A Simple Agent Model of Dunning-Kruger

Natalie Wolchover at a site called Life’s Little Mysteries (which I will not be adding to my now-doomed Reader feed) reports (h/t Desdemona Despair) on an agent-based simulation of politics in a Dunning-Kruger world.

David Dunning, a psychologist at Cornell University, shows that incompetent people are inherently unable to judge the competence of other people, or the quality of those people’s ideas. For example, if people lack expertise on tax reform, it is very difficult for them to identify the candidates who are actual experts. They simply lack the mental tools needed to make meaningful judgments.

As a result, no amount of information or facts about political candidates can override the inherent inability of many voters to accurately evaluate them. On top of that, “very smart ideas are going to be hard for people to adopt, because most people don’t have the sophistication to recognize how good an idea is,” Dunning told Life’s Little Mysteries.

He and colleague Justin Kruger, formerly of Cornell and now of New York University, have demonstrated again and again that people are self-delusional when it comes to their own intellectual skills. Whether the researchers are testing people’s ability to rate the funniness of jokes, the correctness of grammar, or even their own performance in a game of chess, the duo has found  that people always assess their own performance as “above average” — even people who, when tested, actually perform at the very bottom of the pile. [Incompetent People Too Ignorant to Know It]

Mato Nagel, a sociologist in Germany, recently implemented Dunning and Kruger’s theories by computer-simulating a democratic election. In his mathematical model of the election, he assumed that voters’ own leadership skills were distributed on a bell curve — some were really good leaders, some, really bad, but most were mediocre — and that each voter was incapable of recognizing the leadership skills of a political candidate as being better than his or her own. When such an election was simulated, candidates whose leadership skills were only slightly better than average always won.

Nagel concluded that democracies rarely or never elect the best leaders. Their advantage over dictatorships or other forms of government is merely that they “effectively prevent lower-than-average candidates from becoming leaders.”

I am not actually thrilled by what I understand from that about Mato’s research – the result is pretty much built in to the way the question is asked. But it is an interesting question.

It also raises the question of whether science’s output is bound by the competence of the average scientist. This would argue for more elitism, not less.

Comments:

  1. Pingback: Another Week of GW News, March 17, 2013 – A Few Things Ill Considered


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>