The idea that test anxiety hurts performance is deeply ingrained in American culture and schools. However, researchers have found that it is actually worry about performance and anxiety--not bodily feelings of anxiety (emotionality)--that impairs performance. Drawing on this insight, anxiety reappraisal interventions encourage the view that anxiety can be neutral or even helpful. Initial evidence--largely from laboratory studies--suggests that these kinds of reappraisal interventions can improve student performance in mathematics. But can they do so in other domains and within the constraints of everyday classroom activities? If so, for whom and how? In an intervention study, we tested whether a minimal reappraisal message embedded in an email from course instructors could improve students' academic experience and performance in an introductory college course. The night before their first exam, students received an e-mail that either did or did not include a paragraph designed to lead them to interpret exam anxiety as beneficial or at least neutral. First-year students, who experience greater test anxiety and are less certain about how to perform well, benefited from the reappraisal message, showing decreased worry and increased performance on the exam the next day as well as increased performance in the course overall. Mediation analyses revealed that the effect on overall course performance for first-year students was partially mediated by reduced exam worry and enhanced performance on the first exam. The message did not affect the performance of upper year students.