Fortunately, the suit isn't working properly today, so Ariely shuts the machine down and I'm left to imagine the rest: His subjects (some of whom are paid $100 for their trouble) are asked to call out numbers, from 0 to 100, at certain intervals to signal their level of discomfort. Then he raises the water temperature, creating pleasurable sensations, to which the subject also assigns numerical values, until the temperature climbs to its flesh-scalding limit -- about 126 degrees Fahrenheit -- and again the victim whimpers for the pain to end.

Welcome to the world of behavioral economics, where such medieval experiments are dreamed up to produce fresh insights into how people behave with money. Ariely is one of a new generation of researchers shaking up classical economics by challenging some of its bedrock principles -- that people act rationally and in their own best interest, and that they know what goods and services are worth to them. In other words, principles on which billion-dollar business decisions are based every day.

In this case, Ariely wants to know how people evaluate their most painful and pleasurable experiences in retrospect. If conventional economic wisdom is correct, consumers emerging from a bad experience -- a wretched restaurant dinner, say -- might average out their most painful moments to come up with an overall assessment. Or they might focus on the single most painful moment, or how long it lasted.

But Ariely came up with surprising results when he asked people to evaluate how painful it was to wear the suit, which he uses as a metaphor for the consumer experience. In their responses, his subjects focused on two things: how intense the pain was at the end of the experiment, and whether the pain increased or decreased during the second half.

The first point makes intuitive sense: It's why fireworks displays always save the biggest blast for last. But Ariely's second point is startling. Pain that increases is considered worse than pain that is constant, even if that increasing pain is at a lower level than the constant pain.

The implications for business are vast: An Internet service provider that delivers consistently lousy service, for example, would be looked on more fondly than one that once was excellent but has recently gone downhill -- even if that deteriorating service is still superior to that of the ISP that has always been bad.

Of course, you'd never be so irrational. Then again, if you've ever wondered why you just shelled out $4 for a cup of Starbucks (SBUX) coffee -- or why Nike (NKE) doesn't pay us to advertise its logo on our clothes -- you might conclude that the behaviorists are on to something. And if people are irrational, wouldn't it be intriguing to map all the myriad quirks, impulses, biases, and boneheaded maneuvers that constitute human economic behavior and find consistent patterns? That, behaviorists say, would represent the Holy Grail for companies trying to figure out what their customers want and how much they're willing to pay for it. It's an intellectual challenge that Ariely and other bright young academics can't resist.

Ariely's study of human behavior started with his own pain, the kind most of us could never imagine. Born in New York and raised in Israel, Ariely was 18 years old and enrolled in a government community service program when a flare exploded next to him. He suffered burns on 70 percent of his body and spent the next two and a half years in the hospital. During an agonizing daily ritual -- having his bandages removed to clean the wounds -- he debated with the nurses whether it was less painful to tease them off slowly or just rip them off.

On his release from the hospital, Ariely began studying the psychology of pain, wondering why some people have different tolerance levels than others. After graduating from Tel Aviv University in 1991, Ariely went on to study cognitive psychology at the University of North Carolina, where he earned a master's degree and a Ph.D. Business school eventually seemed more appealing, and in 1998, Ariely earned another Ph.D. (in marketing) from Duke University.

Behavioral economics, by then, had come a long way from its fringe origins in the '70s. Spearheaded by Princeton's Daniel Kahneman, founding father of what became known as "behavioral finance," it was all the rage on Wall Street in the '90s. Behaviorists including Richard Thaler of the University of Chicago (who discusses the psychology behind retirement savings in "How to Make Americans Save More," page 147) argued that markets were driven by nonrational factors such as hubris and passing fads. Some went so far as to claim that they could take advantage of predictable flaws in investors' thinking to earn higher-than-average returns on stocks.

