Testosterone and risk: change may bemore important than the level

Abstract

We tested the relationship between women's testosterone (T) concentration levels and their risk-preferences, their time discount rates, and their social behavior with dictator game. As for the risk-preferences, we utilized both choice task and matching task. We artificially decreased subjects' T levels with a dose of pure glucose and compared their economic response with the controls. We found no statistical relationship between T concentration and risk-preferences, and between T concentration and time discount rate. However, short time spontaneous changes in T levels had a positive relationship with matching task, but not with choice task. This may reflect that matching task is similar to the risk situation whereas choice task to uncertainty.