Dataset 3: Dickey-Fuller tests and Cumby-Huizinga tests

The primary interrupted-series analyses evaluated all targeted and total border closures by country-level intervention points. As robustness checks, Dickey-Fuller tests were run for every time series, with none exceeding the 5% critical value of the t-distribution for a unit root. Cumby-Huizinga general tests were run to correct results for each lag found to have serial autocorrelation present for all global and country interrupted time-series.