Why Do We Make Mistakes?

 Knowing Your Limitations

If you've ever seen the movie Magnum Force, you no doubt remember Clint Eastwood's role as the tough-guy detective Harry Callahan. At the end of the movie Eastwood has the bad guy right  where he wants him—at the end of his barrel—when he utters one of his more famous lines: “A man's got to know his limitations.”

Good advice, that.

When it comes to judging just how overconfident someone is, social scientists have devised a term for knowing one's limitations: it's called “calibration.” Calibration measures the difference between actual and perceived abilities. If you're as good as you think you are, then you are said to be well calibrated. If you are not as good as you think you are, then you are said to be poorly calibrated.

Most of us tend to be poorly calibrated— we're not as good as we think we are.

Most of us tend to be poorly calibrated, even (and perhaps especially) when it comes to important skills, like those we need to do our jobs. The U.S. Army discovered this years ago when it asked soldiers at Fort Benning, Georgia, the military version of the layman's question “How good a shot are you?” Most of the soldiers, naturally enough, thought they were pretty good shots. They predicted they would score well on the Army's annual qualification with M16 rifles. Then the Army asked the same soldiers to step onto a firing range. After the shooting was done, the soldiers’ scores were tallied and compared with their predictions. The results were not good: 75 percent of the soldiers predicted they would hit more targets than they did. In addition, more than one out of every four soldiers shot so poorly that they failed to qualify. The soldiers “generally overestimated their actual performance,” concluded a report of the results, “and were biased heavily toward predicting success.”

Oddly enough, the Army noted, the predictions by one group of soldiers proved to be on target. That group? The poorest shots. To be sure, this was a small group. Of the 153 soldiers participating in the annual qualification, only 5 predicted that they would fail to qualify. But those 5 were dead-on. Three of them did, in fact, fail to qualify; the other 2 just barely passed.

“Those who predicted failure,” the researchers noted drily, “were quite accurate.”

Researchers have found equally poor calibration for people engaged in other tasks, regardless of factors like income, intelligence, and education. Not long after the Army tested the soldiers at Fort Benning, for instance, researchers conducted a similar test of students at the University of Wisconsin, Madison. Instead of having the students shoot at targets, though, the researchers asked them to do the academic equivalent: read one paragraph of text and then rate themselves in terms of their confidence in their ability to draw the correct inferences from it. Then the students were tested on what they had read. As you might expect, the students, like the soldiers, didn't do so hot.

“Calibration was strikingly poor,” said the researchers. “Our readers were unable to distinguish between what was understood and what was not understood.” Go, Badgers.

Why We Make Mistakes

Joseph T Hallinan