Models in the Post-Truth Age: Politics and predictive statistics in the US COVID-19 Pandemic Response

Abstract

The COVID-19 pandemic in the United States marked one of the first national crises in which predictive mathematical models played a significant role in the day-to-day life of the nation, impacting political, economic, and social structures and relations to an unprecedented extent. Through digital ethnography, interviews, and participant observation in the uniquely online and open spaces of scientific practice and meaning-making during the early quarantine phases of the US COVID-19 pandemic, this dissertation explores the role that mathematical models played in shaping the national discourse, guiding individual-level beliefs and behaviors, and influencing the decisions and actions of policy makers and public health officials. The specific means, processes, and pathways through which COVID-19 models’ projections operated at various scales to facilitate numerous direct and distributed effects on belief, behavior, policy, and more are explored to shed light on the general potentials and dangers of predictive models in the context of the highly uncertain crisis environments like the global COVID-19 pandemic. I argue that predictive models, by their very nature, trouble the traditional dichotomies that undergird our contemporary political and epistemic orders, constituting neither “fact” or “fiction, “science” or “ideology,” nor “science” or “policy.” I argue that the dual science-policy, fact-ideology, science-policy character of models mirrors the broader challenges of the “post-truth” age and model-driven policies (e.g. climate change) and can thus provide insight into new and possible ways of existing in the contemporary world.