Earlier quoted context omitted.
> After a lot of arguing and furious paper writing in the late 80s-90s, everyone began working on how to deal with that bias, moving it front and center where it can be handled instead of subtly hiding it in a bunch of numbers and pretending it didn't exist. And the result of that is... what exactly?
Vastly improved methodology and understanding of results. One of the issues of modernist research was that objective data was trash, but no one knew it. For instance, anthropologists would throw out / fail to collect artifacts that weren't usable with the analyses available at the time. You can't recover that from a table that says 50 bone shards were found in a pit. You can begin to recover that if the researcher fl…
I would think that the foundation behind your argument - which I agree with - implies that this isn't actually possible. You have an infinite regress of biases.
Empirically, what you can do - to take your example of those menstruation researchers - is to get more data from researchers of different backgrounds to try and statistically isolate those biases. This is part of "standard empiricism". You also always have to live with the fact that data is sometimes too noisy to be meaningful.