Reliance on emotion promotes belief in fake news

Contributions

• Demonstrates that elevated emotional engagement—across a wide range of specific emotions, not just a few—predicts both increased belief in fake news and decreased ability to discern real news.
• Provides large-scale causal evidence that experimentally manipulating reliance on emotion (vs. reason) causally increases belief in false news.
• Shows that inducing reliance on reason decreases belief in fake news, supporting a dual-process model where emotion and reason pull in opposite directions on misinformation susceptibility.
• Challenges prior work claiming emotion only influences a narrow set of beliefs; instead, emotion broadly heightens credulity toward implausible content.

Method

Study 1 (N = 409) assessed the relationship between baseline emotionality (measured via the PANAS 20-item scale before exposure to news headlines) and belief in fake and real news headlines presented with images and source labels. Participants rated headline accuracy on a 4-point Likert scale. The authors used mixed-effects regression to isolate the effect of each emotion dimension while controlling for participant and headline random intercepts.

Study 2 (N = 3,884) experimentally manipulated reliance on emotion versus reason using priming instructions. Participants in the emotion induction condition were told "Many people believe that emotion leads to good decision-making"; those in the reason induction condition heard "Please assess the news headlines by relying on reason, rather than emotion"; control received no induction. Participants then rated accuracy of 10 headlines (5 fake, 5 real; half Democratic-aligned, half Republican). The authors examined whether induced reliance on emotion causally increased false-news belief and decreased discernment, using linear mixed-effects models with random slopes and intercepts nested by participant and headline.

Studies 3 and 4 replicated the experimental design across different participant samples (Studies 3: N = 1,129 from MTurk; Study 4: N = 1,175 from Lucid panel) with similar manipulations and headline sets.

Results

Correlational (Study 1): Nearly every emotion measured by the PANAS (enthusiasm, interest, determination, excitement, inspiration, alert, active, strong, proud, attentive, scared, afraid, upset, distressed, jittery, nervous, ashamed, hostile, guilty, irritable, positive, negative) showed significant positive associations with belief in fake-news headlines and decreased discernment (real minus fake belief). The relationship held across political concordance conditions.

Causal (Study 2): Inducing reliance on emotion (M = 3.47) versus control (M = 2.06) significantly increased belief in fake-news headlines. Inducing reliance on reason (M = 1.41) decreased fake-news belief compared to control. The emotion condition also significantly decreased discernment (reduced ability to rate real news higher than fake). Crucially, inducing reliance on reason did not significantly affect real-news belief, suggesting the effect of emotion is specific to increased credulity toward false content. The causal effect of emotion reliance on fake-news belief remained substantial even when controlling for headline familiarity and political concordance.

Meta-analysis across studies: Consistent findings across all four experiments confirmed that both emotional processing (broad across specific emotions) and induced reliance on emotion causally promote false-news belief.

Notes

Strengths:

• Large-scale causal evidence across four preregistered experiments with sample sizes > 400 per study.
• Broad measurement of emotions (PANAS scale) shows the effect is not limited to anger or anxiety, but spans positive and negative emotions.
• Clear causal chain: induced emotion → increased false-news belief; induced reason → decreased false-news belief.
• Rigorous controls for participant and headline random effects, headline familiarity, and political concordance.
• Replication across distinct participant populations (MTurk, Lucid) strengthens external validity.

Weaknesses and open questions:

• Headlines were selected from 2017–2018; unclear how findings generalize to more recent misinformation ecosystems or to visual/multimedia false news.
• The emotion-induction manipulation is brief (single-sentence prompt); real-world drivers of emotional engagement may operate differently.
• Study 2's emotion induction instruction is somewhat blunt ("Many people believe emotion leads to good decision-making"); more naturalistic inductions (e.g., affective priming with images, music) might reveal different effects.
• Unclear whether the findings apply equally to all political orientations; the paper notes some variation by Trump vs. Clinton supporter partisanship, but doesn't fully decompose this.

Implications for interventions:

• Suggests that encouraging analytical thinking or even brief prompts to rely on reason could reduce susceptibility to false news.
• Raises questions about online environments that systematically prime emotional responses (e.g., sensational headlines, outrage-inducing framing, social engagement metrics tied to reactions).