The Counter-Attitudinal Illusory Truth Effect

Abstract

In the digital age, rapid dissemination of information has increased the challenge of truth assessment, particularly given the proliferation of misinformation and its potential societal consequences. To aid understanding of truth assessment processes, two distinct research literatures have identified repetition and attitudinal alignment as critical factors influencing truth judgments. The Illusory Truth Effect (ITE), first documented by Hasher et al. (1977), demonstrates that mere repetition enhances a claim's perceived truthfulness. This phenomenon is commonly evidenced through experimental paradigms comparing truth ratings between repeated and new claims across diverse materials, from trivia to marketing claims. While the robustness of the ITE is well-established, existing research has not systematically examined the interaction of the ITE with attitudinal alignment - another influential factor in truth assessment. The alignment bias literature consistently demonstrates that individuals evaluate pro-attitudinal claims as more truthful than counter-attitudinal ones, reflecting influences from the alignment of material with prior attitudes (review: Ditto et al., 2025). However, this research stream has yet to critically examine the role of repetition, despite the likelihood of repeated exposure to claims in digital environments. This gap presents a critical opportunity for investigating the potential interaction between repetition and attitudinal alignment in truth perception.

In this thesis, we explore this question by investigating the ITE for material that varies in attitudinal alignment. In the first study, we measured the ITE for claims related to climate change science or skeptic attitudes. For our samples of climate science endorsers, repetition consistently increased perceptions of truth for both pro-attitudinal and counter-attitudinal claims (Experiment 1 & 2), even for the strongest endorsers of climate science and when people could later identify claims as skeptic-aligning (Experiment 2). In the second study, we investigated whether the pro- and counter-attitudinal ITE was moderated by increased alignment salience of claims at the encoding phase, through either an active alignment sorting task (Experiment 3) or the passive consumption of alignment labels (Experiment 4). We found that increased alignment salience had no effect on the pro-attitudinal ITE, but eliminated the counter-attitudinal ITE. In the third study, we examined the pro- and counter-attitudinal ITE in the context of political attitudes, by presenting political sentiment claims to both Democrats and Republicans. We found that while Republicans showed a robust ITE in both experiments, Democrats showed a smaller ITE in Experiment 5 and no ITE in Experiment 6. These patterns occurred regardless of claim alignment with people's attitudes, and regardless of whether people were primed with affiliation identification before the ITE phase (Experiment 6). To further examine this pattern, I conducted a mini meta-analysis examining the role of political affiliation on the magnitude of the ITE, across different materials including trivia, climate change claims, and political sentiments. We found some evidence that the role of political affiliation in the ITE was dependent on material type.

Together, these studies extend understanding of the role of attitudinal inputs in the ITE, with main takeaways from the current evidence being 1) an ITE can emerge for both pro- and counter-attitudinal content, 2) alignment salience eliminates the counter-attitudinal ITE, and 3) Democrats show a smaller ITE than Republicans depending on materials. Our findings provide some evidence for a form of attitude neglect in assessments of truth. This research also presents opportunities to further explore boundary conditions of the role of alignment salience in the ITE, and highlights the need to systematically examine potential group differences in susceptibility to the ITE. Show more