I have been thinking of modeling this in terms of possible worlds semantics and also metric spaces, but haven't proceeded much so far.
Any approach is welcome. Please let me know if clarifications are needed.
Thanks!