Miriam Ruiz
random thoughts on technology and life











{April 07, 2008}   Dunning-Kruger effect

People who have little knowledge tend to think that they know more than they do, while others who have much more knowledge tend to think that they know less. That’s what Justin Kruger and David Dunning (both of Cornell University) demonstrated in a series of experiments they carried out.

Incompetent individuals tend to overestimate their own level of skill, fail to recognize genuine skill in others and fail to recognize the extremity of their inadequacy. If they can be trained to substantially improve their own skill level, these individuals can recognize and acknowledge their own previous lack of skill.

In one of their investigations, they discovered that 98% of university professors believed they were above the average, while that is obviously impossible statistically. Bright students, far above the others, considered themselves below their real skills, standard students saw themselves as above the average, while real bad ones were fully convinced of being among the best. In fact, the worse the person was, the most their conviction was that they were right. Even more, the most incompetent ones were incapable of realising the superiority of others, and they tended to think that the answers of the tests were the wrong ones, and not themselves. “Ignorance more frequently begets confidence than does knowledge”, quoting Charles Darwin.

Most of the people react to this study with a smile, thinking about the incompetent people they know and how well this study describes them. It should be noted that everyone is incompetent about many things, and this effect affects all of us.

How can a person realize that they’re wrong? The lesson that comes out of the study is that it’s really hard to find out, according to Dunning. His recommendation is not to trust just one’s own thoughts, but to ask for other’s opinions, especially before taking important decisions. Nobody should ever stop trying to improve and to learn, because it’s really difficult to know when to do it.



Dennis Krøger says:

Couldn’t another plausible reason for the high amount of “I rule” sucky students, be that those who realized their incompetence quit a long time ago?

If you study at a level much higher than your abilities, you probably already have trouble facing reality.



Edward "the pedant" Allcutt says:

“98% of university professors believed they were above the average, while that is obviously impossible statistically.”

Oh really?

On my arbitrary scale of whatever-it-is-we’re measuring from 0 – 1

98 professors get 1
2 professors get 0

Average (arithmetic mean): 0.98

% professors above average: 98%

Reproducing result with less skewed numbers is left as an exercise for the reader :P



Felipe Sateler says:

I read that paper a while ago, and the conclusion didn’t convince me. What I saw was more “Everyone estimates their ability in the same range, independent of actual ability.” (Not sure if that makes any sense in english ;) . This is because in the paper I saw, all estimations were around the 60-70 percentile range, in all subjects.



Octavio Alvarez says:

The result of self-estimation is directly related to the actualy knowledge, but with the willingness to learn about a particular subject, which does directly relate with the knowledge of the subject, as people usually climb the learning curve faster than others. Thus, There is an indirect relation between self-estimation and actualy knowledge.

People who are really interested in learning (and does learn), quickly realize that the more they know on the subject, the more unanswered doubts and questions they have and the more people they know with deeper knowledge of that same subject, realizing their lack of knowledge on that given subject.



Andy Price says:

The safe strategy seems to be to always assume that you might be wrong. This makes you check your facts, read up on subjects, and generally enhance your knowledge of a subject before you start to write/speak about it. Outwardly it shows humility and depth of intelligence. I believe it was Plato who famously claimed that he knew nothing, despite being one of the great ancient philosophers ( http://plato-dialogues.org/plato.htm ). Inspiring.



Theodotos Andreou says:

As for me, all I know is that I know nothing.

– Socrates –



Miry says:

Okay Edward, mathematically you might be right, but you know as well as me that your situation is not a realistic one. The results are more likely to be distributed in the shape of a Gaussian Normal distribution, as long as enough data are collected. Thus, it can be mathematically possible, but it is statistically impossible (with a high level of certainty).

Anyway, thanks for your remark.



Russell Coker says:

To avoid the risk of being one of those “too stupid to know they are stupid” is to look at the objective results of formal tests. If you do an IQ test and get a result significantly above average then it’s an indication that you probably aren’t stupid (in regard to mathematical, logical, and pattern-recognition tasks).

For skills that have highly subjective criteria (such as the ability to tell funny jokes – something that has been researched in regard to this issue) it’s more difficult – but every comedy club has an “open mic” night that allows every amateur to test their skill.

One of the harder areas seems to be “computer science” (of which most practitioners don’t know anything about the basics of science such as reproducable results, falsifiable hypothesis, etc). The web site http://thedailywtf.com/ has many examples of projects that were regarded as successful even though most people who read your blog would regard them as failures.

Although it could be that many professions could be regarded in the same way if you look at the real practice rather than the theory of the work.



are you one of the best? The Dunning-Kruger effect. « io41 says:

[...] Am I competent enough to pull off learning enough C and enough about the kernel & glibc to be able to hack them to do what I want? Miriam Ruiz pointed out and formulated exactly what my worries are. [...]



Leave a Reply


about

This is a personal webpage that belongs to Miriam Ruiz.
If you want to contact her, you can do at:
webmistress(at)miriamruiz(dot)es.

pages
categories
archive
twitter
calendar
May 2019
M T W T F S S
« Nov    
 12345
6789101112
13141516171819
20212223242526
2728293031  
credits
WikiLeaks

La Lista de Sinde