Do prebunking and debunking protect against novel misinformation?

Misinformation can negatively influence belief formation and reasoning. To counter these influences, a range of countermeasures have been developed, the two dominant approaches being (proactive) prebunking and (reactive) debunking. While psychological research has revealed important insights into the efficacy of these interventions, the literature (1) lacks comparative analysis contrasting the two approaches, (2) has mostly used time-constrained designs that provide conducive intervention contexts, and (3) has focussed predominantly on cognitive measures, neglecting behavioural outcomes. We addressed these limitations in an experiment (N = 793) that provided two misleading articles on separate topics. We contrasted prebunking and debunking interventions that targeted the initial misinformation (article 1), and tested to what extent the interventions conferred protection against novel misinformation (article 2) either immediately or after a one-week delay. We included standard questionnaire measures of inferential reasoning as well as a text-generation task (writing a social-media post). With no delay, prebunking effectively protected against novel misinformation on both measures, while debunking effects were statistically significant only on the cognitive measure. Neither intervention was effective after a seven-day delay.