I don't understand their conclusion. They claim that the effect is a positive bias, but if that were the case, you'd expect the two lines to move roughly in parallel. Instead, they overlap, so that the bottom group is positively biased and the top group is negatively biased.
There is a suggestion that that's caused by "regression to the mean", but that's doesn't match my understanding of regression to the mean. "Regression to the mean" implies some kind of outlier. And the outlier here is a consistent effect where the least knowledgeable quartile was the most overconfident. That is, the Dunning-Kruger effect.
This isn't my area of expertise so perhaps I'm getting this wrong. And yes, I'm aware of the irony.
Surely it's going to far to conclude from this that Dunning-Kruger isn't real. Seems to me the strongest conclusion we can draw from this is that there are at least two possible explanations for the result. How do we determine which explanation is correct?
2 comments
[ 4.1 ms ] story [ 18.1 ms ] threadThere is a suggestion that that's caused by "regression to the mean", but that's doesn't match my understanding of regression to the mean. "Regression to the mean" implies some kind of outlier. And the outlier here is a consistent effect where the least knowledgeable quartile was the most overconfident. That is, the Dunning-Kruger effect.
This isn't my area of expertise so perhaps I'm getting this wrong. And yes, I'm aware of the irony.