Maybe Firefox usage statistics could be distorted, because of its various mechanisms of blocking trackers.
User-Agent headers are mandatory; and it would be counterproductive for everybody, including Mozilla, if Firefox were sending invalid User-Agent headers.
UA switcher is a pretty popular addon in the Mozilla ecosystem. So it may well be more popular than UA stats alone can surmise. (Though probably not enough to matter to businesses too lazy to test anything besides the top two.)
I didn't think that that blocked the user-agent header. But I guess I could be wrong.