“This study claims to support mask mandates, by demonstrating that there was a statistically-significant reduction in the case development rate in mandate-locales, but an insignificant reduction in non-mandate locales. This is an invalid statistical argument. Just because variable ‘A’ changes significantly over time, and variable ‘B’ changes less significantly, does not mean the two variables are statistically different. This analysis requires that the confidence intervals (error bars) between the two variables don’t overlap. If you look at the two graphs presented, and overlap them, you will see that the confidence intervals overlap, proving that there is no statistically-significant difference between the mandate and non-mandate locales. It therefore, if anything, argues against mandates. This is a rather rudimentary error in statistics. Usually when such data is presented, both variables are presented on the same graph, so that this comparison is easier to see. I cannot explain why the authors made the mistake, and presented the graphical data as they do. Even if this can be seen to argue for mandates, the effect was very transient, and of very small degree.”