BLACKSBURG, VA., FEB. 21, 2000 -- Do celebrity endorsements influence how people make decisions? Advertisers who pay big bucks for celebrity endorsements seem to think so. And now an innovative economics research project has demonstrated that most people will follow the example of a successful "celebrity."
Speaking at the American Association for the Advancement of Science (AAAS) 2000 Annual Meeting in Washington, D.C. on Monday, Feb. 21, Catherine C. Eckel, professor of economics at Virginia Tech, will explain how she and Rick K. Wilson, professor of political science at Rice University, tested whether people can be influenced to make economically beneficial decisions.
"In many economic situations, there are gains to cooperation; people can do better if they cooperate and work together than if they make decisions independently." says Eckel. "However, individuals tend not to collaborate, in part because they don't know what other people will do. People may interact on a small scale, among a few neighbors or friends, and base their idea about what is best to do on their limited experience. These groups then interact in a way that may have an eventual good result for the economy -- or not.
"The individual tendency is to be conservative -- to avoid risky activities, even when they can lead to a better outcome," she says.
Eckel and Wilson have introduced the concept of one person whose behavior individuals and groups of individuals watch. In their study, they show that, while people are influenced by neighbors or others whose actions they observe, the influence is stronger when the person they observe has high status, i.e., a celebrity.
The celebrity's success can get others to change their behavior -- to a point. "The celebrity -- or higher-status agent -- can get people to make a better choice, such as select a better product or investment, but rarely will the agent influence people to make a choice that is clearly inferior," says Eckel. In an experiment in Virginia Tech's Laboratory for the Study of Human Thought and Action, 144 subjects played a game where they chose among three strategies. One could give them 550 points -- considered a good equilibrium of the game -- but was somewhat risky; the second strategy would pay only 350 points -- considered a bad equilibrium -- but was less risky; The third strategy resulted in a potential 600 points, or higher if the player cheats, with very high risk -- considered a non-equilibrium.
The researchers established one player's higher status before the game began. When all of the subjects took a trivia quiz, it was announced that player C got all 15 questions right and completed the quiz in less time than anyone During the game, the subjects interacted anonymously, choosing among the three different strategies. In general subjects were tempted by the good equilibrium, but its riskiness meant that most chose the strategy that earned 350 points. When the researchers told the group about player C's strategy for achieving 550 points, most of the other players made the same choices as C, and had a better outcome on average.
Thus, having a commonly observed celebrity can help an economic system get to a good equilibrium, Eckel explains.
However, there are limits to a celebrity's influence, the researchers learned. "A commonly-observed agent cannot get subjects to choose a highly-risky strategy that does not lead to an equilibrium," says Eckel.
"Celebrity endorsement is valuable. It can get people to change what they do. Our work suggests that celebrity endorsement can get people to chose a better product, or one of two equally good products, but cannot influence people to make a foolish choice."
In general, economics research has neglected social status, Eckel says. "The influence of people of high status has not been studied." Much of her research addresses this issue.
The paper being presented at the AAAS conference (Social Interactions in Economics Experiments) is the first experimental economics paper to look at social status in learning, or its role in effecting the dynamics of a market system.
Eckel's and Wilson's research is sponsored by the National Science Foundation. It is one of five presentations selected for the Dynamics of Social Interactions program (Monday, Feb.21, 3 to 6 p.m., in the Omni Shoreham Hotel, Level One West, Governor's Room), organized by Barbara Boyle Torrey of the National Research Council, National Academy of Sciences, and Charles Manski of Northwestern University
To contact Dr. Eckel: email: eckelc@vt.edu or phone: +1 540-231-7707. Learn more about her work at www.econ.vt.edu/Eckel/vitae.htm