[Cavernio]

Originally Posted by kaiten123

I think stargroup is mostly right about Dunning-Kruger.

Caverino, I think you're reading of the Dunning-Kruger effect is off. I don't think anyone (not even Dunning and Kruger themselves) suggested it was as extreme as

since an exact quote from the paper: "We do not mean to imply that people are always unaware of their incompetence". makes it clear that you're misunderstanding something.
Rather, its simply that because the skills needed to be good at something are mostly the same as the skills needed to judge that thing, people who are bad at something don't have the skills to accurately judge their abilities.

You are misreading what I've said then, or I've not explained myself properly. Kruger Dunning has not coined this idea. I have my own idea with regards to what they say, and it differs somewhat from how they present the idea.

Originally Posted by kaiten123

They also did sub-experiments specifically to separate the effect from some other effects. So your comments there (specifically, you claim they only showed the above average effect, and that there was nothing to show it only effeted those with less skill, etc.) betray that you never read the paper or even a decent summary of it. In fact, the popular video you mentioned probably mentioned it as well if its the video I'm thinking of so it sounds like you didn't even finish that much.

Just because you don't understand my position or that my position differs from the authors or yours does not mean I'm ignorant or that I've not actually read their papers.

Originally Posted by kaiten123

While everyone initially rated themselves a little above average, there was another test where the people were allowed to see a few other people test answers before rating themselves.
So if people with less skill fall victim to the Dunning-Kruger effect, their ratings will be unaffected since they don't have the skills to judge the tests. At the same time, the people with more skill should be able to judge other people's tests accurately against their own to see more accurately how skilled they are.
In this test, the people who did poorly still rated themselves above average, but people that were more skilled rated themselves more accurately...[This] first sub-experiment rules out all effets that would affect both skilled and unskilled participants since only the unskilled failed to accurately judge,

Right, I alluded to this in my very first post regarding Dunning Kruger, and I've discussed this since then as well. This part of the experiment does not, I believe, show what people say it shows...the general ability gauge where you fit into a group, when you are directly and specifically shown that you are worse than other people in that group, is not measuring the ability that the test is measuring. It's measuring a more general ability to perceive where oneself fits in compared to others, and my theory is that it is largely irrelevant to whatever skill is being measured. The refusal to acknowledge that you are bad at something is a psychological phenomenon.
If you take my interpretation of what they tested, this part of the experiment furthermore does not separate the skilled and unskilled group as you say it does, because it itself seems like it would be a phenomenon of refusal to admit that you're actually that bad at something. Take a person who performed mid-range or higher on this test and compare them to someone who scored near the very bottom; the exact same mechanism could work in both people except that because the first person did better on the test, they still do not have to admit that they are actually bad at it. Whereas the person who is bad at it, in order to be accurate enough, would have to.
This explanation of the results of that study, I feel, makes a lot more sense than the KD is.

Originally Posted by kaiten123

There was also yet another test, where former participants were invited back for another test after minimal training and were all able to better predict their performance after the training.
and the second rules out all effects due purely to the individuals involved since merely increasing their skills made them as good at judging their performance as the people who had high skills to begin with.

This does support the KD effect as far as I can tell, I don't remember it being address in the study of theirs that I read. Was it a statistically significant difference?

Originally Posted by kaiten123

At the very least, this is perfectly consistent with the DK effect, and not easily explained otherwise. (there have been some notable attempts to pin everything on regression to the mean, task difficulty, and a few others, but they've all been debunked since like 2008)

Have you found something that addresses what I suggest though? I haven't, not that I spent a whole lot of time looking. Nor would I say that those studies were 'debunked'. They just also have issues with them, like KD studies do too. That's generally how good research goes, people critique the shit out of it so that eventually we'll hopefully take into account all possibilities that explain the results.
Nor, again, have I found any Kruger Dunning study that doesn't not clump people into quartiles. This is my most serious concern with their studies because ultimately, as long as they only analyze effects when they do this, I will not ever say that their studies are good support what the internet now calls the KD effect.
Like, it would be really interesting to see how those few individuals who are horrible at something and who say that they're horrible at something with relatively good accuracy, differ from those other individuals who are also bad at the given skill but who rate themselves much more favorably. Because I'm 100% positive that there exist some people in these studies who know and report that they are bad at something. I'd wager that some sort of general test that measures personality traits or emotional states, like arrogance, self-confidence, self-esteem, etc, would find a significant difference b/w these 2 groups of poorly scoring people.
Whatever factor separates these poor judges of their poor score from the good judges of their poor scores, we can then measure that factor in everyone else in the group and see if scores of that factor differ significantly for everyone else who are in the top 3 quartiles. Like, if 90% of low scorers have a lot of self-confidence, would 90% of the rest of the participants also have high self-confidence.

There's so much that I think needs to be teased apart. And my theory, again, is that the KD effect is probably only significant at the lowest end of ability on a scale.

Originally Posted by kaiten123

You also seem to think DK claims to be more than it actually claims to be. They go out of their way in the paper to make it clear that the DK effect is not the only effect in play so questions like "If someone is just unable to perceive that they're bad at something, why would they therefore think they are good or decent at it at it?" are adresed trivially by other well known effects such as the above average effect which you seem to already be aware of.

I've said it twice now what I'm complaining about, and I'm not complaining at what Kruger and Dunning say in their paper(s). I'm complaining about people who perceive that the Kruger Dunning studies show strong evidence in support of what is now known as the KD effect.