How to Write a Recommendation Letter

Today’s post is a special request post for all the people who have written in the past few months asking for a post on writing recommendation letters. A few of these folks have been letter-writers, but most of them are the sorry subjects of letters that they they were horrified to have the chance to see.  Others are those who have been asked to write their own recommendation letters by irresponsible and lazy recommenders. I disapprove of this practice completely.  Go ahead and tell me I’m wrong. I don’t care. Recommendations are to be written BY THE PERSON WHOSE NAME IS AT THE BOTTOM.

Anyway, recommendation letter-writing is really a little-understood art, considering the major role that these letters play in every scholar’s career. And if anyone thinks those letters aren’t really read—think again. They are taken very, very seriously, and pored over, and deconstructed, and discussed. And sometimes, acknowledged as works of writing in their own right. One time I overheard a colleague say to another: “I wanted to tell you that the letter you wrote for XXX was really terrific. It was a beautiful piece of writing.  Nicely done!”

So, what needs to happen in a recommendation letter?

I will first lay out a paragraph structure. This is just a suggestion. Obviously this kind of writing will be completely personalized.

  1. An opening that indicates warmth. “I am pleased to write this letter on behalf of xxxx.” Followed by explanation of who writer is, how they know the subject of letter, in what capacities, and for how long. Brief sketch of the subject in terms of discipline and topical/thematic focus. General comparison of the subject within the field/fields. “XXX is among the very top young cultural anthropologists of Japan, and promises to ….”
  2. Substantive description of major research work (diss or otherwise), describing the topic, but then, more importantly, providing critical context for the topic within a field or fields, or body of literature. The most important thing a letter can do is contextualize a candidate’s research for its larger scope, import, and promise. Many times on searches one of us would remark, “the advisor understands the candidate’s project better than the candidate does.” By this we were referring to the ability to see not just the trees of the project, but also the forest. This of course was not good for the candidate’s chances….but it does indicate a powerful and memorable letter.
  3. A second paragraph continuing from above. The most effective letter will reassure skittish search committees that the dissertation is in fact done or almost done, and/or that the book manuscript will be quickly forthcoming. Pace of work and productivity are valuable here. The key here is that the candidate be described as a (soon-to-be) scholar, not as a graduate student still enmeshed in the minutiae of a graduate program or dissertation.
  4. Evidence of the wider success of the candidate in the profession—the grant support, awards, and of course publications associated with the primary research. Conference activity can be mentioned here.
  5. A brief indication of next steps in research and publishing, so that search committees feel reassured that the candidate has a long-range plan.
  6. Discipline-specific attributes, such as experience in country of research, language fluency, technical skills, or other such things. Because I sometimes visited my Ph.D. students in the field in Japan, I had the opportunity to see them operate in their field settings, in Japanese. I always made a point to provide my “objective” account of their skills.
  7. Description of teaching abilities. This will be specific, mentioning course names, and methods used by the candidate, and departmental observations, evaluations, or feedback, to the extent the writer has this information available. Ideally the writer will have personally witnessed teaching by the candidate. If he/she has, this must be specified in the letter. Awards given for teaching should be mentioned.
  8. Service if applicable. One of my Ph.D. students was active in mobilizing the graduate students to overturn an outdated requirement in the program. The student’s effectiveness in this work demonstrated a real ability in departmental administrative responsibilities, and I made a point to mention it in my letter. One of the fears of search committees is that a candidate will be tiresomely focused on their research, and will be unable to assist in the running of the department. The letter can assuage those fears.
  9. A brief final remark about character and personality, insofar as these pertain to the candidate’s potential collegiality. This must be non-emotional and strictly non-gendered. Attributes to emphasize include resourcefulness, responsibility, good humor, organization, energy, etc. For women candidates it is essential to avoid anything that depicts them as any of the following: nice, selfless, giving, caring, warm, nurturing, maternal, etc.
  10. In sum, I expect XXX to have a career in the first rank of xxxx scholars in the country, and give him/her my unqualified recommendation. Please don’t hesitate to contact me at [phone number and/or email address] if you should require any additional information.”

Some basic expectations of the recommendation letter:

1. It will be at least two solid single spaced pages long.

2. It will be on letterhead

3. It will not gush or wax emotional

4. It will stay strictly at the level of evidence and substance

5. It will not rely on cheap and empty adjectives such incredible, remarkable, extraordinary, amazing, etc.

6. It will not damn with faint praise (“XX is one of the better graduate students we’ve had in the department”)

7. It will emphasize depictions of the candidate as a professional scholar, NOT a graduate student.

8.  It will remember to include the wider context, providing a discipline-level view of the candidate’s accomplishments and promise.

