Here's the actual study. https://www.acpjournals.org/doi/10.7326/M20-6817 https://www.acpjournals.org/doi/10.7326/M20-6817 The intervention being tested was m

Here's the actual study. https://www.acpjournals.org/doi/10.7326/M20-6817 https://www.acpjournals.org/doi/10.7326/M20-6817 The intervention being tested was merely how a RECOMMENDATION to wear a mask (and other advice and a pack of 50 surgical masks) may affect infection rates. Of the mask wearers, and based on asking them and accepting their answers as truthful, 46% wore the mask as recommended, 47% "predominantly" as recommended, 7% not as recommended. The 7% were excluded from the results. The infection in the area at the time was 2%. This gives relatively fewer chances for the efficacy of masks to be tested and so it is not surprising that the results are inconclusive. The benefit masks bring in this case is keeping the infection rate low and only actively do their job during the relatively rare times when an infected individual is near a healthy one. The low infection rate also leads to an inflation of any systematic differences between the two samples. A very small difference applied to the expected 98% healthy subsample can lead to a strong contribution to the the 2% infected subsample. For example, mask wearing at the home was not covered and a few reported covid infections at the home. Testing in places where infection rates are higher would have a better chance to show an effect. Of course that would be stymied by sample bias (if "random" self-selection of mask-wearers and anti-mask idiots were allowed in the procedures) or it would be faced with unethical practice of forcibly withholding masks from individuals in a non-mask sample. My personal conclusion is that these results are not conclusive but will be hugely appropriated for whatever sick fantasy the outgoing US administration is orchestrating.