Staff Pick Comment:

You essentially can’t perform a randomized controlled trial on masking during a pandemic for a LOT of reasons.

1. Your institutional review board would never approve it because there’s enough evidence of masks reducing disease spread generally that it would be considered placing participants at risk of harm for limited scientific benefit.
2. You would have largely insurmountable confounders, ESPECIALLY during a pandemic when there are likely to be rules people need to follow regardless of what mask group they were randomly assigned to.
3. Behavioral confounders would be enormous. Since people would be randomly assigned to a mask group, and participation must necessarily be voluntary; the only participants would be people who don’t care about wearing or not wearing a mask or truly don’t have any opinion about if it’s safe or dorky looking or socially unacceptable in their circle to be seen wearing/not wearing a mask. The sample would not be at all representative of most people.
4. Adherence would be terrible. The same applies to virtually any study about diet, activity, changing habits, dependence on things like caffeine or nicotine, etc. If participants don’t follow the prescribed behavior of their randomized group the data gets weaker and weaker until there is no longer any visible signal.
5. Even if all else was solved by magic, the study would still not be blinded. People would know if they’re wearing a mask or not and those around them would also know. This would create it’s own set of confounders.

Observational data is likely to be the best and only data you can get on this topic. To throw it out because it’s not randomized and controlled is to admit that you have no idea what any of these words mean or how such studies are conducted.