You're talking about a base rate fallacy, which is a different thing than correlation/causation errors. (Both are potentially problems given the high-level description of the study, the full text of which is behind a paywall)