Having them expose themselves more to the virus is a "good thing" (though tell them how long should they wait for the immunity to set in)
Yes, with the control group you would have a base for how many people you would have expected to be infected, so there's that, but you "could" get that from other observations
> Yes, with the control group you would have a base for how many people you would have expected to be infected, so there's that, but you "could" get that from other observations.
Getting this number for a distribution of people as similar as possible to the people in the experimental group is, in fact, the primary purpose of having a control group. There is no better way to predict this number, as you can't predict how likely a particular person is to catch coronavirus at a particular time, in a particular location.
You also need a control group to get the number of various medical problems expected in the experimental group, even if the vaccine doesn't cause them. This also depends on time, place, and behavior.
So, if you create a reason for people in the control group to behave differently from experimental group, for instance by letting people find out which group they are in, you defeat the purpose of having a control group.
I'm not denying there's the need for a baseline to compare to, but the need to have an dedicated designated group to be the 'official control'. I agree it makes things easier statistically.
> So, if you create a reason for people in the control group to behave differently from experimental group
Yes and here lies a minor catch, both those groups might behave differently from the population in general (even though they aren't "supposed to")
>Why are so many people commenting when they clearly don't understand knowledge-creation/statistics/methodology?
Because (a) we want to understand. and (b) reasonable experts differ on what constitutes proper methodology which is constantly being improved.
Surely the effect size of a useful vaccine is large enough that modified behaviour of the vaccinated does not matter at all. We're not looking for a small incremental improvement in contraction rate of the virus here, we're looking for a very large percentage of those who have been administered the vaccine not getting covid. Yeah it could change behaviour but so what? How is that going to bias the result to suggest it isn't working if it is? (or vice-versa)
Anti-scientific sentiment is at high levels because you don't get someone to personally explain basic statistical concepts that you could have googled yourself?
Having them expose themselves more to the virus is a "good thing" (though tell them how long should they wait for the immunity to set in)
Yes, with the control group you would have a base for how many people you would have expected to be infected, so there's that, but you "could" get that from other observations