The Dunning-Kruger effect

Latest post in my blog on popular science:
The Dunning-Kruger effect


1 Like

Interesting, thanks. I’m generally familiar with the Dunning-Kruger effect, but wasn’t familiar with the details of the experiment you described. I’m not sure, though, that their conclusions necessarily follow from the data–they place a lot of weight (unavoidably, I’m sure) on the students’ self-assessment of their abilities, but that leaves me wondering how seriously the students took those questions. An alternative explanation for the results would be lack of serious consideration of the question, combined with perhaps a measure of humility on the part of the highly-capable students. The reason for my concern here is that everyone rated their ability about the same, which I don’t think is really explained by their conclusions.

With respect to your opening text, I hadn’t seen quite that phrasing before, though I’m well familiar with what are at least closely-related categories which I’ve seen referred to as levels of knowledge (generally with “that” replaced by “what”):

  • You don’t know what you don’t know. Not so much that you’re a fool, but that (with respect to the subject at hand) you don’t even know the different domains of knowledge. Lewis, I think, might analogize this to the schoolboy who’s been enthralled by Homer buckling down to learn Greek–I believe he used this analogy for a slightly different purpose in Screwtape.
  • You know what/that you don’t know. You’ve made progress. There’s still a great deal you don’t know, but you have a reasonably good idea of what it is.
  • You know what/that you know. You’ve learned, and you’re consciously aware of your knowledge.
  • You don’t know that you know. You’ve learned the material so completely that you aren’t consciously aware of it any more.

And, of course, leave it to the Germans to come up with a nice tongue-twister on the subject: Er, der nichts weiß, und weiß daß er nichts weiß, weiß mehr als der, der nichts weiß, und weiß nicht, daß er nichts weiß. (He who knows nothing, and knows that he knows nothing, knows more than he who knows nothing, and does not know he knows nothing).

But look at the second part of the experiment: it shows (at least it seems to show) that the best students were able to improve their estimation of their own capacity after they saw a few of their colleagues’ answers, while the worst students couldn’t. Doesn’t this show that there was a difference between those two groups, which would tend to explain why all of them thought that they were above average?