Impact of Repeated Exposure to Polarized Health-Related News on Explicit and Implicit Attitudes Toward Dietary Supplements: Online Experimental Study.

Résumé

BACKGROUND: Repetition is a central feature of digital news consumption, where engagement-driven algorithms often expose users to similar health-related content. Prior research suggests that repeated exposure can influence perceived truth and evaluation, but most studies use brief or decontextualized stimuli and have rarely distinguished between explicit and implicit attitudes. Little is known about how repeated exposure to full-length, polarized health news shapes explicit and implicit attitudes, particularly toward familiar products such as dietary supplements.OBJECTIVE: This study aims to examine whether 2 weeks of exposure to positively, negatively, or mixed-valence health news articles would alter explicit and implicit attitudes toward dietary supplements, and whether engagement mediated these effects or initial attitudes moderated them.METHODS: In a preregistered 4 (group: PRO, CON, MIX, and control) × 3 (time: T0, T1, and T2) mixed experimental study, 228 participants (174 women, PRO: n=68, CON: n=51, MIX: n=52, control: n=57; mean age 2.81, SD 4.88 years) were randomly assigned to receive one full-length article per day for 2 weeks. Articles presented positive (PRO), negative (CON), mixed (MIX), or neutral space-related information (control). Explicit attitudes toward dietary supplements (perceived efficiency, harmfulness, and willingness to recommend) were assessed with visual analog scales, and implicit attitudes with an Implicit Association Test at baseline, 1 week, and 2 weeks. Mixed analysis of covariances (ANCOVAs) controlled for self-reported exposure to health-related news. Mediation, moderation, and moderated mediation analyses examined whether time spent reading and baseline attitudes influenced change.RESULTS: The change in implicit attitudes was not significant (F6, 428=1.90; P=.08). In contrast, explicit attitudes showed a significant group × time interaction (F5.04, 428=14.34; P