No, they don’t all do it the same way. Many elite privates, possibly most, exclude all graduate students from their s/f calculation. But some also exclude the faculty in stand-alone graduate/professional programs, as per the instructions. Harvard, for example, counts only undergraduates, but excludes about half of its faculty–presumably all the law school, med school, business school, divinity school, and Kennedy School faculty as those are all graduate-only programs, and probably more that I’m not thinking about right now.
No, for two reasons. First, they are plainly violating the instructions, so they are putting up numbers that simply aren’t comparable to those put up by schools that do follow the instructions. That’s extremely misleading, and that must be their intent. Second, graduate students do take faculty time and attention–probably more, on a per capita basis, than undergraduates. So even if you’re using the CDS primarily to evaluate schools at the undergraduate level, you’ll want to know how many students are competing for faculty members’ time and attention, and that includes graduate students as well as undergrads. Look, if you have a LAC with no graduate programs and no graduate students, you can get a pure s/f ratio: just compare the number of FTE students, all undergrads, to the number of FTE faculty, all teaching only undergrads, and that’s your s/f ratio. But as soon as you start adding graduate programs, faculty in those programs are going to start to devote some of their time and attention to the graduate students, and that means they’ll have less time for undergrads. That’s why the CDS draws the line where it does. It says if the faculty are teaching only graduate-level students, as in a law school or a medical school, neither those faculty nor those students should be included in the s/f ratio, because they’re just irrelevant to the undergraduate experience. But if the chemistry department or the political science department is teaching both undergrads and graduate students, you need to assume some of the department’s faculty time is going to be devoted to the grad students, so you need to count the grad students along with the undergrads and the faculty in the s/f ratio. That strikes me as eminently sensible and a fairer representation of how much time faculty are likely to have for undergraduates. A history professor at Amherst is going to spend 100% of her teaching, supervising, and mentoring time with undergrads. A history professor at Harvard is going to spend only a fraction of her teaching, supervising, and mentoring time with undergrads, and the rest with grad students. You can’t count them the same way. But by deliberately excluding the history grad students from its “student” count for purposes of its s/f ratio, Harvard is creating the impression that its history faculty work only with undergrads, and that simply is not the case.