9. It will provide specific information or examples about research, argument, methods, teaching, or service–not vague generalities. The exception to this is (as noted in the comment stream): if you are in a position of a certain amount of obligation to the subject (as in, on their committee), but don’t feel able to write a detailed letter, and are also aware that the absence of your letter would send a strong but unwarranted negative message about the subject’s position in the department, and also feel positively disposed toward the subject in general, then write a generic letter that “first, does no harm,” but–and this is important–make sure the candidate knows that he or she needs another letter that will be more detailed.

10.  It will be unfailingly (although not gushingly) positive.  Any writer who cannot be 100% positive about the subject should not write a letter. In the event that you are ambivalent about the subject, it is better to tell the subject that you cannot write the letter.  A tactful method is to say you don’t know the record well enough to write a detailed letter.  

And indeed, if the subject is really problematic in your view, you are doing the subject a favor by stating that clearly, and not “enabling” self-destructive or substandard performance through misplaced “niceness” or conflict avoidance.

Bonus tip for European, Asian and other international letter-writers:  American letters of recommendation, like American tenure file external review letters, must be entirely and energetically and overtly positive (but without degenerating into gushing or encomium).  “Objective” and “realistic” are not qualities of this genre of writing.  As a department head, I had more than one tenure case almost derailed by European external review letters that very reasonably provided a “strengths and weaknesses” assessment of the candidate.  In the U.S. context, there can be no mention of weaknesses.  I’m not saying this is good.  I’m just saying this is true.  So international writers for the U.S. market: please be alert to this cultural difference, and have an American colleague review your letters when possible.  Candidates with international references:  consider sharing this post with them!

When candidates write their own letters, they typically fall afoul of #3 and #5 in particular. Remember: stay at the factual, don’t go emotional!  And #8, which, after all, no candidate is really positioned to write.

Now, some of you will wonder where the warmth comes in. Indeed, warmth is necessary for an effective letter. But in reality, the warmth comes through, even without a lot of cheap adjectives. Any writer who can speak with great care and thoroughness and respect about a candidate’s achievements is a writer demonstrating warmth about and investment in that candidate. Extra efforts to “sound warm” just end up muddying the message and in the case of female candidates, overly-gendering the profile in ways that do not work to women’s benefit on any professional job market.

 

 


Comments

How to Write a Recommendation Letter — 57 Comments

  1. Excellent… this is great from the grad student’s perspective. It’s quite helpful to see how we can “help” our referees craft strong letters.

    Your point of seeing your students using their language skills hit home for me. My adviser regularly uses terminology and phrases in their original language, almost never in English (though there are translated words/phrases that I’ve always used), and she shared a few stories of how she used those languages to “prove” herself as a competent scholar. After reading this post, it hit me that I’d better damn show her that I can capable of utilizing those languages in scholarly conversations too! Perhaps, not quite wait until I actually am in such a situation where my language skills may be tested in front of her.

  2. Quick question about when you cannot write such a letter, either because you don’t know a student well enough or your regard is not high enough. Is it best to demure rather than write a half-hearted or overly general letter?

    • When I get into that situation with an undergraduate I just tell them plainly that I can comment on this or that, but that they need a stronger letter than I can produce. Then I ask them about their contact with other professors and help them find someone else — or cultivate someone else. If it’s a grad student going on the job market, sometimes one really needs to write the letter since it will look strange if one doesn’t. Example: there was a woman on my dissertation committee that I needed to have a letter from so as not to raise suspicion but I had had only arm’s length contact with her (she moved to another university a year before I was done). So, she didn’t know me very well but I got a letter from her in any event suspecting it might say something to the effect that I was a fine student but that she just hadn’t had that much contact with me. To solve this, I got four letters (instead of three) so that she was not absent but the other three (I’m guessing?) were just more detailed.

      There are key people who must write a job candidate a letter and if for some reason they can’t write the strongest letter, candidates should just get more letters, I think. Indeed, who writes letters for whom can be highly political and I’ve seen PhD’s made terrified, miserable and unhire-able when key faculty on their committee refused to write them a letter of any kind. So be frank about what you can do but also respectful of the power you wield as a referee.

