The disputed role of knowledge in truth illusions: A critical test based on comparative judgments

It has been argued that the illusory truth effect - the tendency to judge repeated statements as more likely true - persists even when a statement is known to be false, whereas a competing view maintains that knowledge shields against such truth illusions. In this study, we reexamine two models of truth judgment that formalize these opposing perspectives: the fluency-conditional model (FCM), which assumes that repetition-induced fluency can override knowledge, and the knowledge-conditional model (KCM), which assumes that knowledge always takes precedence over fluency, such that fluency only becomes relevant in the complete absence of knowledge. Utilizing a novel two-alternative forced-choice paradigm, we derive and test critical qualitative predictions of each model. The results disconfirm the FCM while the KCM can be maintained. Quantitative assessments of both models corroborate this conclusion and extend it to some alternative model specifications. We further show that the FCM, as per its qualitative predictions, is incompatible with the assumption that repetition leads to a uniform increase in perceived plausibility. A continuous single-process model operating on this assumption aligns with the KCM's predictions but implies a different psychological mechanism. We conclude by outlining avenues for empirically dissociating these theoretical accounts.