Dunning on Dunning-Kruger

4 minute read

Dumb people think they’re smart. Smart people think they’re dumb. This is a simplistic overview of the Dunning-Kruger effect. David Dunning explains more in this interview, including important nuances that people miss.

even though your belief about the way the world is just seems so compelling or so self-evident, it doesn’t mean that it really is true

Just because we want to believe in something doesn’t make it true. Most people seek comfort by sticking with their beliefs and attacking anything showing otherwise. See flat earth deniers, anti-vaxxers, white chocolate fans [1]. If you strongly believe in something, you should be seeking out arguments that disprove it, rather than further adding on to your confirmation bias. When you do find a credible opposing view, steel manning the opposing side’s arguments is a better approach.

The first rule of the Dunning-Kruger club is you don’t know you’re a member of the Dunning-Kruger club. People miss that.

I first read of the Dunnning-Kruger effect many years ago, still fall to it often, and only realise on hindsight. This applies to both areas I’m unskilled at and areas I like to think I’m skilled at [2].

Well, it teaches us both about the limits and the genius of human understanding. Which is, we can take some idea and spin a complete and compelling story around it that is coherent, is plausible, makes a lot of sense, is interesting — and it doesn’t necessarily mean that it’s right. So it shows you how good we are at spinning stories.

A major point of the bestseller Sapiens by Yuval Noah Harari is that humans succeeded through their ability to tell stories.

people who think not in terms of certainties but in terms of probabilities tend to do much better in forecasting and anticipating what is going to happen in the world than people who think in certainties

He references Philip Tetlock and Philip’s work on superforecasting here, which is interesting to read more about. I would like to believe this is true, though Taleb has major issues with Philip’s work. Will leave you to decide whether probabilistic thinking matters, though I think the point on tracking your predictions and scoring them makes sense.

That’s an interesting question, because people seem to be uncomfortable about saying, “I don’t know.” That’s one thing we’ve never been able to get people to do. I have to admit that over 30 years of research, I often think the correct answer to the question I’m asking you [in a survey] is, “I don’t know.” And people give me an answer [other than “I don’t know”]. How do you get people to say, “I don’t know”? I don’t know.

I try to avoid this, with debatable success. I find that people seem to lose confidence in you when you say you don’t know, which is sad. This applies to both situations in which I say “I don’t know but can find out” and “I don’t know but that’s un-knowable”. If anyone has advice here I’m interested!

It’s up to you whether you think this attitude is worth adopting. I personally think too many people pretend they know about subjects they don’t for fear of seeming dumb. On a related note, you start to notice how many errors there are when news covers a topic you actually know about. This is related to the Gell-Mann amnesia effect though I’m unsure how scientifically valid this has been proven.

Here’s the key: The consequential decisions tend to be the ones that we don’t come across all that often. Like, what houses do we buy? What people do we marry? What kids do we have? And so consequential decisions tend to be the ones we don’t have experience with. They’re exactly where there’s stuff we don’t know, and that’s exactly those types of situations where we should be seeking outside counsel.

I like this point. There are many large but infrequent decisions to be made over the course of a lifetime. This is where getting advice from multiple, diverse sources will be helpful. Note the emphasis on diversity, since it doesn’t help if everyone you talk to within your family adds on to your confirmation bias [3].

An active social life, active social bonds, in many different ways tends to be something that’s healthy for people. Social bonds can also be informationally healthy as well. So that’s more on a top, more abstract level, if you will. That is, don’t try to do it yourself. Doing it yourself is when you get into trouble.

Again an emphasis on seeking out others. Feedback from qualified people is helpful, but that depends on how you select the people to get feedback from.

Update 2019-02-10: The entire effect has been called into question in this paper here, which is an interesting development. Most interesting quote from the abstract below:

Our data show that peoples’ self-assessments of competence, in general, reflect a genuine competence that they can demonstrate. That finding contradicts the current consensus about the nature of self-assessment. Our results further confirm that experts are more proficient in self-assessing their abilities than novices and that women, in general, self-assess more accurately than men.

Footnotes:

  1. Before I’m attacked, I actually do like white chocolate, in addition to all other chocolate.
  2. Remains to be seen whether I’m skilled at anything at all though
  3. It is a delicate balance between people who all say the same thing, and people who say irrelevant things