Tag Archives: Overconfidence

Overconfidence is a Social Signaling Bias:

We test three different theories about observed relative overconfidence. The first theory notes that simple statistical comparisons are compatible with a Bayesian model of updating from a common prior and truthful statements. … Data on 1,016 individuals’ relative ability judgments about two cognitive tests rejects the Bayesian model. The second theory suggests that self-image concerns asymmetrically affect the choice to get new information about one’s abilities. … We test an important specific prediction of these models: individuals with a higher belief will be less likely to search for further information about their skill. … Our data also reject this prediction. The third theory is that overconfidence is induced by the desire to send positive signals to others about one’s own skill; this suggests either a bias in judgment, strategic lying, or both. We provide evidence that personality traits strongly affect relative ability judgments in a pattern that is consistent with this third theory. Our results together suggest that overconfidence in statements is most likely to be induced by social concerns than by either of the other two factors.

This of course means that you should expect to risk looking less impressive to others if you correct for your overconfidence.

In a 3-point elicitation procedure, an expert is asked for a lower limit, upper limit, and best guess, the two limits creating an interval of some assigned confidence level (e.g., 80%). In our 4-step interval elicitation procedure, experts were also asked for a realistic lower limit, upper limit, and best guess, but no confidence level was assigned; the fourth step was to rate their anticipated confidence in the interval produced. … We … found average overconfidence of 11.9%, 95% CI [3.5, 20.3] (a hit rate of 68.1% for 80% intervals)—a substantial decrease in overconfidence compared with previous studies.

More here.

In Moore’s experiment, volunteers were given cash for correctly guessing the weight of people from their photographs. In each of the eight rounds of the study, the guessers bought advice from one of four other volunteers. The guessers could see in advance how confident each of these advisers was (see table), but not which weights they had opted for.

From the start, the more confident advisers found more buyers for their advice, and this caused the advisers to give answers that were more and more precise as the game progressed. This escalation in precision disappeared when guessers simply had to choose whether or not to buy the advice of a single adviser. In the later rounds, guessers tended to avoid advisers who had been wrong previously, but this effect was more than outweighed by the bias towards confidence. …

Moore said that following the advice of the most confident person often makes sense, as there is evidence that precision and expertise do tend to go hand in hand. For example, people give a narrower range of answers when asked about subjects with which they are more familiar.

There are times, however, when this link breaks down. With complex but politicised subjects such as global warming, for example, scientific experts who stress uncertainties lose out to activists or lobbyists with a more emphatic message.

That is from New Scientist back in June.  So the key hard question is: when does confidence credibly signal expertise, and when it is just empty cheap talk?  Hat tip to Toby Ord.

Gladwell:

The psychologist Ellen Langer once had subjects engage in a betting game against either a self-assured, well-dressed opponent or a shy and badly dressed opponent (in Langer’s delightful phrasing, the “dapper” or the “schnook” condition), and she found that her subjects bet far more aggressively when they played against the schnook. They looked at their awkward opponent and thought, I’m better than he is. Yet the game was pure chance: all the players did was draw cards at random from a deck, and see who had the high hand. This is called the “illusion of control”: confidence spills over from areas where it may be warranted (“I’m savvier than that schnook”) to areas where it isn’t warranted at all (“and that means I’m going to draw higher cards”).

I was once a shy badly dressed person who didn’t understand why people kept underestimating him.  Now I know.