When looking at the origins of affective polarization, political communication scholars have frequently pointed to social media. In this article, we theorize that the relationship between social media use and affective polarization depends on the ways in which social media are used. Based on two-wave panel data collected during a national election campaign, our findings suggest that only active political uses of social media (i.e., sharing, posting, or commenting) foster affective polarization; in contrast, passive uses (i.e., informing oneself) do not. Looking at reciprocal relationships, we found that affective polarization did not significantly predict active or passive political social media use over time. Overall, our findings support the argument that social media are neither unconditionally detrimental nor beneficial for society and democracy.