I looked through the enrollment reports at the total number of students who enrolled for the first time as first-years (four autumns before) or transfer students (three autumns before) for the putative graduating classes of 2016, 2017, and 2018, and compared them to the number of discrete individuals awarded bachelors degrees in all quarters of calendar years 2016-2018.
So . . .
Entering students:
Class of 2016 – 1527 + 18 transfers
Class of 2017 – 1426 + 21 transfers
Class of 2018 – 1442 + 29 transfers
Total: 4463
Persons awarded bachelors degrees:
Class of 2016 – 1484 (1369 in spring)
Class of 2017 – 1577 (1396 in spring)
Class of 2018 – 1430 (1216 in spring)
Total: 4491
It’s not precise. One of the summer reports was missing, so I used the halfway point between the number of summer graduates the year before (30) and the year after (26). More importantly, some of the degrees awarded during those years were doubtless awarded to people who first enrolled prior to September 2012, and some of the people who enrolled during those years will get their degrees in 2019 or later. The fact that 28 more degrees were awarded in those classes than students enrolled is probably attributable to the fact that there was a huge jump in the entering first year class in 2015 (putative graduating class of 2019), which was 90 people larger than the previous year’s entering class, and some of them may well have graduated early. (If I hadn’t interpolated a number of summer degree recipients for the one missing report, the number of students enrolling and the number receiving degrees would have been exactly the same.)
But it’s pretty clear that there is no significant number of students “vanishing.”