Abstract
32 subjects were presented 4 items involving decisions either to take a certain amount of money or select a probability level for winning a larger amount. In a 2 × 2 design, subjects either made these decisions hypothetically or actually played out their choices for money. For half of each group, the items described a fixed payoff regardless of the probability level selected, while for the other half there were variable payoffs where the amount won increased as the probability of success decreased. Data indicated that subjects were more conservative when playing for real money with a fixed payoff but not when there was a variable payoff. These results suggest that the presence of conservatism in real risk-taking situations may be a function of the type of payoff.