Modeling a network of beliefs surrounding parents' endorsement of COVID vaccines for children

Cognitive science offers powerful tools for addressing pressing public health needs. Here we apply the cognitive science of intuitive theories and the tools of Bayesian networks to shed light on why so few children in the US have received COVID vaccines and how we might encourage caregivers to seek out these and other life-saving vaccines for their children. 1700 US parents completed 13 belief scales on a range of topics likely to influence their endorsement of pediatric COVID vaccines. We deployed structure learning techniques to develop a cognitive model of the relationships among beliefs and their influence on endorsement of vaccines for children, which accounted for 70% of the variance in participants' beliefs in a held-out testing split. Model-based simulations suggested that educational interventions focused on the effectiveness of COVID vaccines for supporting individual and community health may be most effective in increasing uptake of the COVID vaccine for children.