  3. Thanks for writing this, Karen!
    It is interesting that the suggested length of a recommendation letter exceeds the suggested length of the research proposal. That is not what I would have expected, so it’s good to know.

  4. Having written many letters (in physical science), I agree with almost all of your prescription. However, I wonder why you think it is the letter writer’s responsibility to describe items that will be found in the applicant’s own application materials, such as items 4 and 5 in the above. Do you think hiring committees read the letters and not the applicant’s cover letter and CV? This is not my experience with hiring committees – although it’s true that, in my field, applications tend to number in the dozens, not the hundreds. Of course, in writing letters, I give my own perspective on those items if I have access to information that the hiring committee may not have.

    • To me being able to describe the achievements shows that you are, in fact, keeping up with the candidate’s record–ie, care enough about the candidate to actually know his or her record well. You definitely don’t have to list ALL the publications or grants, just the highlights. Ie, “She is the only person in the program to win both the AAUW and the Fulbright…” That kind of thing. That lends credibility to the letter.

  5. I’d like to suggest a couple of do’s and don’t’s if the person you’re recommending is an adjunct colleague or member of your department. I recently got a full-time position after working as an adjunct at the same school for four years, so I am sensitive to these issues. I also got to see all of my letters of recommendation once I was hired.

    First, do NOT belabor the person’s adjunct status. Mention it ONCE ONLY, in the first paragraph, when you are describing how you know the person. After that, stick strictly to describing the candidate’s teaching, research, collegiality, and so on, as described in Karen’s post. Please avoid saying things like “Susie carries on an active research agenda even though she’s an adjunct” or “Jeffrey was regarded by his colleagues as an equal member of the XYZ Committee despite being only an adjunct.”

    Secondly, do NOT, under any circumstances, refer to how small the candidate’s salary is in your letter. Never say things like, “Susie’s diligence and generosity are clear because she worked hard to provide quality instruction despite making so little money as an adjunct.” I had a letter-writer who said something along those lines and I was mortified when I saw it. It’s like telling someone I am an intellectually stimulating cheap date.

    These types of comments send the wrong message, IMO. I know you may think that you’re praising someone’s ability to overcome adversity, or to make something out of nothing. But you aren’t doing that if you write those kinds of things. You’re actually diminishing the candidate’s value. You’re telling the search committee that this applicant is a superior adjunct, and I can’t help but assume that will undermine the candidate’s chances. As I said, I did get a job from the only interview I actually scored after sending out dozens and dozens of applications. I do wonder if I wouldn’t have gotten more interviews without letters that made those types of comments.

      • This probably won’t come up as much in a job recommendation letter, but for highly competitive fellowships please also do not belabor the fact that a candidate may have struggled with poverty or other difficult circumstances. It is fine to mention it — as noted above, abililty to overcome adversity is to be admired and many review committee’s are looking for economic as well as other types of diversity — but the primary focus of the letter should be the candidate’s stellar academic credentials, great potential (that, given your observations of their strong character, they are likely to live up to), and solid research question/ project design.

  6. Pingback: Weekly List Bookmarks (weekly) | Eccentric Eclectica @ ToddSuomela.com

  7. Pingback: Sunday Link Encyclopedia and Self-Promotion « Clarissa's Blog

  8. Pingback: How to Write a Letter of Recommendation | UCSD Graduate Funding Blog

  9. I have heard in the past that giving excessive praise in a recommendation letter can be function as a sort of code that the recommender is not actually in support of the candidate. It sounds a bit ridiculous, but it’s made me concerned about being perceived by readers as excessively praising a student and inadvertently hurting their chances. Any sense of this?

  10. Is the 2 page single line recommendation for graduate students the same as those for undergraduate students applying for grad school? Or is only for masters students applying for PhD positions?

    • I was definitely envisioning the letters that Ph.D.s need for jobs and postdocs and fellowships in this post. When you’re writing for someone applying to a grad program, a one page letter is often totally sufficient.

  11. This is a wonderful how-to guide. I read (and write) a lot of recommendations myself and can second your point that hiring, fellowship, and graduate admissions committees read them very carefully. I would say, however, that external letters for tenure and promotion files are a different genre. Those that read like standard recommendation letters (i.e., letters solicited by the candidate) aren’t given the same weight as relatively dispassionate “strengths and weaknesses” letters–at least in my experience.

