You have to allow for some level of pragmatism. They found a method that seems to provide better data without actively biasing their results. Rodents make convenient test subjects, and no scientist wants (or more likely it doesn't even occur to them) to include an unproven methodology in their published work.

Are they publishing "make sure only [female name] handles the mice"?

Or do they just publish, "X caused effect Y"?

Frankly, identifying why it requires a female, or female non-cat owner sounds like it may be more interesting than whatever other result they are working on...

If results were thrown out because of a confounding factor then it would hopefully be mentioned in the paper.

I've had a similar thought about these little insights. There are literally thousands of times more biology undergrads than there are people doing biology research. Biological experiments are notoriously difficult to conduct. There are so many confounding variables, data is hard to collect, etc, that experiments bomb as often as they succeed.

A lot of these experiments are extremely well supported, so you might actually be able to discover something novel if you were able to track the failures.

Anyways, I don't do anything with biology anymore, but I strongly believe that biology is a field where citizen scientists could make a meaningful impact.

Another thing to consider is epigenetics.

https://www.newscientist.com/article/dn24677-fear-of-a-smell...