I didn't think I could accurately self-report the frequency with which I remember small favours compared to everyone else, because my perception of it was bound to be skewed.

Oh, I'm sure of it. I was at least partially interested in people's possibly inaccurate self-perceptions. (If I'd wanted objectively accurate data, an LJ poll is not how I'd have gone about getting it! :-)

Somewhat tangentially but still related, fanf pointed out this article (http://www.apa.org/journals/features/psp7761121.pdf) (pdf) on twitter, which argues that incompetent people are less aware of their own incompetence than capable people.

That is to say, fanf used Twitter to link to it - the article itself is considerably more than 140 characters.

I was reminded of it on bofhnet and realised I had never saved the link. A classic paper.