> I also can't help myself but point out that it is very bad to extrapolate from a sample of outliers.
I don't think that's exactly what's happening here.
As an analogy, look at twin studies: they're undeniably based on "a sample of outliers" in a sense, yet they're also key tools in many sciences because the outlier-ness of the participants isn't [directly related to] what's being measured. Rather, it's an external factor that allows the experimenter to control for other factors that might otherwise be difficult or impossible to account for.
The accounts mentioned in this article are far from an objective result. Yet, who better to understand the differences between how men and women are treated in society than somebody who's existed as both? Assuming they "passed" before and after their transitions (a fair assumption IMO), they offer a unique and highly relevant perspective.
With all the modern political strife around these issues, which includes a lot of misinformation and misrepresentation, I think it's easy to draw certain lines between "trans" and "cis" people that may not have much basis in reality. I'm not trans, but it's my understanding that for many trans people "trans" isn't an identity in principle or practice. Rather, they are—for all useful intents and purposes—representative "members" of the genders they identify with.