When the behaviorists proved no better than anyone else at investing, traditionalists took rhetorical revenge, arguing that the newcomers had proven only what everyone already knew -- that folks sure act strange sometimes. "After a while, people started saying, 'So what?'" says Eugene Fama, a finance professor at the University of Chicago and an outspoken critic of the trend. "You can't just poke fun at a theory. You have to come up with one of your own."

And so Ariely, now 35, and others have begun to expand on the work of Kahneman and Thaler. During the last two years, Ariely has done 20 studies, coauthored with Carnegie Mellon's George Loewenstein and MIT's Drazen Prelec, that challenge another bedrock economic principle: that people can determine how much something is worth to them. In one experiment, Ariely sold a variety of products -- keyboards, chocolate, wine -- to his subjects but began by asking if they would buy the products for a price determined by converting the last two digits of their Social Security numbers to dollars (for example, 10 became $10). Then he asked the maximum price each would be willing to pay. The results were striking: Subjects with high Social Security numbers were willing to pay nearly twice as much for the products, even though they were reminded that the suggested price was essentially random. Those with numbers in the top half of the range paid on average $19.95 for a 1998 Cotes du Rhone Jaboulet "Parallele 45," while those in the bottom half paid only $11.62 for the same bottle.

In another study, Ariely and his colleagues asked one group of subjects if they would listen to him read 10 minutes of Walt Whitman's Leaves of Grass if he paid them $10. He asked another group if they would pay him $10 for the same experience. Then, with each group, he negotiated prices for readings lasting one minute and three minutes. Again, the initial question wielded tremendous influence. Those asked to pay knew they should pay more for 10 minutes of poetry than for three minutes, but none suggested that they should be paid. The reverse was also true: Those who were paid agreed to get $3 for three minutes but never insisted on paying for it. It didn't matter with either group whether they even enjoyed poetry or Walt Whitman.

Ariely draws two conclusions from these studies: First, people are terribly suggestible and have no innate sense of what a product or service is worth. Instead, they take cues from their environment -- even something as random as a Social Security number. Second, once people act on such cues, they tend to repeat that behavior without questioning it. Ariely calls this "self-herding." Basically, we're worse than sheep.

To prove his point, Ariely brings up Starbucks. Not long ago you might have considered a cup of coffee to be worth less than $1. Now you happily fork over $4 for a white-chocolate mocha. Why? Because Starbucks added some new flavors, opened some trendy cafes, and told you that coffee is worth much more. Picking up on various external cues -- you've heard that the coffee is good, say, or that the brand is associated with status -- you accept the new price. And soon you're hooked: You begin seeing your own past behavior as proof that a cup of coffee is truly worth $4 to you.

Old-school economists like Fama reject this scenario as simplistic. If the price of coffee is so malleable, he says, why not charge $1 million per cup? Starbucks can charge more because it offers higher-grade coffee, upscale furniture, sophisticated sound systems -- even new wireless networks in hundreds of stores. Competition is so fierce, Fama explains, that if Starbucks's prices didn't reflect the cost of delivering the product, somebody would cut in and offer it at a lower price. "Markets," he says, "protect people against that kind of irrationality."

Focusing on the supply side of the equation misses the point, Ariely counters. A Lexus doesn't cost twice as much to make as a Toyota, but people are willing to pay twice as much to buy one. Consumers rely on external cues -- the brand image, the prices of comparable cars, and so on -- to assess what a product is worth. Eventually that creates inefficient markets that reflect the irrationality of customers.

All of this is worthy of debate, old-school economists say, but it doesn't change their views. If behaviorists are so smart, they ask, how come they're not rich? Ariely is not a corporate consultant and has no blueprint for how to make millions from his research. In fact, talking to him can feel like a late-night bull session at graduate school: lots of fun, but what's the practical value?

And yet anybody who has gotten caught up in a bidding war on eBay realizes how capricious we all can be with our money. Ariely may have to torture a few more people before some smart company discovers precisely how to turn his surprising research into profits, but stay tuned: When his subjects start offering to pay him for the privilege of wearing the pain-and-pleasure suit, we'll know he's really on to something.