This is a small analysis that gives a first insight into the following question : When a random quartet is created, when will people have more fun : if the quartets are formed fully randomly, or if the people rate themselves into three categories and then the quartets get picked randomly while trying to remain in a group.
It is a gross oversimplification that should NOT be taken too seriously. It does give a bit of insight at most.
Source is quartet analysis. The results are at the bottom of it.