With colleagues, I'm applying for an NSF-funded mathematics REU site at my institute. In putting together the budget, it's easy to see how much is an appropriate stipend for student participants, because student stipends are published on the recruiting sites for other REUs. However, I wondered if other faculty on this site who have been involved in an REU could provide a ballpark as to what is a usual rate of compensation for faculty mentors (over the course of a 10-week program, pairs of faculty mentors in complementary disciplines would guide teams of 4 students). I don't want to undercompensate faculty mentors, but at the same time, the main portion of the budget is meant to support student participants, so there must be some sweet spot in there. Any input based on recent experience would be much appreciated! Thank you!
The NSF's guidelines are that you can generally pay one faculty-month of salary. In principle, you can propose to divide this money up however you like, but if you try to divide up the compensation among all the faculty members involved, it may not be enough money going to any single person to really be worthwhile.
As the PI of an REU site, I do not pay the other faculty supervisors anything. However, my field is not mathematics but physics. The faculty mentors, most of whom work in experiment but some of whom are theorists, would all be on campus doing research even if the REU students were not present. Most of them are getting paid summer salary off other research grants. The only direct faculty compensation goes to me, since I manage all the advertising, admissions, and logistics for the program, which really do amount to about a month of full time work. Some other programs have a PI and co-PI who share the responsibilities and compensation basically equally, but I would suggest not dividing the money up too finely.