richard_nisbett's picture
Theodore M. Newcomb Distinguished University Professor of Psychology, University of Michigan; Author, Thinking: A Memoir
Graceful SHA's

1. A university needs to replace its aging hospital. Cost estimates indicate that it would be equally expensive to remodel the old hospital vs. to demolish it and build a new one from scratch. The main argument offered by the proponents of the former is that the original hospital had been very expensive to build and it would be wasteful to simply demolish it. The main argument by the proponents of a new hospital is that a new hospital would inevitably be more modern than a remodeled one. Which seems wiser to you — remodel or build a new hospital?

2. David L., a high school senior, is choosing between two colleges, equal in prestige, cost and distance from home. David has friends at both colleges. Those at College A like it from both intellectual and personal standpoints. Those at College B are generally disappointed on both grounds. But David visits each college for a day and his impressions are quite different from those of his friends. He meets several students at College A who seem neither particularly interesting nor particularly pleasant, and a couple of professors give him the brushoff. He meets several bright and pleasant students at College B and two different professors take a personal interest in him. Which college do you think David should go to?

3. Which of the cards below should you turn over to answer to determine whether the following rule has been violated or not? "If there is a vowel on the front of the card then there is an odd number on the back."

"If there is a vowel on the front of the card then there is an odd number on the back."





Some considerations about each of these questions

Question 1: If you said that the university should remodel on the grounds that it had been expensive to build the old hospital you have fallen into the "sunk cost trap" SHA identified by economists. The money spent on the hospital is irrelevant — it's sunk — and has no bearing on the present choice. Amos Tversky and Daniel Kahneman pointed out that people's ability to avoid such traps might be helped by a couple of thought experiments like the following:

"Imagine that you have two tickets to tonight's NBA game in your city and that the arena is 40 miles away. But it's begun to snow and you've found out that your team's star has been injured and won't be playing. Should you go or just throw away the money and skip it?" To answer that question as an economist would, ask yourself the following question: Suppose you didn't have tickets to the game and a friend were to call you up and say that he has two tickets to tonight's game which he can't use and asks if you would like to have them. If the answer is "you've got to be kidding, it's snowing and the star isn't playing," then the answer is you shouldn't go. That answer shows you that the fact that you paid good money for the tickets you have is irrelevant — their cost is sunk and can't be retrieved by doing something you don't want to do anyway. Avoidance of sunk cost traps is a religion for economists, but I find that a single college course in economics actually does little to make people aware of the sunk cost trap. It turns out that exposure to a few basketball-type anecdotes does a lot.

Question 2: If you said that "David is not his friends; he should go to the place he likes," then the SHA of "the law of large numbers" has not been sufficiently salient to you. David has one day's worth of experiences at each; his friends have hundreds. Unless David thinks his friends have kinky tastes he should ignore his own impressions and go to College A. A single college course in statistics increases the likelihood of invoking LLN. Several courses in statistics make LLN considerations almost inevitable.

Question 3: If you said anything other than "turn over the U and turn over the 8," psychologists Wason and Johnson-Laird have shown that you would be in the company of 90% of Oxford students. Unfortunately, you — and they — are wrong. The SHA of the logic of the conditional has not guided your answer. "If P then Q is satisfied by showing that the P is associated with a Q and the not-Q is not associated with a P. A course in logic actually does nothing to make people better able to answer questions such as number 3. Indeed, a Ph.D. in philosophy does nothing to make people better able to apply the logic of the conditional to simple problems like Question 3 or meatier problems of the kind one encounters in everyday life.

Some SHAs apparently are "graceful" in that they are easily inserted into the cognitive toolbox. Others appear to be clunky and don't readily fit. If educators want to improve people's ability to think, they need to know which SHAs are graceful and teachable and which are clunky and hard to teach. An assumption of educators for centuries has been that formal logic improves thinking skills — meaning that it makes people more intelligent in their everyday lives. But this belief may be mistaken. (Bertrand Russell said, almost surely correctly, that the syllogisms studied by the monks of medieval Europe were as sterile as they were.) But it seems likely that many crucially important SHAs, undoubtedly including some which have been proposed by this year's Edge contributors, are readily taught. Few questions are more important for educators to study than to find out which SHAs are teachable and how they can be taught most easily.