Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The author states, as regards the interpretation of the Dunning-Kruger diagrams, that

  [i]n two of the four cases, there’s an obvious positive correlation between
  perceived skill and actual skill, which is the opposite of the pop-sci 
  conception of Dunning-Kruger.
In my corner of the universe, you don't get to cherry-pick which pieces of data (ie "what instances of two sets of random variables") you bestow the golden twig of correlation upon. If I'm not entirely mistaken, correlation is very much a global feature, not a measure of proximity of two points on a chart.

So, yes, Dunning-Kruger (as evinced from the diagrams sported here) indeed seems to make a weaker claim: that there's no correlation between “perceived ability” and “actual ability”. As such, this claim is as far from the "pop-sci conception" of Dunning-Kruger as it is from the author's.



The referenced graphs measure performance and perceived ability on 4 different tasks. You're right that ideally you'd pool this data to get to an overall correlation, but for the point the author makes, eyeballing it and taking a mental average does the trick, no?

Also, what corner of the universe are you from? Loess regression, hierarchical modeling, conditional analyses... methods for finding "non-global" correlations aplenty.


Yeah, the valuable conclusion to make from the Dunning-Kruger paper is that self assessment isn't worth a lot.


Err, no. There is clearly a pattern to the data, and drawing the conclusion that self-assessment and actual skill are uncorrelated is not what you should do (the simplified pattern being that unskilled individuals overestimate their skill, while skilled individuals underestimate theirs, or a regression to the average if you will).

At any rate, even if we don't take that into account self-assessment is worth plenty as demonstrated by many studies which manage to get coherent data from self-assessments. Sure, you should take them with a grain of salt, and you can expect biases, but no need to throw them out.


I apologize for not writing an essay carefully specifying what "worth a lot" means to me.


self assessment is worth plenty.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: