I think there are a lot of issues in the methodology of the paper that you linked. Not in terms of what they did, but what it means to the GRE.
The GRE verbal section tests whether you can read academic writing. Like the math section, this varies a lot from discipline to discipline. For example, if you don't get just about perfect without studying in the math section and you're applying to math programs, then something is wrong. The verbal section is like that as well: if you've done undergraduate philosophy, creative writing, or English, this part of the test is a shoe-in where it's easy to expect close to perfect without studying. Because, with the proper background, these parts of the tests are truly pushovers, I don't see why they would lower the ceiling on the difficulty (if anything, they could raise it so it's not so truncated by 100%s).
However, it's not expected that someone in the sciences does well on the verbal section of the GRE because scientific writing, though containing a lot of jargon, does not contain as much "high vocab" as these liberal arts disciplines. That's why low verbal scores are accepted.
But the paper you linked didn't take into account how many journal articles are in different disciplines. If you look at publication volume, I am pretty sure that the broad biology/chemistry/medical field accounts for the vast majority of the literature output. There are reasons for this, including the fact that the departments are larger and generate more function, much of liberal arts research is published in the form of books instead of in academic journals, acadmeic articles in the liberal arts usually much longer with the first publication/book being published at the end of a PhD, etc.
Compare this to how often people in Bioinformatics/CS are publishing and it's pretty clear that the study favors "academic section" to mean the sciences (especially since it they did not include academic books), and then it's clear how it gets a conclusion that the tested words are not used very often in their sample. But again, I don't see why a broad mix of "spoken, fiction, popular magazines, newspapers, and academic journals" should be used as the high end marker for GRE vocab instead of 19th/20th century continental philosophical writings.