I'm currently enrolled in a PhD program in the US. Before picking up my master's degree, I'll be applying to a different PhD program that's more suited to my interests. I will be able to get two letters of recommendations from my undergraduate university and one letter of recommendation from my current graduate institution. Does this combination seem all right? If possible, I can get a second letter from my current graduate intuition as well, but I suspect that letter won't be very strong due to my not having done a lot of work with that faculty member.

Does this combination look all right?

2 Answers 2


(Again, for the U.S., especially in mathematics, my field...) Generally, the more recent the basis for letters, the better, by far. For that matter, having 2/3 of your letters be pre-graduate program might make admissions committees wonder what happened in your graduate work.

If you can get an absolutely stellar letter from undergrad, that might be helpful, but also really try to find two solid letters from your graduate program, attesting to your capacity for graduate work, rather than just saying you aced undergrad material.

  • I'm applying to applied math and statistics programs. Here's the breakdown of my letters: one's from a professor with whom I completed my undergraduate project, one's from a math professor who supervised my undergrad coursework and mentored me, and the third letter would be from my graduate program adviser with whom I took one class and two reading courses. Haven't really been able to do any research here. Does the combo look right in context? It seems the best I can do.
    – Nobody
    Aug 19, 2019 at 20:04
  • 1
    For math and stats, a strong letter from a graduate course instructor in a serious course, in which you did excellently, would be better than almost any letter from a person from your undergrad institution. Both because it's more recent, and because it would refer to higher-level material. Aug 20, 2019 at 1:03
  • Do reading courses count? For example, I did a reading course in probability theory from Rick Durrett's textbook because a probability course wasn't running this academic year. This helped me figure out I'm interested in probability theory and its applications, and hence I'm looking to move to an applied math/stats school. The other course was an applied math/numerical linear algebra survey course; the course itself wasn't as serious it could have been, honestly. I probably won't be able to get letters from other professors with whom I studied/will study measure theory and functional...
    – Nobody
    Aug 20, 2019 at 1:21
  • analysis, probably because I don't know them well enough as this professor (vice versa, more so).
    – Nobody
    Aug 20, 2019 at 1:22
  • 1
    Sure, reading courses count, if the sponsoring faculty person really had any idea what you did. Again, higher-level, more-recent, is vastly, vastly better than anything two years old and undergrad level. Aug 20, 2019 at 2:50

Totally normal combo.

Also, don't send them letters they don't ask for. Some universities only look at the asked-for documents, nothing extra, to make sure all applicants are on even footing in the process. There's a chance they will have to toss one of the recs in the bin and it might not be the extra that you want tossed.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .