Historically, our department has been a small, theoretically-heavy, niche department, with a tiny cohort (~3 phd, ~10 masters per year). In recent years, we've started expanding both faculty and student numbers, boosted by our successful marketing plot to turn into Data Science (~6 phd, ~20 masters per year). That being said, it's not like we've sold out (looking at you, Columbia) – we still take pride in our small-ish cohort sizes, and phd and masters students take the same classes.
However, the increase in applicants has been disproportionate in size (hello, Data Science), and as a result our acceptance rates are way out of whack. Like with anything that gets deluged, the result of this is that we are much more selective (and go with easy, obvious signals for an initial cull).
~
Oops, I forgot to respond to the OP: we definitely spent relatively more time considering each domestic student. Fewer in numbers, oftentimes more interesting backgrounds. Of course, we're talking about the top schools here. We also have *other* considerations that make domestic students sometimes more compelling.