In this preregistered study, we investigated the type of knowledge people use to discriminate between true and fake news by asking participants (N = 327 Prolific users residing in the United States) to rate the veracity of different news headlines and indicate what decision strategy they used to make each rating (guess, intuition, familiarity, prior knowledge, rule, or other). We found that participants discriminated well between true and fake news headlines, and predominantly chose decision strategies that suggested they were using tacit knowledge (knowledge that is not easily articulated) rather than explicit knowledge (knowledge that is easily articulated). For example, guess and intuition were chosen 63% of the time, and participants’ discrimination was good even when they claimed to be guessing. The fact that tacit knowledge formed the dominant basis of participants’ discriminative ability speaks to the types of interventions that are likely to be successful in improving this skill.