Positive Psychology Interventions (PPIs) are specifically aimed at improving well-being via increases in positive emotions, positive cognitions, or positive behaviors. Despite the proliferation of PPIs, few studies have specifically tested whether increased positive emotion mediates effects of the interventions on health and well-being. In this keynote, I present results from several randomized trials of a theoretically-based, multi-component PPI. I will report when positive emotion is a significant mediator of effects and when it is not, and discuss possible explanations for discrepancies across studies including variations in measurement and operationalization of positive emotion, different delivery mode (e.g., online vs in person) of the intervention, and differences in stress experienced in the different samples (e.g., HIV diagnosis, dementia caregiving, depression). I will conclude with recommendations for emotion assessment to advance the science of PPIs.