When Evidence Exists but Is Not Used: Diagnosing Policy Responsiveness in COVID-19 Education Policy

During the COVID-19 pandemic, governments drew heavily on scientific evidence to inform education policy under urgency and uncertainty. Prior studies show that recent research was visible in policy documents, but it remains unclear whether policy systems updated their evidence base as new, relevant studies became available. We introduce a method to diagnose policy responsiveness by separating the availability of substantively relevant research from its incorporation into policy texts. Using large-scale semantic similarity analysis, we identify cases where newer, closely related studies were available before a policy was issued. We find that policy documents often continued to cite established pre-pandemic research even when relevant post-2020 studies were already available. This pattern suggests that evidence-use systems updated incrementally, with lags shaped by advisory routines and institutional infrastructures rather than by a lack of relevant evidence. By distinguishing availability from uptake, our approach offers a generalisable and scalable way to assess how quickly policy systems incorporate newly available research, and to compare responsiveness across domains and institutions.