“…The three cities were compared according to temporal trends in the rate of positive test results, as assessed in the two waves of the pandemics which strike such regions during the time course of sampling. For practical reasons, these waves were detached by the lowering of the daily incidence rate around the October to November 2020 edge, as detected using external, unofficial data 13 . The probability of a positive test result (response variable) was assessed by generalized linear models with a binomial error distribution, using wave, city, day of collection, age, sex, and race/ethnicity as risk factors (predictors).…”