Do trait empathy measures predict how people experience empathy in daily life? Despite considerable research on empathy, we know surprisingly little about how trait measures relate to real-world empathic experiences. In this preregistered analysis of 7,343 experience sampling surveys from a near-representative sample of 246 U.S. adults, we map the connections between validated trait empathy measures and state experiences of empathy. Each component of state empathy-including emotion sharing, perspective taking, and compassion-was significantly predicted by theoretically relevant trait measures. However, trait empathy explained limited variance in daily experiences overall, ranging from just 3% for emotion sharing to 15% for perceived empathic efficacy. Adding emotional valence as a predictor improved model fit and variance explained for most state experiences, highlighting the crucial role of context. Our findings validate trait empathy measures while revealing their limitations in predicting real-world experiences.