🎓 Dr. Freemo 🇳🇱 @freemo@social.cleverlibre.org

@spazzpp2 you are missing the point. If people in a study are wearing a mask because the choose to do so then you cant be sure if the mask is the thing effecting the transmission of the disease or if it is something else that mask wearers are likely to do non-mask wearers dont.

For example say people who take the virus seriously are the people who wear masks, they are also the people who wash their hands frequently. That means people who wear masks are likely to be the people washing their hands. So even if masks have no effect but hand washing does then you would observe people who wear masks get sick less even though the masks arent the reason.

@spazzpp2 In what way is it not observational? The problem here is your drawing inference from things happening in the wild where wearing a mask is a choice

@spazzpp2 exactly, thus the problem. Interviewd people are "observational data" and their choices are not controlled. Thus the studies are victim to the fallacy mentioned

@spazzpp2 yes but you;d have to do so in a controlled setting where you know the **only** factor that is different is the mask. That would mean intentionally exposing a group to the virus which is unethical and thus impossible.

Any attempt to do the experiment using observational data, and thus other variables are not controlled would lead to the post hoc ergo propter hoc fallacy i mentioned.

So in effect there is no way to actually test the hypothesis and thus the problem.

@spazzpp2 but it doesnt change how we analyze the contribution of any factor. so the reason current studies are invalidated remains true regardless of the R0 value.

@spazzpp2 yea but since the infection rate doesnt effect the logic here you dont really need to imagine it at all

@spazzpp2 Well not really a thought experiment so much as established practice when it comes to statistical analysis.

Not sure why extremely high infection rates would change the principles here.