Originally Posted by
bobbytables
the hypothesis is that temperature affects infection (transmission). There is an unknown and variable delay between infection and recording as a case. In particular that delay can vary a lot between different places. This, temperature vs. number of cases is a deeply problematic - I would say almost useless - statistic.
A good study will take all these into consideration, by stratifying the cases in a sufficiently large sample. Among the simpler analysis mentioned on the COVID-19 thread before the split, one paper merits a peer review. See
Coronavirus / COVID-19 : general fact-based reporting.post 3846)
Past experiences with other coronaviruses suggested a relationship. Since it is difficult to imagine transmission affecting temperature, the empirical evidence of thse other viruses suggests temperature influences transmission. We are eagerly waiting to see if this is also true with COVID-19 viruses. I was pleased to see one decent attempt (the paper mentioned in the above post) to support this relationship. I do not have time to ask for the details of the original data or to review the statistical significance, but the work looks solid.