Don’t ask for letters of recommendation up front

I’ve been sitting on this for a while, not wanting to piss off hiring committees who might evaluate my own applications. I have applied for over a hundred tenure-track faculty positions, and I’ve had over a dozen in-person interviews. One of the things that has always struck me is the incredible degree to which the time of anyone not on the hiring committee is devalued. This is true of the applicants’ time, but today I’m mainly talking about the time of their references.

It’s typical to request three letters of recommendation with an application, but some departments will request four or (rarely) five. In most cases, the application is considered incomplete and will not be considered until all of the letters are submitted. If one of your references has a hard time with deadlines, tough.

In most cases, your references will be fairly senior scientists, Associate Professor or above. It would be nearly impossible, for example, for your Ph.D. advisor to be an Assistant Professor, since that position typically lasts about as long as a Ph.D. In my case, all three of my references were full Professors, one a department head. In nearly every case, all three references will be highly productive scientists. Their time is valuable, not just in terms of salary but in terms of advancing science. Requiring letters of recommendation up front wastes it.

Letters of recommendation should be requested by the committee, and they should be requested after the first cut.

But we need them to decide who makes the first cut!

Bullshit. Bullshit bullshit bullshit. I have been on search committees for two faculty positions, and I have advertised for and ultimately hired three postdocs. Three quarters of the applications are not even close to qualifying for serious consideration. Either the applicant has no record of publications or demonstrated interest in the field they’re applying to work in, or they’re in their third postdoc and have published two junior-authored papers in crappy journals. One glance at their CV is enough to ensure that they’re not going to make the short list no matter what their references say.

But it will slow down the hiring process!

Uh-huh. The hiring process moves glacially slow in any case, since the members of the hiring committee have lots of other responsibilities. There are large periods of time between successive committee meetings when nothing is happening. If the references are requested at the end of the first committee meeting, with a deadline of two weeks, it’s not going to slow things down much at all. If you think it will, set the application deadline two weeks earlier.

But once they’ve written you a letter, additional letters are easy!

That’s true, but it still takes time. Ideally, your references are customizing your letter per position, at least to explain why you’re a good fit for that particular position (just as an example, I have applied for positions in evolutionary biology, botany, molecular biology, evo-devo/developmental biology, microbiology, genetics/genomics, phycology, biological limnology, integrative biology, and astrobiology). More likely, your references are simply changing the name of the department and university. In some cases, the letters will be sent by email; more often, your references will have to navigate the university’s online system to submit the letter.

Each letter (after the first) probably only takes a few minutes on average. Let’s say ten. A tenure-track position at a good university will receive hundreds of applications, depending on the school’s reputation, how broadly they’ve advertised (evolutionary biologist versus fruit fly geneticist, for example), location, etc. If we say 200 as an average, I’m confident that that’s a low estimate. Three references per applicant, 200 applicants, 10 minutes per reference; that’s 100 person-hours. 100 highly-productive-scientist-hours. If only 20% of them are read, 80 of those hours are wasted. We can dicker about the precise numbers, but I think it’s a fair, probably conservative order-of-magnitude estimate.

How long would it take someone on the committee to request references for the 20% or so of applications that make the first cut? About an hour, I think. One person-hour to send (using my numbers above) 120 form emails. Copy paste copy paste send. Copy paste copy paste send.

So if you’re heading up a search committee and you make the decision to require letters of recommendation up front, what you’re saying is that your time is ~80 times as valuable as those of the applicants’ references. That you won’t take an extra hour to save other scientists two person-weeks. Collectively, over the hundreds of positions advertised each year (I’m only talking about evolutionary biology here), it’s a massive waste of highly productive person-hours, most of which are ultimately paid for from taxes or tuition.


  1. says

    I totally agree with you.

    For the opposite extreme, I’ve recently served in two hiring committees at my university where the discussion of recommendation letters was explicitly and totally forbidden by the central administration! (We managed fine, for the record, although letters would have certainly helped. Also, it was frustrating to not be able to mention at all the letters that had been sent to me in support of candidates I liked.)

    For context: this is France, where hiring of tenured faculty (like everything else) is strictly regulated by national law. The law obliges committees to fairly evaluate all candidates based on the same detailed list of documents and doesn’t mention recommendation letters, so you see the problem here. Because of course most committees would like to see letters, so they typically find a way around the regulations in order to get them. University administration vary in the way they handle this, from creating official websites where you can have your letters uploaded in parallel to the official application procedure, to outright banning their use like they did here two years ago, in a sad show of preemptive and unnecessary ass-covering.

    • Matthew Herron says

      That (from a U.S. perspective) is bizarre. There are definitely problems with letters of recommendation. Just as an example, some PIs write over-the-top letters for anyone who’s better than mediocre, while some are more frank and measured, and it’s often impossible to know which one a lukewarm letter represents. But I can understand why committees want to hear from the people who are best positioned to comment on all the things that don’t show up on a CV but are crucial for evaluating someone as a potential colleague. Forbidding them seems likely to have all kinds of unintended consequences, for example an unfair advantage (or disadvantage if you’re a jerk) for applicants who happen to have met committee members.


Leave a Reply