An Inductive Learning Intervention to Improve News Veracity Discernment

Across three preregistered experiments (total N = 1,135), we tested whether an inductive learning (IL) intervention improved participants’ ability to distinguish between true and false news. IL involves learning categories by observing or classifying exemplars. Therefore, in this research, IL involved observing true and false news headlines and classifying them as either “true” or “false”, with immediate feedback on accuracy. In each experiment, participants took part in either an IL or control condition, the latter of which involved no training or playing Pac-Man. To test participants’ ability to distinguish between true and false news headlines, a final test was administered where all participants rated the veracity of true and false news headlines, which were different from those used during IL, on a scale from 1 (high confidence false) to 6 (high confidence true). In Experiment 1 (N = 214), the IL intervention significantly improved participants’ news veracity discernment compared to control, but the Bayesian evidence was only anecdotal. In Experiment 2 (N = 483), we incorporated game-design elements into the IL intervention, including performance-contingent badges. Unexpectedly, the effect of the IL intervention decreased. We reasoned that, because the IL intervention involved easy-to-hard training, the provision of performance-contingent badges inadvertently made participants more aware of their declining performance. This awareness may have undermined their motivation to learn as the training progressed. Therefore, in Experiment 3 (N = 438), we implemented hard-to-easy training instead, which resulted in the IL intervention significantly improving participants’ news veracity discernment, now with strong Bayesian evidence.