Please Stop Mentioning the Dunning-Kruger Effect
Announcements and apologies
I haven't written in a while, but that's mainly because I'm in the middle of switching jobs and it's pretty over-whelming right now. I'll tell you more about that later! I have couple of texts in the pipeline, but I need more excess energy to finish them.
I'm writing this newsletter, because this subject gets my blood boiling instantly, and, well, yeah, the blood is boiling right now. I have no real excuses as to why this is related to software development other than the misinterpretation of D-K seems quite popular in these circles too.
Please Stop Mentioning the Dunning-Kruger Effect
An absurd, wild, exaggerated, insane, and incorrect interpretation of The Dunning-Kruger effect has become an Internet meme and it has no basis in the reality. But unlike most memes, people seem to actually believe in this meme. I think it's just best we stop referring to it alltogether, because it has been butchered so heavily that it is beyond saving.
There are also actual, valid criticisms of the original research paper, which suggest that even the mild interpretation that the original paper gives is probably not a reflection of the reality, but my rage does not stem from that at all. Even if the original result was 100% accurate and reflects the reality, it is still irrelevant to this discussion. The real problem is that people have come to believe this exaggerated interpretation of the D-K effect, which has no scientific support whatsoever.
Please stop googling for D-K effect too
So, I just searched for the effect in Google Images Search and all of the results point to a graph of the exaggerated interpretation. I don't want to be a part of the problem and include one of these pictures here, because I guess in the end it would be "another data point" for the support of this absurd and incorrect interpretation of the research result.
But if you're not sure if you're seeing the graphs of the wildly incorrect interpretations let me offer you one from the original paper and you can compare to that:
Now, I've had multiple discussion about this discrepancy in the Internet (because nowadays it is my hobby to correct the misinterpretations of the D-K effect) and these discussions make me even more depressed and confused.
When I point out that the original paper has graphs of their results and that they look nothing like this exaggerated and incorrect graph of the results, people say to me that "of course I know it's an exaggeration, but the point is to better illustrate the effect!"
This... I just can't ... even ...
I have trouble finding words to illustrate how frustrated I'm about this response. Maybe I should draw a graph of that? It fucking makes me question my own sanity.
So first you have this scientific research and you read about it and you are like "yeah, I know this phenomenom", because we all have met a person who thinks he is a genius but in fact is clueless. Then you tap into this interpretation of it and then draw a graph of that and then refer to the "D-K effect". Which is just plainly false.
The. D-K. Does. Not. Support. Your. Interpretation. OK?
All in all, my suggestion is that you should not illustrate an effect that does not exist.
In a serious discussion, at least.
Other people have written better takes on this
Dan Luu writes about this and also other misinterpretations of scientific results.
(PS. This newsletter was not proof-read or reviewed by anyone, which probably results in incoherence and ramblingness, but I just had to get this out of my system, once again. Sorry. Next draft I'll send to someone to comment on.)