...make the world go round. Alternative title: What you should do to maximize your chances of getting a good letter of reference from a pre-tenure PI*.
When you are a grad student or postdoc and you need a letter from someone it seems like a Big Deal. But it really isn't. As a PI, it is part of your job to write letters for folks that have trained with you (plus others that you interact with). In my short time as a prof, I have had to write a few letters. For folks that want to go to grad school, grad students (in and outside my lab) applying for training grants, and the occasional postdoc applying for fellowships or jobs. I am not (yet) the person that asks someone to "draft" a letter for themselves. I kind of find that disturbing, though I now understand why one might do this.
I try very hard to write a letter that accurately reflects the accomplishments and talents of the person for whom I'm writing. In a perfect world, I would make sure to mention all the things the referencee (who I'm writing the letter about) has done that would be appreciated by the granting institution. There are two main things required for me to do this: 1) I need to know all the awesome things the referencee has done and 2) what the granting institution is looking for. As a letter-writer, the more information that I get from the referencee that can help me fill in these blanks the better (with the best possible view of you, of course). In the best case, the referencee will give me the answer to ALL the questions that are asked in the instructions to letter writers. This is not as easy as it seems, especially from folks that are not actually in my lab**. There are different ways around this. When I was a postdoc, some folks had me "write my own letter". I HATED THIS. But now I can see how it is a great way to get all the info you need from the referencee. I didn't know what was supposed to be in a letter, and it is really, really hard to write a letter about yourself.***** But I was always very careful to address all the important points based on the application. I also had folks ask to provide a bullet-point list of my accomplishments. At the time, this seemed vague. But in retrospect this would be a great way to make sure that all the information that the letter writer needed was easily accesible. Other folks just wrote me letters. They (I think) knew me pretty well. We would talk before they wrote me a letter, and I sent them a current CV and a copy of my grant, so hopefully they had all the information they needed.
ANYWAY. Writing letters is not that hard. After you read a few you get a sense of what people are looking for. And honestly, it is part of my job that I don't get pissed about. Folks wrote letters for me, I'll write letters for students and postdocs that I train. But if you want to make sure to get the best, most accurate letter possible you have to help out the folks that are writing for you. Especially if it's not your primary PI. Here are some easy tips:
1. Send your current CV, your grant, and a copy of the funding announcement to your letter writers.
2. Include the instructions for letter writers.
Seriously, make sure they know what the grant/position is and WHY you are applying
3. If there is a bullet list of things the agency is looking for from letter writers, ask the writer if they would like a bullet list response. They may not use all your answers, but knowing how you see your past experiences and accomplishments fitting in can be very helpful.
4. Give the letter writer a friendly reminder by email or phoe. Don't go all crazy emailing every day. But you should send a reminder 2-weeks-or-so before the letter is due and again 2-days-or-so before it is due (depending on if they have turned it in).
*IME, YMMV (obv). Also, that second title sucks.
**because I can yell out of my office and ask them for info I don't have.***
***Also, I know them better
****REALLY HARD. In fact, I ran the letter by some faculty I trusted and many thought that I wasn't positive enough about myself. WTF?