COVID-19 treatments, causal inference, robustness, policy
omitted variable bias, coefficient of proportionality, exact