My favourite example of this is the Dunning-Kruger effect, in which people who are incompetent at tasks do not realize they are incompetent -- because they don't know what competence looks like. So they assume they're doing "okay", or even above-average, even when objective tests show otherwise:
I read at the time that the reverse is also true -- the very competent tend to assume similar levels of competence in everyone, so they are unaware how good they are. However, this wikipedia page makes no mention of that part.
Building nonsuperficial relationships with people is a great way to get uncomfortable opinions/data about your actions.
I have often been criticized/enlightened by my wife (usually in a nice way ;-)) and ended up eventually agreeing. This happens less often with my good friends but it still happens.
I think there are always going to be illusions in the way of this "help" in every nonsuperficial relationship (whether perceived by zero, one, or both parties). But that does not mean it is a fundamentally untrustworthy source of feedback to me.
Regarding "cognitive dissonance", I wonder if there is a way to recognize and welcome it more (not sure how delusional this is, but I have experienced deep contradictions and they do seem to "feel" a certain way).
It isn't enough by itself, but the best aid I have read to this is PG's essay on keeping your identity small (http://www.paulgraham.com/identity.html). The fewer things you self-identify as, the less dissonance new information is likely to cause.
In fairness people you have "nonsuperficial relationships" with will just as often not tell you things to spare your feelings and, as my grandpa used to say, "a thermometer that's only right half the time is no better than one that's never right at all"
I thought you probably did but I figured I'd make the point a little more bluntly because it set up my real response which was that advice with illusions isn't all that usefull when trying to see yourself accurately.
Some of the most valuable feedback we can get is of the 'unfiltered' variety. Unfortunately, according to this article, when we watch videos of ourselves, we have a filter. The camera, it seems, is only an objective witness when we have someone who is objective (a neutral 3rd party) to critique the video with. In the Toastmasters club I was a member of, we allowed members to videotape their own speeches. Coupled with the detailed written evaluation from a single assigned evaluator and the anonymous mini evaluations from other members of the audience, I think it was an effective foil for our own self-biased critique.
Most people also spend very little time watching themself on video -- I imagine this is a learned skill like any other -- learning to know yourself as a 3d object viewed from a perspective other than a mirror!
http://en.wikipedia.org/wiki/Dunning-Kruger_effect
I read at the time that the reverse is also true -- the very competent tend to assume similar levels of competence in everyone, so they are unaware how good they are. However, this wikipedia page makes no mention of that part.