Science / Medicine : Sophisticated observers can be led astray by irrelevant similarities. In fact, the ability to reason well by analogy might be a rare talent. : Analogies : ILLUMINATING OR ILLUSIVE?
Learning how electricity works can be a challenge. One helpful trick is to use an analogy: electricity flows through wires in much the same way that water flows through pipes. Batteries and switches can be thought of as pumps and valves. A good analogy allows people to see that a new problem resembles an old one that has already been solved.
Cognitive psychologists, however, are finding that using analogies, or at least using them well, is surprisingly difficult. Experiments reveal that good analogies can be elusive and bad ones seductive.
Cornell University psychologist Thomas Gilovich has shown how poor analogies can influence decisions. In one study in the Journal of Personality and Social Psychology, Gilovich asked professional sportswriters to evaluate fictitious college football players on the basis of short biographies. Players were rated higher when they were described as coming from the same hometown as a famous professional player.
This result shows that even sophisticated observers can be swayed by an irrelevant similarity. The fact that a college player and a professional have a common birthplace does not mean they are likely to be analogous in other areas, such as football skill.
Gilovich, who is also in charge of graduate admissions for the social/personality psychology program, thinks this kind of shallow analogizing is a part of everyday life. In college admissions, he says, a certain applicant might “remind you of someone who was a disaster or someone who was a terrific star.” Once the analogy is made, it will bias the candidate’s rating even if the analogy was based on irrelevancies such as accent or hair style.
No one is immune to unsound reasoning. Recent studies have shown that even world leaders have been misguided by weak analogies. Deborah Larson, a UCLA political scientist, has studied Harry Truman’s relationship with Joseph Stalin. At their August, 1945, Pottsdam meeting, Truman remarked to an aide, “Stalin is as near like Tom Pendergast as any man I know.” Pendergast was a Missouri machine boss who helped get Truman elected to the Senate.
For some superficial reason--maybe the way both Stalin and Pendergast would pound their fists on a table--Truman saw the two men as analogous. According to Larson, Truman concluded that, like Pendergast, Stalin was a man one could deal with, a man of his word. “It led Truman to believe,” says Larson, “that Stalin would hold free elections in Eastern Europe.” U.S. foreign policy in this case was shaped in part by a president’s idiosyncratic and shallow analogy.
While slipshod analogies are all too readily used, another line of research is showing that good analogies can be almost impossible to spot. University of Michigan psychologist Mary Gick and her colleague Keith Holyoak, now at UCLA, asked people to solve the following problem:
Suppose you are a doctor treating a patient with an inoperable stomach tumor. Although there is a kind of ray that will destroy the tumor, if it is intense enough to do the job, it will also destroy the healthy surrounding tissue. At lower intensities, the ray is harmless to all tissue, including the tumor. How can the patient be saved?
Before seeing this radiation problem, some subjects read about a general who wished to capture a dictator’s fortress. Though the general had enough troops to win, all roads leading to the fortress were mined. The mines would be detonated by the weight of a large column of soldiers, but a small platoon could pass safely. The general devised a simple plan. He divided his troops into small groups and sent them to the heads of the roads that led to the fortress. At a prearranged signal, the troops began to march. The army converged on the stronghold and vanquished the dictator.
The doctor’s quandary in the radiation problem is analogous to the problem solved by the general. Once this is seen, the general’s solution can be recast in medical terms: A number of low-intensity rays could be aimed at the tumor from different directions. These beams would damage no tissue except the tumor, where their effect would add up.
When given a hint to use the general’s solution, most people were able to solve the radiation problem. But surprisingly, when no hint was given, only one person in five used the analogy to solve the problem.
Outside the psychology laboratory--where people people are not sensitized by being part of an experiment--analogies may be even harder to spot than that. Two Temple University psychologists found that virtually no one notices the analogy if unaware that the two problems are part of the same experiment.
So what does it take to get people to use a good analogy spontaneously? Clobbering them over the head with it seems to do the trick, as Holyoak and Richard Catambrone of the Georgia Institute of Technology have recently shown.
Subjects in their experiment first read two analogous problems and answered questions about how the solutions were similar. Next, subjects solved a third analogous problem using what they had learned from the first two. Finally, one week later, a fourth analogous problem was given to the subjects, who were unaware they were in the same experiment. Under these conditions, a large majority of people were able to solve the problem.
One reason a good analogy can be so hard to discover, says Holyoak, is that people rely on superficial details more than logical structure.
To support this notion, Holyoak and University of Michigan psychologist Kyunghee Koh asked people to read a new problem and its solution: The broken filament of a fragile glass light bulb is mended by several weak lasers aimed from different directions. Afterward, people tried to solve the radiation problem.
In contrast to the earlier experiment featuring the general’s problem, most people who read about the light bulb were able to make the analogical leap and solve the radiation problem without being told to use the light bulb problem. The reason, says Holyoak, is that the two problems are more similar at a superficial level. Both involve kinds of rays and a delicate substance--glass or tissue--that must be protected.
It would be comforting to think we could do better than Holyoak’s and Gilovich’s subjects, or, for that matter, Harry Truman in his appraisal of Stalin. One approach that might help, says political scientist Larson, is to pay closer attention to the causes of events. For example, she argues that it is inadequate merely to note that Nicaragua is analogous to Vietnam from the U.S. point of view. To avoid errors, we must go one level deeper. “We lost Vietnam for certain reasons,” Larson says. “Do those reasons hold in Nicaragua?”
Another political scientist, Yuen Foong Khong of Harvard University, is more pessimistic. “We are beholden,” he says, “to the things we are familiar with.” Khong, who argues that policy makers of the ‘60s were misled by a bad analogy between Vietnam and Korea, says that one pitfall that might be avoidable is getting stuck with the first analogy that happens to come to mind. To this end, Khong suggests, “you might want to let a hundred analogies bloom. Take all contesting analogies seriously and see what happens.”
Still, there is no foolproof recipe for sound reasoning. In fact, the ability to reason well by analogy might be a rare talent. William James, one of the founders of psychology, might have had it right when he wrote a century ago: “A native talent for perceiving analogies . . . (is) the leading fact in geniuses of every order.”
