When surveys rely on repeat participants, this raises the possibility that survey participation may affect future responses, perhaps by prompting information acquisition between survey waves. We show that these “learning-through-survey” effects are large for household inflation expectations. Repeat survey participants generally have lower inflation expectations and uncertainty, particularly if their initial uncertainty was high. Consequently, repeat participants may be more informed about or attentive to inflation. This has important implications: for example, inflation expectations of new participants are more influenced by oil prices, and estimates of the elasticity of intertemporal substitution are lower for new participants. (JEL C83, D84, E31, E37, E58)