3 ms·
It looks to me like the consensus is that this data artifact, while real, only accounts for part of the effect, so the Dunning Kruger effect is still a thing.
by EdwardCoffin 4y ago
It looks to me like the consensus is that this data artifact, while real, only accounts for part of the effect, so the Dunning Kruger effect is still a thing.
- naasking 4y agoDunning-Kruger is not just an observation but a purported explanation for those observations. That only qualifies as "a thing" if only that explanation can account for the effect independently of confounding factors. The Dunning-Kruger effect has not reached this threshold, not only because data artifacts account for a large portion of it, but also because of the other confounding factors described there that could account for the remainder. See also this review for a more detailed analysis: https://replicationindex.com/2020/09/13/the-dunning-kruger-effect-explained/ https://replicationindex.com/2020/09/13/the-dunning-kruger-e...
- EdwardCoffin 4y agoI'm unconvinced. There are other studies that show that there's more to it than confounding factors. For example: > not only is the Dunning–Kruger effect not merely a statistical artefact at the group level, it also cannot be explained solely by Bayesian shrinkage in the rational estimations of individual participants. [1] Note too that we've veered way off from the original assertion that DK had failed replication, through the fallback claim that it was an artifact of analysis itself then finally to the claim that though data artifacts are part of it, there could be confounding factors. [1] https://matanmazor.github.io/files/papers/mazor2021dunning.pdf https://matanmazor.github.io/files/papers/mazor2021dunning.p...