Some of our more recent trolls have reminded me of an article, Unskilled and Unaware of It: How Difficulties in Recognizing One’s Own Incompetence Lead to Inflated Self-Assessments, by Justin Kruger and David Dunning of Cornell University. Briefly, Kruger and Dunning demonstrated that college students who scored in the lowest quartile on several tests grossly overestimated their own abilities compared to everyone else’s, probably because they did not know enough to know that they did not know. Oddly, students in the highest quartile slightly underestimated their own abilities.
Figure 1 is a graph redrawn from the paper by Kruger and Dunning. It plots the students’ predicted scores on a particular test on the x-axis versus their actual scores on the y-axis. On average, 66 % of the students thought they were better than average at performing that test; in reality, only 50 % of the students could have been better than average. More interestingly, the students in the lowest quartile estimated their abilities at around the 60th percentile, whereas in fact they should have got the 12.5th percentile. By contrast, the students in the 4th quartile estimated their abilities at around the 75th percentile, whereas in fact they should have got the 87.5th percentile. Kruger and Dunning provide 3 more graphs that pertain to other tests, but they all look roughly the same. There is more, but I won’t bother you with it; let it suffice to say that the least competent students vastly overestimated their own abilities. Remind you of anyone?
Figure 1. Students’ actual scores on a certain test versus their predicted scores. The weaker students substantially overestimate their abilities compared to everyone else’s, whereas the stronger students slightly underestimate theirs. [After Kruger and Dunning’s Figure 1.]
Interestingly, the people who actually are above average underestimate their scores. So, those above average underestimate their abilities and those below average overestimate their.
And perceptions can be telling. The bottom 25% thinks that it is almost as competent at a variety of skills as the top 25% (60th percentile vs. 75th percentile). Yet in reality it is 12.5 vs. 87.5. The top 25% really can do things better.
So the bottom 25% can complain about an ‘elite that thinks it really is better’ because the bottom really does ‘think’ its abilities are almost as good as the top. Thus those ‘elites’ are putting on airs.
Unfortunately for the bottom, the guys at the top really do score better, their abilities really are greater and the difference in abilities really exists.
The Dunning-Kruger effect has been well described. Incompetent people do not recognize their inabilities. Competent people underestimate theirs.
Remember, if someone tells you just how good they are at something, the are chances are they aren’t. At least not as smart as they think.
This sort of bias seems to be socially driven, as this great a dichotomy between perceived and actual abilities is greatest in America. In fact, in Asian cultures there seems to be a better matching of the two curves, although they actually seem to underestimate their abilities at all levels.
Only in America do the bottom percentiles not really perceive a need to improve, because they already see themselves as almost as good as everyone else. So they actually never learn to try and get better. If someone shows them data indicating that their perception is not correct, they simply ignore the data. Ignoring data that does not fit perceptions is a classic case of a Cargo Cult World.