I don't understand where the first graph graph came from. The peak of mount stupid doesn't come from Dunning-Kruger at all (the first peak is not in their data set), and your graph is one of four they included in their paper (self-assessment of grammar). While it's true that at the very upper end there's between 10% and 25% underestimation of skill, it's interesting there's almost no underestimation in logic and reasoning:

https://graphpaperdiaries.files.wordpress.com/2017/08/logicgraph.png

But there's a significant overestimation at the low end.

Edit: By first graph, I'm referring to the one in the post you responded to.

I don't understand where the first graph graph came from. The peak of mount stupid doesn't come from Dunning-Kruger at all (the first peak is not in their data set), and your graph is one of four they included in their paper (self-assessment of grammar). While it's true that at the very upper end there's between 10% and 25% underestimation of skill, it's interesting there's almost no underestimation in logic and reasoning:

https://graphpaperdiaries.files.wordpress.com/2017/08/logicgraph.png

But there's a significant overestimation at the low end.

Edit: By fist graph, I'm referring to the one in the post you responded to.

I don't understand where the first graph graph came from. The peak of mount stupid doesn't come from Dunning-Kruger at all (the first peak is not in their data set), and your graph is one of four they included in their paper (self-assessment of grammar). While it's true that at the very upper end there's between 10% and 25% underestimation of skill, it's interesting there's almost no underestimation in logic and reasoning:

https://graphpaperdiaries.files.wordpress.com/2017/08/logicgraph.png

But there's a significant overestimation at the low end.