I’ve been passing out links to the first Dunning-Kruger article since 1999. Now, I have a link to an article that tested Dunning-Kruger on college students who were taking the LSAT – the admissions test for wannabe lawyers. I like studies that use standardized tests as the test for the hypothesis.
The first article I read on the Dunning-Kruger Effect had the long title “Unskilled and Unaware of It: How Difficulties in Recognizing Incompetence Lead to Inflated Self-Assessments”.
Testing the hypothesis against the LSAT scores yielded this graph:
So the next thing was to sort out the statistical noise:
The results showed an astonishing level of inaccuracy in expected scores. For example, on the grammar test, the lowest performers who got either zero or one question right (out of twenty questions) expected on average to have scores of approximately twelve right (!). By comparison, high performers who got all twenty questions correct underestimated their performance by a much more modest amount, expecting on average to get approximately sixteen right.”
As the article goes through the mathematica that support the reality of the Dunning-Kruger effect, the article ends with:
“But perhaps the hardest lesson to learn about DK—often stressed by David Dunning himself—is that this problem is not limited to some other group of people safely off in the distance. Depending on the area of expertise, the Dunning-Kruger effect applies to us all. If ignorant enough, under many circumstances, we will fail to recognize how ignorant we are. Ideally, the message of the DK effect is both troubling and humbling.”
Both articles are available online, at the links, and well worth reading.
Trego's Mountain Ear 