Monday, January 17, 2005

Recommendation Letters

A few random comments as I read and write recommendation letters for various academic positions:

Back in the old days, a candidate would send a department a list of references and the department would send to each reference by postal mail a request for a letter which would be sent back by postal mail and followed up by a thank you sent again by postal mail. Faculty had a lot more secretarial support back then. This year I only got one such request, from a math department. Many departments have the candidates ask the recommendors to send letters directly by post and/or email. The best organized departments send the recommendors a link that leads to a secure upload page that puts the letter directly in the department's database.

Some misguided people like to send letters by post because they worry about the security of email and the electronic storage of their letters. Letters sent by post are far less secure. Copies of these letters must be made and these copies get left, on copy machines, in mailboxes in public areas, on peoples desks and in conference rooms. In any case it doesn't make sense to go crazy over security, any student with a little imagination can find a way to see their letters. Students: Don't do this. No good will come out of it.

There is an old saying "If a price is advertised as under $30 you can rest assured it is not $19.99." I take this saying into account when I read recommendation letters, particularly lines like "among the top half of complexity theorists graduating this year". In general I read letters more for what is not said than what is said.

"Strong potential" looks good for a fresh Ph.D. and the kiss of death for a senior candidate.

I ignore negative recommendations as probably personal issues. If you really dislike someone write a lukewarm letter. Seriously, if you don't feel you can write a strong letter for someone make up an excuse for why they shouldn't list you as a reference. Don't refuse to write if you get a request directly from a department. No letter is seen as a negative letter.

"I recommend" is a weak recommendation. "I very strongly recommend" is a strong recommendation. "I give my strongest recommendation" is a meaningless lie. "Don't hire this person because we plan to make him/her an offer and we want him/her for ourselves" is as strong as they get.


  1. I know somebody IN A DIFFERENT FIELD who managed to see one of her letters... it was not really worth it. On the whole it was positive, but there was a nugget of criticism which this person found unfair.

    She had to work with the letter writer for a few more months, and she really wanted to challenge this statement in the letter, but she could not do so because the letter writer believed that his letter had been secret. So she just kept it bottled up and tried to act like there was nothing bothering her. It probably served her right for breaking the guy's trust.

  2. On the student side (at least for getting into PhD programs) there's an "optional" waiver that we can sign that says "I know it's my right to read the letters when I enroll in the school, but I waive that right." I've always feared that by not providing such a waiver that I would send a message to my letter writer that I didn't trust them enough not to read the letter.

    Obviously, sending such a message would be worse for me than the pleasure of reading a letter, so I always sign the waiver; but I want to ask someone on the other side: Do you usually expect the waiver? How often are you given them? Would it change the letter you wrote if you didn't get it?

    My guess is that most students would sign them, if only because it looks formal and is just yet another part of all of the other paper work. Thanks...

  3. For me, I use the same basic letter for everywhere a student applies. I usually don't even check if the waiver is signed. Nevertheless I suggest signing the waiver as a sign of trust.

  4. Regarding the "waive/do not waive" check box on recommendation forms, it is my experience that Ph.D. admissions committees tend to heavily discount the value of a letter with "do not waive" checked. This is partially due to suspicion that the writer felt restricted, but more importantly, it is due to the suspicion that the student simply "went fishing" for letters and sent the best three he could find.

  5. Regarding "grade inflation" in recommendation letters, I recall reading a letter where the writer started out with a couple of sentences explaining his policy of not inflating things and being frank. That letter was taken as positive, even though the comments included praise and some real criticism. She was hired and her career since then has been rather stellar.

  6. So on a 1-4 scale, it looks like your translation from expressed quality to actual quality is roughly:


    With that kind of evaluation voodoo, why bother reading the letters at all?


  7. This comment has been removed by a blog administrator.