To a biologist, "micro" means bacterium-sized. To an economist, "micro" means people-sized. And this year's Bank of Sweden Prize in Economic Sciences, given in honor of Alfred Nobel, goes to two researchers who gave the field of microeconomics--the study of individuals' economic behavior--new tools to help draw conclusions from imperfect data.
As any scientist knows, statistical investigations are prone to error; any inadvertent biases in choosing the sample or systematic errors, and the whole exercise might be doomed. The situation is even dicier for economists who take statistical samples of complex, semirational objects like human beings.
James Heckman, an economist at the University of Chicago, shares this year's prize for his methods of dealing with selection biases. He came up with two ways of handling those biases and then used them to analyze things such as how wages affect the behavior of married women in the labor market. "Economics is a field where you're solving real problems," says Heckman from Brazil, where he and his students are studying education and economics. "Being able to tackle real problems has always been an attraction for me."
Daniel McFadden of the University of California, Berkeley, tackled a different conundrum: how to quantify discrete choices rather than continuous ones. "Before McFadden did his work, economists were concerned with buying amounts--how many oranges a consumer buys, et cetera," says Charles Manski, an economist at Northwestern University. "But many important choices are discrete: Do you go to college or not? Do you buy an auto or not?" McFadden looked at discrete choices in terms of probabilities; given certain conditions, a consumer might have, say, a 20% chance of taking a car to work, a 40% chance of taking a bus, and a 40% chance of taking the train. This method, inspired by similar approaches used by psychologists, gave McFadden a way to handle discrete problems--and these methods led to his helping design the Bay Area Rapid Transit system.
Marketers, sociologists, political scientists, and others are all indebted to McFadden's methods, says Steve Lerman of the Massachusetts Institute of Technology: "The number of applications, in both marketing and economic analysis, must be in the thousands--maybe even higher."