Recommendation letters

Believe it or not, many professors write upwards of 50 recommendation letters each Fall semester. This means you cannot expect professors to keep track of your details or find out when & where to send these letters.

Instead, you should make it easy and super-quick for your letter-writers to write you a great letter, on time.

Specific 'How to ask for a letter' instructions - yes, these apply to you regardless of whether you are a postdoc or an undergrad

Please follow these instructions to the letter when asking for a recommendation letter from me. If you don't, I assume you don't really care that much about whatever you are applying for.

- Email me the complete information in one email, i.e. where you are applying, to get what, where do I send my letter, and what is the deadline. All of this needs to be repeated in EVERY email you send!!

- include the text of your application, your cover letter, or whatever document you have that tells me what you are telling them why you want this thing that you are applying for (so that I can write why you deserve it, really need it, etc.)

- include your latest CV or resume or something that reminds me of all the great things you've done

- send all this about a week in advance of the deadline. If you have it, send it earlier, but send a reminder email with all the same information again about a week before.

- send me a reminder on the day of the deadline (or the day before that), repeating all the same information. This is important, because it saves me from trying to find your old email, which, if I have to do it, will make it more likely that I postpone the task to a later time.

How to get stronger letters

- the kind of adjectives that a very strong letter has are: independent, enthusiastic, reliable, a good communicator, a good problem solver, creative, scholarly (understands research background, i.e. literature, well), well-liked, easy to work with, professional skills/maturity. I always have at least a paragraph on what is cool about your specific research. If you want me to use those adjectives, make sure they actually reflect your work & interactions in the lab. I always focus on everyone's strengths (it is really, really rare for me to mention any weakness in a letter), but (!) people experienced in reading these letters will nevertheless notice which words are absent. It cannot be overstated how important professional skills are - advisers want students who reliably read their emails (to the end), who come prepared to meetings, who seek and use feedback, and respect their adviser's time (e.g. by taking notes and using a calendar, not forgetting about meetings or deadlines nor asking the adviser to remind them repeatedly). 

- the longer you've been in the lab, the better your letter gets. This is because I'll know you better, have more example stories to tell on how you did well, and with experience your work quality will actually improve. So the top strategy for getting a really good letter is to work in the lab for at least 3 semesters.

- for undergraduates: many of you think of the graduate student or postdoc you are working with as an independent scientist, and you are asking them for letters. While that is certainly allowed, I do not recommend it for both your and their sake. First, writing lots of letters is an administrative burden that I want to protect my graduate students from. Second, graduate students and postdocs do not have a lot of experience either writing or reading such recommendation letters; thus their phrasing might seem odd to people who have, and they'll have a hard time calibrating what they want to say about you compared to what might be said about other students. Third, related to this last point, more junior people are often harsher in their judgments because they are more comparing everyone to themselves; as you get older, you realize more how everyone has their own path. I realize I am totally stereotyping here but this is my experience, as in I certainly learned this more as I got older/more experienced, and reading reviews and such and having conversations with my students I think there is truth to this. Lastly, whoever is receiving your letter will be more impressed with a letter on university letterhead from a more senior person, and even better if they personally know the letter writer and think highly of them. Whether you think this is a good thing or not, from a selfish point of view all this means that your recommendation letter should come from me not a graduate student or postdoc in the lab, with few exceptions. 

- if you are worried that I don't know you well enough, work on the following: (1) make sure your CV is very detailed. (2) Feel free to remind me in your asking-for-a-letter email about great stuff you've done in the lab. (3) Feel free to ask other students, graduate students or postdocs who know you well to send me a couple of bullet points on what your skills/strengths are, and/or particularly what events or activities or stories prove this. I will consider all of this information in my letter. (4) Make sure you schedule individual meetings with me to talk about your career goals and current activities; and make sure you speak up and participate actively in lab meetings. On the whole, I probably know you a good deal better than you think - remember that I am also asking your grad student/postdoc about you on a regular basis.

General tips on how to improve your application anywhere

- read the advice in 'applying' and possibly 'interviewing'

- if what you are applying for involves a proposal, statement, or some kind of longer text that will be judged by a committee or the people where you are applying, send me this early enough that I can give you feedback on it before you actually send it in. I may or may not manage to do this if the deadline is close, but it is always worth trying, because there are some standard phrases and standard mistakes that I can help you include/avoid, respectively.


Specific letters

NSF GRFP: in addition to the above, send your letter-writer this link to the NSF hints on letters.