Writing letters of recommendation must be a fine art. Culture across fields and countries vary wildly in almost any respect, and this is likely to the case for LoR.

How/when do professors learn how to appropriately convey the value of the person they recommend?

  • As they write them, by trial and error seeing which students land the position. But this would require a large number of trials, given so many factors come into play (including circumstances that have nothing to do with letter writing)
  • As they read them, while sitting on admission committee. I am then almost completely oblivious to the procedures, but as above, many observation might be required before one can be confident on its ability to write a proper letter.
  • 8
    "Writing letters of recommendation must be a fine art." Bad assumption. Writing letters of recommendation has close to no priority, incentive, or time for exploration on the part of the letter-writer, IME. May 2, 2016 at 2:05
  • 13
    @Daniel: IME, I somewhat disagree with that. Not all LoRs are created equal. For instance, grad school letters are easy to write: basically they can say "Yes, the student is strong enough to succeed in a program like yours, as I know because..." When it comes to LoRs for faculty jobs for my own students or affiliates, I often spend several hours per sitting over several sittings to try to get something that will stand out from the pack. I also read a lot of these letters, so I know that while mine are at high end of time spent / length, there are many other such letters being written. May 2, 2016 at 2:14
  • 3
    While there is an art to it, it is also sad but true that many professors never learn how to write an appropriate letter.
    – Kimball
    May 2, 2016 at 3:40
  • 1
    I don't think this is worthy of a full answer since I don't know of this being done anywhere else, but my department will proactively train the postdocs as a group in this regard.
    – user4512
    May 2, 2016 at 3:51
  • 2
    Culture across countries varies so wildly in this respect that in many European countries LoRs are ignored, or have almost no weight in an application. May 2, 2016 at 5:00

4 Answers 4


Ideally, you'll discuss the first few letters you write with a mentor, who can offer feedback and keep you from writing an ineffective letter (getting advice certainly helped me tremendously). The same situation can come up again later in your career, for example the first time you write a letter for a liberal arts college if you are used to recommending people for jobs at research universities. If you aren't confident that you can write an effective letter in certain circumstances, then you should seek advice from someone with more experience, even if you are used to writing other sorts of letters.

Serving on hiring and admissions committees is extremely valuable, both in figuring out what people reading letters want to know, and in calibrating how strong the competition is and how much enthusiasm one should express in any given case.

Trial and error does not seem useful, at least in my experience. There's just too much randomness in the decision-making progress and too many unknown factors that differ between students.

Writing good letters also gets easier with age, as you develop more experience and have a broader basis for comparison. As an incoming faculty member, you haven't seen how careers progress over time, and you've never had the experience of finding out which of the smart young people you know will live up to or exceed their early promise and which will not. Of course you'll never be able to predict this with certainty, but over time you'll start to see patterns, and you'll be able to make more nuanced and illuminating comparisons in letters.

  • 3
    Mentors are important. Even if you have a pretty good idea of what it ought to sound like they have experience. I lucky in that my place has an official mentor and I have the informal support of several additional senior members of the faculty. May 2, 2016 at 1:39
  • 1
    I agree that trial and error will not work. The channel is just too noisy to get any feedback, you could spend years writing crappy letters and never know it.
    – Tom Church
    May 2, 2016 at 18:37
  • @TomChurch I assure you, "could" is too hypothetical for this situation. I wouldn't want to feed the academics who can't write an adequate letter of recommendation.
    – Nzall
    May 3, 2016 at 11:13
  • 1
    I believe this to be the best answer, providing information on how they DO learn. The answer with most votes is probably a good preamble to this and very informative.
    – Three Diag
    May 7, 2016 at 10:20

As with most professorial duties, it's a sink-or-swim process. There is no formal or official training they receive for this task. They're just thrown into the situation with no experience and are expected to perform well at it immediately. If they've seen other recommendations during their other duties (that they likely also weren't trained for) they can use that, otherwise they can try asking colleagues or do google searches. So for the most part it's a bunch of untrained people relying on other untrained people and crossing their fingers that the other guys were able to figure out a good approach.

  • 10
    This is by far the most accurate answer at this time. May 2, 2016 at 2:03
  • 2
    I pity anyone who tries to use what they find through google searches (I did about two weeks back, and quickly realized that the examples ranged from merely passable to absolutely atrocious). And the advice accompanying the examples was very superficial.
    – Ben Voigt
    May 2, 2016 at 4:11
  • Are letters of recomendation often a lot more about who sent them and less about the exact content? Almost any unambigiously positive text from someone whose opinion I respect would carry a lot of weight with me in deciding between candidates. Sophisticated rhetoric would perhaps be the icing on the cake, but I wouldn't necessarily expect an expert in any field other than English Composition to be able to impress me with prose. If I got a letter from Stephen Hawking that said merely, "this kid's a genius", I'd put a lot of weight behind that. May 2, 2016 at 12:45
  • 1
    @ToddWilcox Certainly the name of some professors carries tremendous weight. But very, very few people are talented and lucky enough to have world-renown talents in their references. In large part because very, very, very few people are world-renown talents. Most people are going to have references from people who are at best well-known by the few handfuls of people who work in their particular specialty. That'll be very helpful for getting post-docs with these people, but faculty job searches are likely to take you outside of places with these few people. May 2, 2016 at 14:59
  • If you're a professor, odds are you've had someone write some good letters for you. With any luck, they'll have read at least one of them, and know what a good one looks like. May 2, 2016 at 22:11

Before I entered academia, I had the benefit of having served on the review board for a major graduate fellowship program. The result of this was that I had the ability to review a number of applications, and the attached letters of recommendation. Mostly they were reasonably strong, but some were definitely better than others. Reviewing what these reference letters (some 250 over 3 years) gave me a lot of insight into how to write good letters of recommendation, and I've been putting that into practice ever since.

(It has also taught me that good letter of recommendation writing is not a skill commonly encountered in most countries other than the US and Canada.)

  • 7
    For what concerns other countries, the issue is not the skill, but how much the writer believes in LoRs. Speaking with a few colleagues of mine, from various European countries, the common sentiment is that we really don't believe in the utility of LoRs, and we wouldn't be interested in reading long LoRs. Moreover, writing LoRs for many people in Europe is seen as a favour to the candidate, not as a duty of the profession, and thus one is not willing to spend too much time on it. May 2, 2016 at 5:44
  • @MassimoOrtolano This also seems to be a matter of field. In math, I can see that most postdoc positions require at least 2 (often 3) LoRs and they are send directly from the writer similarly to how it is done in the US (and these certainly seem to be given high weight). On the other hand from reading some guidelines for some grants, this is not the case elsewhere, as for example one such stated to include at most 2 LoRs and they had to be included by the applicant themselves (and even being allowed to include LoRs felt "tagged on" in some way). May 2, 2016 at 9:08
  • 5
    Well, on the other side of the coin US-style letters are sometimes viewed as excessively adulatory and I have heard of committees that "normalize" for excessive praise. May 2, 2016 at 15:27

I've seen a few good ones that were written about me. I use similar thoughts and structures without copying and by supplying my own knowledge about the subject of the letter. In time as a professor, you will see letters written about others during your service on hiring, student selection, and tenure committees. You can borrow (without plagiarising) ideas and structures from the ones that make strong impressions on you.

  • 4
    I had to write one without ever having served on an admissions committee, and it was hard. I imagine it helps immensely to have been on the other side of the table.
    – user37208
    May 1, 2016 at 23:04

You must log in to answer this question.

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