Temperature dependence of COVID-19 transmission

Alessio Notari

is the (which was not peer-reviewed) The copyright holder for this preprint Table II : Best-estimate, standard deviation (σ) and 95% C.L. intervals for the parameters of the quadratic interpolation, for the base set of 42 countries. We repeat then the same analysis for the extended dataset of 88 countries. Results for the linear fit, together with our data points, are shown in fig. 4 . The parameter estimation contours are shown in fig. 5 , and the estimate, standard deviation and confidence intervals for the parameters are shown in Table III . The slope β is smaller in absolute value, but the significance remains high, since a zero slope is excluded at 99.86% C.L. (p-value 0.0014). Now R 2 = 0.11 and R 2 adjusted = 0.10. In this enlarged sample the quadratic trend is not visible anymore, and indeed the AIC does not prefer the quadratic fit: ∆AIC = +0.9 compared to the linear fit, in disfavor of the quadratic model. The R 2 is also practically the same as in the linear fit. Table III : Best-estimate, standard deviation (σ) and 95% C.L. intervals for the parameters of the linear interpolation, for the extended set of 88 countries.