I am applying to PhD programs in astronomy this fall, and I will need letters of recommendation. I have been working with a research group for about a year, have done a poster session at a major conference based on work I have done with this group and have written a piece of software that will be used in a collaboration with another group at a different university. In these examples I worked most closely with a post doc in the group, but the PI was aware of what we were doing. The PI is well respected and widely known in the field, and the post doc recently received a prestigious and well known fellowship.

I am unsure whether to ask the post doc (who knows me better) or the professor (who knows me but is also more established in the field) for a letter of recommendation. I am in the fortunate position that I have more potential letter writers than I need, so I would prefer to have only one from this research position to show the breadth of my other research experience. Would I be better suited by asking the PI or the post doc? Is it common to have jointly written letters?

  • 2
    Why don't you ask both ?
    – Our
    Sep 26, 2016 at 5:54
  • I have additional people that I would like to write letters for me and my understanding is that grad apps accept three letters.
    – user62376
    Sep 26, 2016 at 6:03
  • As far as I remember, though it can vary institution to institution, you have to provide at least 3 recommendation letter.
    – Our
    Sep 26, 2016 at 6:08
  • you should check this kind of obligation before applying for the program.
    – Our
    Sep 26, 2016 at 6:12
  • Also see academia.stackexchange.com/questions/20274/…
    – adipro
    Sep 26, 2016 at 18:45

1 Answer 1


Ask your PI to write the letter, and consider suggesting that the postdoc consult with the PI or even help write the letter. That way, you get the best of both worlds: the reputation of the advisor, and the specifics from the postdoc (who could be named within the letter).

  • 2
    It's not really the place of the student to tell the PI how to write the letter. Maybe the PI doesn't like working from someone else's draft and would rather talk to the postdoc. I agree with you that this is the best outcome, but it's not clear how to ask. Sep 26, 2016 at 12:37
  • Good point - I have updated the answer. Sep 26, 2016 at 14:04