Research suggests that the structure of trainee reactions is captured with as few as one or as many as eleven dimensions. It is commonly understood that reactions contain both affective and cognitive components. To date, however, training research focuses largely on affective reactions that range from pleasant to unpleasant (i.e., valence). Here, we expand and further refine the construct of affective trainee reactions by including reactions that are more and less activating versions of pleasantness (e.g., excitement and calm, respectively) and unpleasantness (e.g., stress and boredom, respectively). We develop and validate a new measure based on this model and argue that the structure of affective reactions has implications for better understanding learning and course reputation outcomes. Results from a short online training indicate that reactions were best explained by four factors: pleasant activation (e.g., excitement), pleasant deactivation (e.g., calm), unpleasant activation (e.g., tension), and unpleasant deactivation (e.g., boredom). In addition to these four factors, results also show a general activation factor underlying all items, which produces a bifactor measurement model. The relationships between these reactions and training outcomes suggest what is most beneficial for course reputation outcomes (i.e., pleasant activating reactions) may not benefit learning; what is most beneficial for learning (i.e., pleasant deactivating reactions) may benefit course reputation outcomes but slightly less so.