I've read the Dunning & Kruger article referred to in the editorial. Another interesting fact is that very competent people are also unable to accurately assess their relative performance - they assume everybody else also does about as well as they do. Intriguingly, (in the study) most people tended to assume that their own personal performance was a little over average, centering around the 63rd percentile.

However, if you show competent people the full range of results, they can recognize good and bad answers and will thus significantly increase their self-assessment. However, incompetent people can't evaluate very well which responses are good, so they only shift their self-assessment slightly.

Isn't this also known as the "Peter Principle?"

Olivia, that sounds fascinating — but quite intelligible, granted the trajectory of the research they report here.

M-i-T, if I recall correctly the Peter Principle (in the strict sense) teaches that people get promoted until they reach a level at which they’re no longer competent. That is, the reward for being good at what you do is that you’re promoted to a job where that’s no longer you’re responsibility. I expect that there are nice correlations to the Dunning-and-Kruger research, but Peter was more interested in the principle that promotion from one job level to another removes us from the work at which we’ve shown excellence and assigns to us work that we may not do as well at.

This reminds me of two buttons I used to wear. It was the 80's, thus buttons were almost cool for five minutes.

"Nothing is foolproof to the sufficiently talented fool."

and the other...

"Sufficiently talented fool."

I can vouch for the existence of the Oblivious Incompetents -- they appear at every level at my workplace (not just managers -- the Peter Principle is something else).

One such person frequently instructed her coworkers on entering data into our membership database, and if I hadn't known better, I would have thought she really knew her stuff, judging by her self-confidence.