UC Berkeley News
Berkeleyan

Berkeleyan

Xlab invites staff to join pool of experimental subjects - for cash
Payment, though 'contingent,' is guaranteed to all who participate

| 17 March 2005

If you got an e-mail message with the subject line "Make money playing computer games on campus!" you'd probably send it to the trash unread, along with the latest offer from someone you've never heard of hoping to sell you a Rolex for $27.95.

The deal is real, however - as is the opportunity to contribute to the advance of human knowledge while pocketing some extra change.

Since it opened for business last year, the Experimental Social Science Laboratory (better known as Xlab), part of the Institute of Business and Economic Research (IBER), has been coordinating computer-based psychological research for several schools and departments. With the blessing of the campus Committee for the Protection of Human Subjects, Xlab's administrators are now actively recruiting Berkeley staffers to serve as subjects in a variety of experiments - in exchange for financial consideration. This wrinkle may surprise those who have previously participated in unpaid social-science research projects with human subjects, because behavioral economists, unlike experimenters in many other social-science disciplines, regard compensation of subjects as the norm, not as an aberration.

"We give our subjects money so they'll take the experiment seriously," says Xlab's director, Teck Ho, professor of marketing at the Haas School of Business, where Xlab is housed. He cites an experiment he was involved with in China in which participants were paid twice their average monthly income for an hour of lab time. "They took their roles very seriously," Teck recalls drolly.

The amount a staff volunteer receives for participation in an Xlab experiment is "contingent," says IBER's deputy director, Bob Barde - that is, it varies according to a number of factors that include the decisions the participant makes, the decisions made by other subjects in the study, and that well-known income modifier, chance. "An earnings rate between $15 and $25 for one and a half hours of participation is common," reads the fine print in Xlab's online protocols, "although in many experiments subjects earn much more (or less)." The amount paid will never be zero, Barde says, and will always exceed the "show-up fee" that participants in a specific experiment receive automatically.

The games themselves, while not exactly Halo or Road Kill, are interesting in their own way . . . and require no cash outlays for pizza and energy drinks. You won't know in advance what type of game you'll wind up playing when you register as a staff subject - but it's likely to be one that will enlighten researchers about the ways in which decisions are made by individuals who find themselves operating in a context where the actions (or assumed actions) of others have a role in the outcome. Those could range from a "buyer's curse acquisition auction" or "ultimatum/dictator bargaining game" to a "traveler's dilemma game" or even a "signaling [aka 'beer/quiche'] game." (And not to worry if you're selected to participate in the latter, only to discover that the "menu of default parameter settings includes one with separation and another with two pooling equilibria, both of which are sequential but only one of which is 'inuitive' in the sense of Cho and Kreps." That level of understanding is required only of researchers, not of participants.)

The decision to expand the pool of experimental subjects to include staff stemmed from the perceived need to broaden the age range of those subjects. "There's a general problem with university-based experiments, whether in psychology or behavioral economics," says Barde. "It's the notion that a pool of students is a suitable proxy for people in general. It's not that students are different from real people - it's that there isn't much variation in a student pool. If you're using students to predict how the student population in general will behave, that's fine - but if you're using a pool of 18- to 21-year-olds to predict how the entire population might behave, that's pretty risky."

To register as a staff subject for future Xlab experiments, visit xlab.berkeley.edu and click on "UCB staff members are now eligible to sign up!" A $10 bonus will be paid to the first 100 staffers who do so.