The reinforcing spirals model (RSM) has quickly become a popular way of thinking about the reciprocal dynamics of media use and effects, not only in political communication. One key assumption of the model is that both exposure to media content and individual attitudes vary over time-if there is no variation, there cannot be any covariation. In this paper, we discuss the assumptions of the RSM regarding intra-individual variation and inter-individual differences in media use and political attitudes. We present empirical results on the temporal stability of different phenomena of interest in political communication, such as media use, campaign involvement, or candidate evaluations, drawing on panel data from two national election studies in the USA and Germany. Our results suggest that, after accounting for imperfect measurement, media use and most political orientations and attitudes are very stable over the course of a campaign, making it rather unlikely to detect patterns corresponding to the RSM. We discuss the usefulness of the RSM and similar dynamic media effects models in light of our results.