Researchers at Carnegie Mellon University have taken the lab to the people. Outside of some Pittsburgh bars and restaurants, exiting patrons become prospective test subjects, lured into a mobile laboratory with the promise of snacks or a few bucks.
Places like airports and other transportation hubs used to be choice locations for conducting behavioral surveys. That is, until Sept. 11, 2001, after which heightened security made such sites mostly inaccessible to researchers.
"That knocked out my biggest source of easy-to-collect data," said George Loewenstein, a professor of economics and psychology at CMU.
Enter the Data Truck. Dreamed up by Loewenstein after Sept. 11, the mobile lab is 36 feet long, 13 feet high and equipped with eight laptop computers used to survey the decision-making behavior of subjects in a range of conditions. The truck also boasts a flat-screen TV, a refrigerator, a table and chairs, internet access and air conditioning. In operation since last summer, the truck has given Loewenstein and his colleagues inroads to a whole population of prospective test subjects.
From inebriated bar-hoppers to people exiting the movie theater, pedestrians are invited into the Data Truck to complete a survey or to participate in a brief experiment. In exchange, they are rewarded with popsicles, candy bars, money or other small gifts, said Jessica Wisdom '04.5, a doctoral student at CMU studying behavioral economics and health psychology.
"It's hard to get people to believe that you're going to pay them," Wisdom said. But the word "Research," emblazoned on the side of the truck in black capital letters, "makes it seem more legit," she said.
In one of many experiments conducted this summer by researchers at CMU's Center for Behavioral Decision Research, Wisdom parked the Data Truck outside of a Subway restaurant, asking passersby to flip through different menus and make choices about what food they wanted. Wisdom observed different choices based on the location of the items in the menu and whether or not calorie contents were listed.
Both Wisdom and Tamar Krishnamurti, another CMU graduate student of Loewenstein's, said the Data Truck has opened new doors in their research. Previously, participants had to register for experiments on the CBDR Web site.
"We were getting people who already knew about (the experiments)," said Wisdom, who added that the Data Truck has made their studies much more accessible to interested subjects.
"The difference is incredible," Krishnamurti said. "Often you're getting student participants ... (but with the Data Truck), you can access a totally different demographic."
Access to a larger participant pool is a problem for more than just CMU researchers. Economics professors Louis Putterman and Pedro Dal Bo, who run the Brown University Social Science Experimental Laboratory, say that they focus on student participants in their research out of "convenience."
"Getting American adults with regular jobs to show up in a university laboratory is difficult," said Putterman. "You need large amounts of money, and even then it's hard to reach them."
"We would like a more diverse pool," Putterman said. But, he adds, "Brown subjects are very smart and cooperative. Our incentive to go far and wide is weakened by the fact that the Brown student subject pool is very pleasant to work with."
Through mailbox slips, advertisements and monetary compensation of up to $25 per student, Putterman and Dal Bo have assiduously courted student interest for over two years now. Getting funding for experiments and obtaining an experimental laboratory, rather than recruiting additional students, is BUSSEL's primary focus, Putterman said.
Although an ideal experiment works from a perfectly random participant pool, "no sample is truly random," Loewenstein said. "Even if you randomly dial the telephone, you're getting a random sample of people who have telephone numbers and answer the telephone when you call. ... In a sample, all data sets are biased."
Though Loewenstein said he and his colleagues have always been able to find test subjects "some way or another," one of the Data Truck's unique qualities is that it allows researchers to target specific pools, such as retirees or citizens of low-income neighborhoods.
"The truck is useful for collecting data from people you might have trouble in getting to come to a lab," he said.
"I wouldn't say that it's fundamentally changed my work in any way," Loewenstein said, but he believes that as a method for making data collection easier, the Data Truck is "a wave of the future."
"I don't think that many people know about it," said Loewenstein, "I think that when word gets around, every major university will want one."
