# A Factoring (Possible) Effect of Dunning-Kruger.

1. Feb 21, 2017

### WWGD

Hi All, I am looking at some data regarding correlation between control measures and (perceived)errors at
a given position. The level of error seems , intuitively, suspiciously low. I suspect the low perceived error may
be at least partially explained by the Dunning Kruger effect:
https://en.wikipedia.org/wiki/Dunning–Kruger_effect

Does anyone know of anyway of factoring the effects of Dunning-Kruger to come up with a more realistic
expectation of error given control? We have a measure of Control from 0-5, same scale for error(both scales are linear; control goes from high to low, error goes from low to high: control near 5 is high, error near 5 is low) . Say a an input of control is 1 , a low measure, and the perceived error at that level is 4. Can I make a more accurate estimate of error by factoring out effects of Dunning-Kruger?
Thanks.

2. Feb 23, 2017

### Stephen Tashi

Do you have data about the "actual" error as well as the perceived error?

3. Feb 24, 2017

### Staff: Mentor

Without at least some comparison of actual and perceived error I don't see how you could do anything useful.

A more detailed description what exactly you have as data would be useful.

4. Feb 25, 2017

### WWGD

No, sorry, I was trying to estimate any possible error. I was wondering if there were ways to do this. I was wondering to what extent ( and if ) DK provided some means to do this. I have data about the perceived error and I was wondering if DK allowed to estimate the margin of under/evaluating or general mis-evaluation of error.

5. Feb 25, 2017

### Staff: Mentor

Not without a study how strong the effect is in your particular setup - and including such a study in your particular setup is probably the only reliable way to estimate that. I guess you cannot say anything about actual errors, just something about perceived errors.