Incompetent people think they're doing fine because they can't tell the difference between good and bad work. Highly competent people, readily seeing their own flaws, think they don't do as well as they are doing. (At least until you have them grade a bunch of other work, at which point they are better calibrated.) Competent people are competent because their metacognitive ability, their ability to tell good work from bad, drives them forward.
So basically, if I already can see the problems in my work, then my metacognitive ability is sufficient to drive improvement. When I improve enough that my work looks fine to me, further improvement is prevented because I literally can't tell whether changes I make are better or worse. There, the path to improvement is to improve metacognition.
As to the latter, my approach may be common sense, but it is not in fact common. Last I was interviewing developers, most people had never done a pair programming interview before. The default is talk-talk interviews. I'm sure some people did it the other way, of course: science frequently trails practice. But the Dunning-Kruger paper in specific is what convinced me that self-assessments were at best highly unreliable and at worst thoroughly misleading.
haven't read it thanks. are you saying the paper is about how one strives to improve? - because it's not titled that. the title is pretty clear. so is the abstract.
i think we agree that self assessments aren't worth much.
i think our disagreement stems from me applying it from a neutral ability perspective and you applying it to your own self assessment, where you don't question your basic competency.
> haven't read it thanks. are you saying the paper is about how one strives to improve? - because it's not titled that. the title is pretty clear. so is the abstract.
I am not. You didn't see how one could usefully apply their work. I'm explaining how I have applied it. You asked "Why in the first instance do you give your negative assessment credence, and have doubts about your positive assessment in the second?" When I say, "It's about how one strives to improve," I am answering your question.
> i think our disagreement stems from me applying it from a neutral ability perspective and you applying it to your own self assessment, where you don't question your basic competency.
Not at all. If for X I am a person of low ability, then I am likely to perceive myself as doing pretty well at X because I cannot see the flaws in my work. So if I'm at a state where my work looks good to me, the thing to do is, as I said, to improve my discernment until I can see flaws in my work. Then I can work to improve, increasing my competency.
This strategy works no matter what you relative competence is. And note that relative competence is distinct from absolute competence. Even people in the top quartile of a given sample can keep improving, and I believe Glass's quote helps explain why: their taste is still ahead of their skill, which if they keep practicing yields increased absolute competence.
I initially stated that you are just doing self improvement and dunning kruger isn't informing it.
you haven't said anything that disagrees with what i said, neither does the paper. at best you are applying half of it, the positive side. cherry picking?
if you're competent and you assess your work and it appears to be bad, then dunning-kruger says that it is not as bad as you think.
if you're incompetent and you assess your work and it appears to be good, then dunning-kruger says you don't have the competence to make that assessment and it's probably not good work.
if you had somebody competent in to assess your work instead of you, then that would probably be a more accurate application of it.
sure self improvement works, but self assessment doesn't really and that's the main take away.
I'll try one more time to get my point across. I agree with this:
> if you're incompetent and you assess your work and it appears to be good, then dunning-kruger says you don't have the competence to make that assessment and it's probably not good work.
Definitely. Ergo, if my work looks good to me, I should not assume that it is. Instead, I should look to improve my ability to evaluate the work. (What D+K call metacognition.) One way to do that, of course, is to have other people evaluate my work and tell me what they see. But there are plenty of others.
> if you're competent and you assess your work and it appears to be bad, then dunning-kruger says that it is not as bad as you think.
Exactly. Therefore, I should not be discouraged, but instead focus on upping my game to the level of my taste.
your overly positive outlook means you are either deliberately ignoring or unable to see the other side of it, and so there's no point in our continuing this conversation.
you've stated your position over and over, i've reworded mine, no progress will be made.
> Exactly. Therefore, I should not be discouraged, but instead focus on upping my game to the level of my taste.
it doesn't mean you should do anything. there's no need for this self improvement. the self improvement tactics are not informed by dunning kruger. they are not an application of dunning kruger.
in the absence of dunning kruger if you run with a self improvement mentality the input, action and outcome are the exact same. your behaviour is not informed by it. if you weren't self improving before dunning kruger, then you might be on to something - but i'm certain that wasn't the case.
in fact you are in danger of spending too much time on the improvement, whereas dunning kruger could be telling you, as you are competent "it's good enough".
I suggest going with either "there is no point in continuing" or "I will repeat again my possibly mistaken notion about why I think you're wrong." Because doing both makes you look like somebody who cares more about "winning" than having an actual discussion where people learn things. That's certainly the feeling I'm leaving with.
http://psych.colorado.edu/~vanboven/teaching/p7536_heurbias/...
Incompetent people think they're doing fine because they can't tell the difference between good and bad work. Highly competent people, readily seeing their own flaws, think they don't do as well as they are doing. (At least until you have them grade a bunch of other work, at which point they are better calibrated.) Competent people are competent because their metacognitive ability, their ability to tell good work from bad, drives them forward.
So basically, if I already can see the problems in my work, then my metacognitive ability is sufficient to drive improvement. When I improve enough that my work looks fine to me, further improvement is prevented because I literally can't tell whether changes I make are better or worse. There, the path to improvement is to improve metacognition.
As to the latter, my approach may be common sense, but it is not in fact common. Last I was interviewing developers, most people had never done a pair programming interview before. The default is talk-talk interviews. I'm sure some people did it the other way, of course: science frequently trails practice. But the Dunning-Kruger paper in specific is what convinced me that self-assessments were at best highly unreliable and at worst thoroughly misleading.