Decoding the Dunning-Kruger Effect

Written by romymisra | Published 2018/08/30
Tech Story Tags: psychology | women-in-tech | recruiting | hiring | dunning-kruger-effect

TLDRvia the TL;DR App

How lack of skill leads to inflated self-assessment

Photo by dylan nolte on Unsplash

I have been a product manager for the last eight years. After talking to many product managers in different companies, it became evident how difficult it is to assess product talent. This led me on a path of conducting product interviews to help companies hire good product managers that fit their needs. At the same time, I went on a parallel mission to master the science of recruiting through all the existing research. The insights were really interesting and I thought it would be helpful to “decode” some of the research I’ve read — or, in other words, break down the highlights to spare you some time and a few thousand words.

The first paper I want to decode is called “Unskilled and Unaware of it — How difficulties in recognizing one’s own incompetence lead to inflated self assessments” by Dunning and Kruger. I got interested in this paper after learning of the Dunning-Kruger Effect, which you may have heard of before if the word “imposter syndrome” rings a bell.

Here is what the paper mainly revealed:

The very skills that make you good at your job also make you good at recognizing whether you are good at it. Therefore, when someone is not skilled at their job, they also lack the skill of awareness, which makes them unable to recognize that they are not good.

Here are the details of how the researchers went about the study, which was carried out in three phases exploring three different questions.

Phase 1:

Primary question: “Are incompetent people incapable of assessing their skills?”

How they went about it: The researchers tested three different groups with a variety of skill levels in a) humor b) logical reasoning c) grammar, and asked them to measure their own test performance. Data points were then recorded for each person.

Results and takeaways:

Results of Phase 1

People in the bottom quartile tend to evaluate themselves as above average — they overassessed their abilities by more than 50%. In fact, the bottom two quartiles overestimated their abilities and thought that they were better than they actually were.

The top 25% actually underestimated how good they performed in the test.

People who lack the knowledge to perform well are often unaware of this fact — the researchers attribute it to a lack of “metacognitive skill.”

Phase 2:

Primary question: Less skilled folks definitely overestimate their own skill and ability. Is their misjudgment limited to their own skill or do they also assess their peers’ skills incorrectly?

How they went about it: The same groups as above were called in and, this time, each person was given the tests of their peers and asked to grade it to assess how good their peer did at the test. After that, the candidates were asked to assess their own test scores again.

Results and takeaways:

  1. People in the bottom 25% were less able to gauge competence of others as people in the top 25%.
  2. Even after seeing their peers’ performances, people in the bottom 25% didn’t assess their own scores any differently, in some cases they tended to inflate their own.
  3. What was interesting was after seeing the performance of their peers, the top 25% actually assessed themselves differently and raised the scores on their own performance.

This is because of what the researchers call false-consensus effect. In the absence of data, the best people assumed that everyone’s performances were similar to their own which is why they underestimated themselves.

Phase 3:

Primary question: Does training less competent people in their core skills actually help their ability to assess their skills? As the former would imply, does increasing your ability to become good at something increase your metacognitive skill as well?

How they went about it: Researchers decided to train the cohorts to improve the skills. After training, they were asked to grade their own and their peers’ scores again.

Results and takeawys:

  1. Training the less skilled candidates in the bottom quartile actually improved their ability to assess themselves.
  2. Training them in their core skills also improves their metacognitive skills as well.

Final takeaways from the paper:

There is a paradox here, that to actually help incompetent people assess their and anyone else’s skills better, we need to train them to make them more competent.

One of the things which the author mentions is that the general absence of negative feedback makes it really hard for incompetent people to grow since they’re unaware that they’re incompetent in the first place. This issue is primarily faced by the bottom quartile of competency. Also this doesn’t apply to every domain and every use case — like, for example, if you stepped into a room of Spanish speakers without knowing Spanish, no one would have trouble identifying that, regardless of their skill level.

Decoding these insights for skills and recruiting

How does this affect hiring and recruiting in companies and the way we look at skills?:

  1. Hiring people in the bottom quartile of any skill has terrible consequences.

Unskilled employees won’t recognize that they need to improve, even after mixing with the best.

Take, for example, a programmer — If someone is not good at programming, apart from not doing the job well, they probably also lack the competence of being able to assess that they are doing a bad job. Worse-than-average programmers tend to think they are above average.

2. If we do hire people who are less skilled, we should train them and give feedback.

Training is the best way to break the cycle. Training will actually help them not develop their own skills but also their overall metacognition.

3. Most importantly, companies should select the highest skilled people to interview and recruit. If you select people who are not skilled, they won’t be able to assess talent.

4. The most skilled people in a company may tend to underestimate their performance (the consensus effect) unless they have worked with people who are less skilled.

5. Finally, recruiters should have the ability to recognize those who are self-aware of their skills, as many people with lower skill level and high levels of self awareness are likely to improve over time.

Do you feel like LinkedIn isn’t truly reflecting your skills? Or are you looking for the next opportunity and having trouble breaking through the noise? If the answer is yes and you would like to hear about exciting opportunities personalized to you, please sign up here.

If you are a company looking for a new way to hire, please sign up here.


Published by HackerNoon on 2018/08/30