Comparing a risky choice in the field and across lab procedures

Controlled laboratory experiments have become a generally accepted method for studying economic behavior, but there are two issues regarding the reliability of such work. The first pertains to the ability to generalize experimental results outside the laboratory. The second pertains to the impact the payment procedure has on observed behavior. This paper adds empirical insight into both issues. Using data from the promotional campaign of a bank and a laboratory experiment that closely mimics the same decision, we find similar levels of risk taking controlling for gender and age. We also compare behavior on this same risky choice across three distinct experimental payoff procedures: a single salient choice as in the field, multiple responses for similar choices with one selected at random for payment, and a single salient choice that has only a small probability of being implemented. We find nearly identical behavior across these three payment procedures.