Projects and Publications
The Role of control in decision making
Much of the research conducted in the Georgia Decision Lab is devoted to uncovering the role that perceived control plays in decision making. If you could sink your entire life's savings into a lottery ticket with a 10% chance of winning millions of dollars, would you do it? Probably not. If you could sink the same savings into your own business, with a 10% chance of making millions? For many people, the answer is yes. At a more basic level, research in the Decision Lab has revealed that people are more willing to accept risk when betting on an event they have control over - like starting a business - than when betting on a random event like a lottery. Participants do this even when the probabilities and the payoffs are identical, suggesting that control itself plays an important role in decision making.
In the Georgia Gambling Task, where people face bets that are fair (have zero average value) if their confidence is appropriate, we have found that people accept bets on their own abilities more often as they become more confident, even though the value of those bets becomes less as confidence increases, whether considered as objective average value or subjective utility. This effect is termed paradoxical betting, and it leads people to lose large amounts of points in simulated computer tasks, and perhaps money and other things of value in real-world tasks. These findings are being analyzed both with knowledge-based tasks like answering trivia questions, and with more skill-based tasks like making golf putts.
Problem and Pathological Gambling
With the explosion of opportunities to bet in casinos, with friends or on the internet, the rates of problem gambling appear to be rising as well. Research in the Decision Lab, funded by the National Institute for Mental Health, is investigating the role that perceived control plays in problem gambling. For example, we know that most people engage in disadvantageous paradoxical betting when they perceive they have control. When they no longer have control, and the task is a random one, problem and pathological gamblers continue to accept disadvantageous risk, whereas non-problem gamblers reduce their risk taking. This suggests that the illusion of control may be a powerful contributing factor in problem gambling. Research conducted in the Decision Lab has been nominated for the 2006 Outstanding Research award of the National Council on Problem Gambling.
With funding from the National Institute for Mental Health, research in the Decision Lab is exploring the neural activity that underlies the processes that contribute to perceptions of control and decision making. For example, when people believe their probability of having answered a question correctly is high (in other words, when their confidence is high), the right dorsolateral frontal cortex is activated more than when confidence is low. These results are depicted at right. Also, non-problem gamblers show a marked difference in processing "oddball" stimuli with more activation in the parietal-temporal junction, which is associated with attention. This makes sense - people need to devote more attention to processing odd stimuli. But problem and pathological gamblers do not show this difference, suggesting that problem gambling has a significant basis in attentional factors.
Personality and Decision Making
We are examining how personality affects judgment and decision making. For example, research in the Decision Lab has looked at how overconfidence (believing you will succeed more often than you actually succeed) is related to the Big Five personality traits of extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience. We have also explored how performance on the Georgia Gambling Task is affected by the personality trait of narcissism.
We found that narcissists are more confident than others, without being any more accurate. This makes them more overconfident, which makes the bets they face in the GGT less favorable than those that others face. Furthermore, they accept those bets more often than others, leading to systematically lower point totals for narcissists.
THEORY OF MIND IN ADULT DECISION MAKING
The literature on recursive reasoning ("I think that you think that I think.") is pessimistic. Individuals have attributed little strategic reasoning to others, performing well against simulated non-strategic opponents, and learning slowly and incompletely against strategic opponents. In the GDL we replicated these results with a previously used game; but, with a game that was made competitive and simpler by having a fixed-sum structure, these results reversed. Against a strategic opponent, participants performed well, whereas against a non-strategic opponent they learned optimal responding slowly and incompletely. Thus, individuals may be less prone to underestimate their opponents than was previously held. We are currently pushing the envelope further to see whether individuals reason at even deeper levels of recursive reasoning.
GEORGIA DECISION LAB