In three studies (ns = 582, 790, 495), we examined whether self-reported agreement with conspiracy theories primarily captures convictions (i.e., firmly held beliefs) or (im)plausibility judgments. In Study 1, most (76.9%) participants who reported complete (dis)agreement with specific conspiracy theories considered that the available evidence definitely proved their position. In Studies 1–2, the relationship between agreement and subjective knowledge about conspiracy theories was robustly “U”-shaped. In Study 3, agreement was more strongly associated with veracity (r = .92) than plausibility (r = .82) assessments. These findings suggest that extreme conspiracy belief scores primarily capture subjectively well-informed convictions rather than mere plausibility judgments. Finally, the absence of moderation by subjective knowledge for well-established correlates of conspiracy beliefs in Study 2, and the strong intercorrelations between agreement, plausibility, and veracity assessments in Study 3, suggest that the investigated distinctions are unlikely to substantially challenge the validity of past research.