I don't know who conducted the original study, but I remember that Martin Seligman opened his book Learned Optimism with a discussion of an experiment exactly a

I don't know who conducted the original study, but I remember that Martin Seligman opened his book Learned Optimism with a discussion of an experiment exactly along these lines. Two groups of people -- a group of clinically depressed people, and a control group -- were asked to operate some apparatus that was programmed to fail a certain percent of the time. They were then asked to estimate how much actual control they had over the apparatus. The control group consistently overestimated, while the depressed people were spot on.