- so that folks who do specialized work can be fairly evaluated if their work is specialized enough that those inside the department cannot evaluate well the work. I could have used such letters once when a political theorist was up for tenure and we had no theorists in the room who could tell us if the person did good theory and the school did not have outside letters at that stage.
- so that the folks up the line can determine whether the department is being fair or not. Any tenure/promotion decision goes through more than just the department's but also the dean', maybe the dean's committee's, maybe a university-wide committee, and then the provost, the President/Chancellor/Principal, and then the Board of Trustees/Regents/Pooh-bahs. These letters are used to evaluate whether the profession, as represented by the letter-writers, is either more or less supportive than the department.
The first big problem is that universities are expecting more and more of these letters to be written--that each candidate for tenure or promotion needs not five or six but ten or twelve ... because that is the way Harvard does it or some other b.s. argument. Some places are now asking for outside letters for folks going up for their third year review. I am pondering refusing to do those at all since it is a waste of time for all concerned. Few folks are fired after a third year review, and if they are, it is because there is nothing to read. So, asking for outside letters means soliciting letters that are going to be positive... unless there is nothing to read. Giant waste of time and a refusal by those universities to make decisions on their own.
The second big problem is politics. Ooops. That is, when people are deciding to vote within a department or within a committee, they may use the texts of letters selectively to support whatever case they want to make instead of using the letters to inform their votes. So, Stavans is correct in noting that "solid" can be used that way, that solid is obviously inferior to outstanding or special or crucial or some other more positive adjective. And this leads to the "tenure code" that Stavans identifies--that letter-writers have to be careful about what they say so that they don't doom a candidate. Which means that most folks write only positive letters and refuse if they cannot do so. And that becomes a problem since there are often good reasons to refuse--too busy, not enough of a background to evaluate a letter, etc.
The second problem has always existed and will always exist. The first problem is new-ish and can be handled if universities can solve the collective action problem of refusing to be sucky on this. There is no need for an arms race of letters--five or seven letters should be sufficient to evaluate the candidate. Doing more is excessive.
How do I deal with these problems? I tend to say yes. It is not weeks of work as Stavans suggests, but it is a heap of work to read a lot of a person's work. I did receive a request to review someone who was clearly going to be denied, and I knew that if I did not write that letter, someone was going to have to do so. Since I was clearly a relevant specialist for that hunk of work, I agreed to do it. I do think I should have exercised better judgment recently as I ended up agreeing to review someone whose work really is not in my research area. I will say no carefully and clearly when I am asked to do a letter when my plate is already overfull. The department may not believe me, but I cannot take on letter x when I am already doing a heap of letters. What constitutes a heap? Depends on the timing. I tend to average three or four of these each year, which makes me luckier than a few friends of mine who get far more requests. I guess being not so mainstream (IR of ethnic conflict, not war) in my work pays off.
Do I pay attention to every single word I write in these letters? Um, sort of. I am just not that careful, but I do try to be careful. I am willing to say something negative if the record calls for it. Which might mean my getting more letters in the future since some departments seek out negative letter writers.... but nearly all of my letters are positive. Why? In large part because I have been asked to review very productive people. And partly because I don't want to be the guy that ends someone's career.
To be clear, my basic standard is: has this person made a significant contribution? Are they likely to have a good trajectory? Getting into what that means requires much more time, which I don't have since there is a tenure review packet on my desk.
Update: see this piece for an excellent take on how to write these letters.