Sure, I agree with you, excluding them is biased as well. A more appropriate phrase to be used to describe "people who have jobs that allow them to work from home" is "a small minority".

And let's be real, biasing your sample by not including a small minority doesn't affect the outcome of a study nearly as much as doing the opposite and sampling mostly from that small minority. Does it make the results less accurate? Yes. Does it make the results nearly as inaccurate as it would be if they excluded most of people but that minority? Not at all.