Student performance and loss aversion*

We match data on performance in a multiple-choice examination with risk preferences from classroom experiment. Students who are more loss averse leave questions unanswered and perform worse the exam when an incorrect answer is penalized compared no answer. Thus, aversion parameters extracted lottery choices controlled experiment have predictive power field environment of decision-making under uncertainty. Furthermore, degree appears to be persistent over time, as was conducted three months prior exam. Important differences across genders partly explained by aversion.