An Online Game to Help Canadian Adults Identify COVID-19 Vaccine Misinformation With One-Month Follow-Up

Overview

COVID-19 vaccine misinformation contributed to vaccine hesitancy and preventable illness. This study tested whether a brief online game helps Canadian adults identify COVID-19 vaccine misinformation, and whether effects last one month. In the game, participants judged whether statements about COVID-19 vaccines were true or false and immediately received the correct answer with a brief evidence-based explanation, delivered in a game-like format with points, ranks, and feedback. Adults from a national non-probability online panel (n = 4,071) were randomized to the game, to the identical corrective content presented as a static blog post, or to a no-activity control. The game and blog groups were assessed immediately after the activity; all three groups were reassessed approximately one month later. The hypothesis was that gamified corrective information would produce more accurate knowledge than the same information presented as a static blog post and, at one month, than no activity. The primary outcome was accuracy in identifying true and false claims; vaccination intention, vaccine confidence, and self-reported vaccination were secondary outcomes. This is the fourth study in a research program whose first three trials are registered in a companion record.

This randomized controlled trial was the fourth study in a research program evaluating digital educational interventions to counter COVID-19 misinformation. The first three studies were conducted under a separate exempt determination from the University of South Carolina Institutional Review Board (Pro00099400) before participant enrollment and are registered in a companion record. This trial received a separate exempt determination (Pro00122534). This record retrospectively registers the completed trial. The trial used the same intervention platform as the earlier studies: a true/false quiz covering a larger set of circulating COVID-19 vaccine misinformation items, with each response followed immediately by item-level refutational feedback (the correct answer plus a brief evidence-based explanation) embedded in a gamified interface with points, ranks, and affective feedback. A six-item subset was administered in the study surveys to measure knowledge accuracy. Participants (N = 4,071) were recruited from a national non-probability online panel in Canada (Prodege, fielded by Nanos Research) and randomized to three arms: Game (n = 1,168), Text - the identical refutational content texts as a static blog post (n = 1,288), or no-activity Control (n = 1,615). The questionnaire was administered in English and French nationwide, including Quebec. All arms completed a Wave 1 baseline (fielded December 8-30, 2022); Game and Text then received their assigned activity and an immediate posttest; the Control arm received no activity and, by design, no immediate assessment. All three arms were reassessed at Wave 2, approximately one month later (January 23-February 13, 2023); 2,079 of 4,071 (51.1%) were retained. Two contrasts were confirmatory: Game versus Text (incremental effect of gamified delivery; evaluated immediately and at one month) and Game versus Control (total intervention effect; one-month horizon, as the Control arm had no immediate assessment). Text versus Control was descriptive. Immediate analyses included all 2,456 participants randomized to the Game and Text arms. No post-randomization data-quality exclusions were applied. One-month analyses included all 2,079 participants who completed Wave 2; nonparticipation at Wave 2 was treated as loss to follow-up rather than an analytic exclusion.