  12. I’m a recent PhD going out on the job market for the first time this year. I’ve heard that if I give my permission, my supervisor can read all of my letters and, if need be, steer me away from using any “damning” ones. Anyone know if this is true?

    • Some departments have a mechanism or culture of doing that. It depends i9f you use Interfolio or such services. If not, and the letters are personal, then no your advisor can’t just ask for them to see.

  13. What are your thoughts if referees request the letter be drafted by the candidate and then make appropriate edits? As I am a first year direct-entry Ph.D. student, I am thinking specifically of letters that apply to competitive studentships and other scholarships. My advisor has informed me that it is common to write a draft reference letter for one’s referees that they can then modify (i.e. he has been very successful in regards to awards, etc. by doing this). . .

    A:)

  14. I’m on a search committee this fall, and we just got in a letter of recommendation from a recently retired Giant in the Discipline. This is the kind of scholar who has a five-page Wikipedia entry. The first line reads: “I am delighted to recommend Joe Schmoe for a position in DEPARTMENT HERE.”

    Maybe Expectation #11 is in order?

    Something like: “11. It will be bloody well proofread.”

    • Not only did I see that type of mistake in recommendation letters in our search last year, but also letters that had the incorrect name of the candidate leading me to believe the same letter was sent for all students working with professor to-busy-to-care. Also, in at least one instance the letter writer identified the students grad advisor incorrectly. I only knew this because the letter was for an applicant who worked with my former grad advisor.

  15. Your article was very helpful. By outlining your points using specifics and examples helped to drive home your intended meaning of each point.

    I have been searching for information and advice for an undergraduate student who is writing a letter of recommendation for a Professor who has been nominated for a teaching award. Any advice or tips for a letter with this relationship is greatly appreciated.

  16. Your post about recommentadion letters is very helpful and has a lot of information! It contains a lot of helpful tips when writting a recommendation letter. I agree with you that the person who signs the letter is the one who should be writting it, however, I also think that there are supervisors who doesn’t have the time or the interest to write a good recommendation letter for you. For instance, my current Ph. D. advisor has asked me to prepare my own recommendation letter and I really don’t think that is because he doesn’t want to provide me with a good recommendation letter but because he is really busy and he doesn’t have enough confidence on his English skills.

  17. Thanks for sharing this with us. I will be applying for my phd. I got my bachelor’s degree in 2006 and I just obtained my master’s degree from other university. For this phd application, I need three recommendation letters. Two of which will come from my master’s thesis advisor and my head of department where I did my masters degree. Can I ask my undergraduate academic advisor to write this letter?

  18. When a student is requesting letters be sent to multiple universities (PhD programs), should the letters be customized in calling out the name of the university and/or program?

    I could see this being true for instances where the writer knows someone in the program (does research with, etc.) but if not, can it be a general letter?

    • I always customized letters and generally found the best letters for candidates to my programs to have been customized. I believe that it’s one of the best things you can do for a candidate.

  19. Pingback: Why I am angry about academic references « Boggler Blog

  20. Very useful.

    My worst/strangest experiences in assessing reference letters:
    -addressed to the wrong institution
    -hand written
    -very strange (and sometimes inappropriate) translations of idiomatic expressions from other languages
    -too informal (‘he’s a good guy’)
    -on letter was written almost entirely ABOUT the referee (1.5 of two pages)
    I feel really terrible for applicants in these situation – who likely don’t know their letters were so out to lunch. I also know of at least one case, where I think a letter template was re-used which (mis) identified the applicant as a member of a very under-represented group. That was a significant factor in the person’s admission.

  21. I hope that you have not covered what I am about to ask elsewhere in the blog. What do I do if the job call asks for 3 letters of recommendation but I want to send four? Ideally I would be sending two letters from my two co-chairs, a teaching letter and a letter from an outside (senior) scholar. Can I go ahead and just send all four? Or do I need to count one out? In that case, do I throw out the teaching letter or the outside scholar letter for R1, R2 and SLAC jobs? Also, any advice on how I can point my letter-writers to this post without implying that I think they are not doing their job well?

    • The basic rule is send only what they ask for and no more in all things job doc related. However, having said that, it is quite common to send 4 recs when they ask for three so I’d say it’s probably ok. Sometimes, though, ads are very specific about not sending anything extra, and in that case, you’ll need to cut one.

