Metacognition and Its Malcontents

Mon Jun 21 2010

For a long time I assumed that most people, if not aware of their own limitations, at least entertained the idea that they had limitations. It turns out that this isn’t true.

Years ago, I ran the tech support department of a software company that made compilers for the Macintosh. I needed people who knew how to program in C and Pascal, and who knew, or could learn, about the Macintosh API. The problem was that most of the people who qualified wanted to be Macintosh programmers, not tech support people.

We worked with headhunters back then (this was long before the web, and e-mail meant CompuServe or AppleLink), and I tried to explain to them that the people they sent didn’t have to be programmers, but they had to have had some programming experience. “You know,” I said, “liberal arts majors who took a couple of Computer Science classes and were really good at it.”

Most of the people I interviewed were open about their interest in the Macintosh, frank about their lack of ability in programming, and enthusiastic about learning more. I managed to find some outstanding people most of whom have made a difference in the software community. I remember all of the people I hired, but only one of the people I didn’t hire.

He was young, fresh out of college, and ready to work. I asked him to tell me about his experience programming. I expected him to tell me about the course in BASIC that he took, or how he was totally in love with his Commodore 64 even though he had studied anthropology at school. Instead, he said “Oh I’ve programmed a lot. I’ve programmed MacWrite, and I’ve programmed Excel, and I’ve programmed Word. I’m really good at it. I pick it up real quick.”

As gently as I could, I tried to explain to him that what he was doing wasn’t “programming,” but he had trouble understanding what I was saying. He’d been using Macs since they came out. He knew all about them. What the heck was my problem. If I’d just let him at the application we made, he’d pick it in no time.

I think about this guy every time I read about Dunning-Kruger effect. Briefly, the Dunning-Kruger effect says that people who are not good at something tend to overestimate their abilities at that thing, and people who are very good at something tend to underestimate their abilities.

Dunning-Kruger came up recently in the form of one of Errol Morris’s long essays for the New York Times. The Anosognosic’s Dilemma: Something’s Wrong but You’ll Never Know What It Is. Like all of Morris’s pieces, it’s a dense, gooey, and thoroughly delicious essay. I’m looking forward to reading the next installments.

I worry a lot about the Dunning-Kruger effect.

There’s the story people tell about competitive universities. When I heard the story, it was about MIT: It’s the big freshman orientation speech. The president of the college is up on the lectern congratulating the incoming class. “Every single one of you,” he says, “was an academic star at your high school. A lot of you were valedictorians. More than a handful of you got perfect scores on the SAT. All of you were at the top of your class.” The kids bask in their brilliance. “Now,” the president continues, “half of you are going to have to get used to being the bottom 50% of this class.”

All through high school I had done pretty well. I was in the upper 3.7% of my class. (I did the math.) By the middle of my junior year in college, it was pretty clear to me that I was much closer to the lower 3.7% of my class. (I didn’t do the math.) But even then, I still thought of myself as being pretty smart. I was, after all, at a top tier second echelon school-- barely hanging on, but hanging nonetheless.

I worry about the Dunning-Kruger effect because I see it all the time. I know people who are great artists, great musicians, great writers, great designers, great programmers. I don’t know whether they know that they’re good or not. I know that they consistently produce good work. I also know people who think they’re good at what they do but are not. And I can’t understand how they can’t tell they’re not good.

The thing I don’t know about Dunning-Kruger is whether being aware of the Dunning-Kruger effect makes you immune to it. Maybe thinking you’re immune to the Dunning-Kruger effect is a sign you’re in the grips of it.

According to Dunning-Kruger, if you’re really bad at something, you don’t know that you’re bad at it, so you think you’re good at it. So are there things that I think I’m good at that I’m actually pretty bad at?

I know I’m not good at sports. There’s no delusion there. I’d be surprised to find out that I’m anywhere out of the “abysmally bad” zone.

I’m pretty good at writing. I don’t think I’m very good. Does that mean that I’m better than I think? I used to think I was really good at it, but I read some of my earlier stuff that I thought was really good, and it wasn’t. So I’ll just leave it at pretty good.

I’d say that I’m a fair to mediocre programmer. I get things done, but not elegantly. In this respect programming is like art: I know what good looks like; I just can’t do it.

But what about the things I think I’m good that I’m not even qualified to know what good looks like. What if instead of knowing a little about a lot of things, I actually know next to nothing about a lot things I know a little bit about?