James's post Believing Headlines Makes You Stupid put me in mind of an informal experiment in a college Speech class. We were discussing small-group dynamics, learning about Harmonisers and Tavistock Observers and the like, and were assigned group projects from a list. Our group measured changes in measured intelligence when the group dynamics changed.
Bear with me in my description of the set-up. 19 in the class, three in my group, leaving sixteen subjects. Everyone took the "Stranded On The Moon" test, in which you have to rank which items in the crashed ship will be worth carrying to help you survive on your emergency trip back to base. It's a flawed test in many ways, but over time a consensus of answers result in stable scoring. You have a parachute What for? That's out - no atmosphere. But wait! We could carry things in it!
We ranked the scores and told people where they had ranked in the class, then divided them into four groups to take the test again. But we did not tell them that we had put the top four scorers together, then the next four, #5-8, then #9-12, and #13-16. We set them the task of taking the test again as a group.
Group B, with scorers #5-8 got the highest score (by a lot), and Group C with #9-12 finished second, a little ahead of Group A, the top four scorers. The professor explained to us that this was a fairly typical result - every few years a group would pick that experiment off the list and it usually ended up that way. His observation was that knowing you had a top score made you unwilling to listen to others as much as you should. Having an only decently high score provided the right balance of confidence and humility. You could appreciate that someone else might have an insight you missed, while still believing that you were basically on track as to what was required here.
The last group, BTW, just got discouraged and fairly random.
This doesn't generalise to everything, of course. Had the top four known they were the top four they might well have done better as a group than as individuals, being more willing to listen to other "experts" than to random others. If Group D knew that Group B could improve so much, they might have worked better together and gotten somewhere. But it is revealing about the default setting people may have when they are told they are a little better or worse than the others.