# Dunning-Kruger Effect ### Pop science vs Actual science https://twitter.com/BernoulliDefect/status/1470461760297738240?t=XvIiMhm3vFlxZkBCvfKmMA&s=19 FYI, the dunning Kruger effect is vastly misunderstood. The pop science understanding of it is something like this graph. I.e, low performers have huge false confidence compared to the humble mid and high performers ![[20211214_122404.jpg]] But the results of the study are much more boring: predictions increasing when competence increased, and all with a strong bias towards the 50th percentile (which is quite understandable). ![[20211214_122410.jpg]] This *could* be a result of the classic conclusion of the study, or it could simply be that the test was quite a noisy measure - some people ended up in the top/bottom 25% because they were just lucky/unlucky and their guess looks like a underestimate/overestimate As the same pattern of results has been found with random data (maximum noise), it’s easy to understand that the Dunning Kruger effect is at best mild, or at worst a completely fake statistics artefact