Figure 2.

Empirical estimates of the association between COVID-19 growth rates and mobility, restrictions and environmental conditions. The cumulative effect of each predictor variable is derived from mixed-effects linear regression models built for high-, middle- and low-income countries, pre- and post-June 2020. Points and lines represent the model estimates and 95% confidence intervals. Non-significant (“non-sig”) estimates are marked with an “x” whereas significant (“sig”) estimates are marked with a solid point. Estimates are expressed as percentage changes in daily COVID-19 cases per standard deviation (δ) increase in the predictor variable.