Concerns about video-based political persuasion are widespread in both popular and academic circles. These concerns are predicated on the assumption that video is more persuasive than text — that is, that “seeing is believing.” To date, however, this assumption remains largely untested in the context of politics. Here we provide such a test. We begin by drawing a theoretical distinction between two dimensions of belief for which there could be an effect of video versus text: (i) one’s belief that a presented event actually occurred, and (ii) the extent to which one’s attitudes are changed. To assess these two possibilities, we ran a high-powered survey experiment (20,428 observations from n = 4808 Americans, quota-matched to the general population on age, gender, ethnicity, and geographic region). Participants were presented with a selection of both political and non-political stories; for each story, they were randomly assigned to view either a short video or an annotated transcript of the video. While we find that video significantly increases belief that the presented events occurred, the effect is very small (Cohen’s d = 0.11). Moreover, we find that video has a small positive effect on the persuasiveness of, and engagement with, non-political content — but not political content. These results suggest that fears about the unique persuasive power of video-based political messaging may be premature. Though video may seem (slightly) more believable than text, this enhanced credibility does not necessarily facilitate political persuasion.