      In terms of how to get them to read this… Perhaps you could say that you were asked to write a letter for an undergrad and in researching how to do it, you found this post, and what do they think of it…

    • Hi Gilberto! Since this blog is aimed at new Ph.D.s and their advisors this isn’t an issue that has ever arisen in my work with them. I’ll write on this, though, at some point since it is one of the mysterious tasks of a new assistant professor…

  22. hi,i am a Cameroonian undergraduate student,i need a recommendation letter before the 15th of march for a scholarship to study medicine in the USA.I went to my high school chemistry teacher for that because he was the leader of the OZONE club in which i was an active member but he told me to write it myself and he’ll only sign.please help me because i dont know how to go about it.thanks

  23. Hi Dr. Karen,
    I have a very morbid question. Recently a prominent faculty in our university passed away unexpectedly. He basically died over the weekend after saying goodbye to his lab the day before at a conference.
    My question is very selfish, I know, but I was wondering what would happen to supervisor X’s student in terms of getting references (of course, the immediate trauma would be worst of all). My supervisor is of the same age as the faculty who passed on and his passing really hit home in our lab and we were quite shaken up. My PI too seemed quite shaken.
    I don’t know what the correct approach to this is. I don’t want to live nervously and imagine the worst, but my PI’s letter is a huge deal for me because I’m working with her full time. And she is getting older. But I can’t tell her to write me one…just in case!
    Any thoughts?
    Thanks!

  24. Dear Dr. Karen,
    Can you explain/expound on why you wrote: “For women candidates it is essential to avoid anything that depicts them as any of the following: nice, selfless, giving, caring, warm, nurturing, maternal, etc.”. I concur, but I’m not sure I fully understand the reasoning (perhaps because these qualities are not objectively useful or relevant for a position, and therefore not helpful to the person applying. Is there a problem with people using these for women??

    Thank you–
    Elisabeth

  25. I have one query if anybody can help. I wrote a recommendation to one of my X-firm HR Director starting with “I am pleased to recommend….”. However, he replied that I should not be pleased to recommend as I am having enough experience to recommend.
    Can I know how to start a line in other way?

  26. I have written all but one LOR in my career (post grad school app). It was particularly draining when applying for k-awards from NIH, on top of the admin/writing burden, to have to do this task as well …. and make them all not sound like they were in the same voice .I’m currently non tt at R1, applying for tt at a slac. My letter writers for my current search have all requested drafts from me up front and my mentor told me that this is ideal because I can control the content. At least the job docs are less burdensome than a grant app. I agree with dr Karen’s sentiment though. I’ve written many LORs for undergrads for awards/positions never considered asking for a draft. I knew based on their conversations what impressions they wanted to give and I think all mentors should have the acuity to see this. For high stakes letters, like med school, I let the students vet, though I doubt they knew what to look for..

  27. Pingback: How to Write Good Letters of Recommendation | UCSD Graduate Funding Blog

  28. Dear Ma’am,
    Can you suggest me how to thank one of my Prof who has given me a excellent recommendation for graduate studies.

  29. I currently work as an adjunct at two community colleges and am applying for full time teaching positions. I have 2 strong recommendations (from the Head of my Dept. and PhD supervisor) and am just wondering if it would be better to have my third recommendation from the second community college where I teach but have very little interaction with the faculty, another faculty member at the same community college as one of my recommendations that I interact with on a regular basis or a short and sweet recommendation from the Dept. Chair in a 4 year college that I worked for many years ago. Thank-you.

  30. I had a dissertation committee member who had briefly supervised me after I completed my MA (until I dropped him due to sexual harassment but had to keep him on my doctoral committee due to a lack of enough faculty in my area), write in my letter that he supervised my MA (he hadn’t) on an article I had published but had nothing to do with my actual thesis. Luckily, I still got a number of interviews and a job, but I was shocked when my department secretary showed me the letter a year or two later.

  31. I’ve been in an investigator at a non-academic institution for 7 years. A couple years ago I started teaching as an adjunct and I am thinking of making a career switch now. Should my letters still be from my PhD advisor and committee member(s)? Or should I use someone from my current institution?

  32. Hi Dr. Karen,
    I read this post with interest. I learned that one of my recommendation letters is only one page long (Interfolio doesn’t let you see the content but tells you how many pages it is), which doesn’t even meet the basic expectations #1 listed here. Do you think it’s still possible that it is a strong enough letter?

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>