Professor School: Writing recommendation letters

Because we were never trained for this job

Introducing a new occasional feature on Science For Everyone: Professor School! ™

While folks outside higher education are sometimes confused when they learn that professors are wholly untrained for the job, if you’re on the job you know what I’m talking about. I think the most overplayed genre in writing about STEM academia is “advice.” We need to understand ourselves, and one another, and this job, and we make better decisions when we know more and have exposure to a diversity of wisdom. So that’s what I’m aiming to provide here. Even if I’m suggesting something, I don’t think of it as advice, I’m not telling you what to do¹, just telling you a bunch of ideas about what works for myself and other people, and then you can figure out what’s best for you and your students given your particular circumstances. Isn’t that what school is really supposed to be about, anyway?

just a pretty photo for no particular reason

Let’s start with a perennial feature of the job: recommendation letters.

One of the joys of being a professors, which simultaneously is a painful irony, is that if you do a good job of connecting with your students and mentees, and prepare them to be competitive for roles beyond your institution, that writing letters of recommendation could genuinely expand to be a full-time job on its own. This means, off the bat, you need to accept the unavoidable fact that you’ll have to ration the amount of time on recommendation letters. Months ago I was suggesting that it’s useful to know how work duties may be sorted into four categories: 1. you won’t do it; 2. it doesn’t need to be done well; 3. it must be done done well but isn’t your primary focus; and 4. it’s your highest priority.

I think it’s possible that depending on context, we could have letter requests that fall into each of those four categories, but in my opinion most of them fall into the third. It is okay to say no to doing a recommendation letter. Perhaps you might not be able to write a strong letter, they might not have given you enough advance notice. I think it is professional malpractice to say, “a lack of preparation on your part does not constitute an emergency on my part,” but it’s not healthy for oneself to perennially grant last-moment requests if you don't have the bandwidth for them.

We can’t spend as much time on them as they deserve (because they deserve our all!), so what matters for us is knowing what makes a strong recommendation letter look strong. If you ask people for advice on writing letters of recommendation, you’ll get a lot of very firm opinions that are often contradict with equally rigid opinions from others. I think most people write recommendation letters that they think, in their own opinion, are what they would want to see when they review applications. However, if you’ve even been on a search committee or a review panel or whatever, you grasp the idea that people read these things very differently from one another. What could be a strong letter of recommendation from one person might be perceived as a weak one by another. It might be worth understanding how your perception of a good rec letter shapes up with what other people consider to be good rec letters. Because your audience is other people!

That said, I think there are some elements that most people look for in a recommendation letter. These include:

• Who you are and what your relationship to the person is. Specifics are a good thing. They were enrolled in plant physiology with you, you attended SACNAS with them, you met in office hours on several occasions. How well do you think you know them?

• They want reassurance that the person isn’t a toxic mess or an asshole. You want to be very specific with anecdotes showing that this person is a responsible member of the community, shows respect to others, and is held in the good esteem of people who they work with. This might be hard to do if it’s just a student in a course that you’ve taught, but you it’s worth remarking on the extent to which you can vouch for their character. A lot of people use recommendation letters simply to look for red flags and evaluate them. I understand why. Be sure to put in the green flags. If someone has an actual red flag that means you shouldn’t pass them on to others, then that’s simply one less letter to write. The way to say this is, “I’m not in a position to write a strong letter for you.”

• People want honesty and sincerity. I think we all agree that a good letter captures the general essence of what it’s like to work with this person and the strengths that they bring to the academic community.

• In the US, the culture expects letters that are highly positive. I’ve heard that in other parts of the world, a good recommendation letters contains qualified praise as a wholly positive letter wouldn’t have an air of credibility. But in the US, the letter of recommendation culture is that if you choose to mention a negative, even if it has a positive twist to it, then most folks will perceive it as a ding. I think if you are writing a long, detailed, and nuanced letter, then you can address areas of personal growth over time that can land as a big positive, but realize how this might fit in with letters about other candidates that don’t have this kind of negative information.

• The fact that you would continue to work with this person if you could. You’d want them in other classes you teach. You’d love to keep them in the lab forever. You’re glad to see they’re moving on to bigger things and they will be missed. I think this is important because whenever I get a phone call on a reference check, this is always the last question they ask. Always.

• One thing to consider leaving out: Someone somewhere told me that we should avoid saying anything like, “If you have any questions, don’t hesitate to contact me,” because for some, that might code as, “I have negative things that I might say in a phone call that I wouldn’t put in a letter.” I wouldn’t read it that way, but apparently some people have in the past?!

• Some people expect a letter to hit a minimum length. Too many people have the opinion that taking the length of the letter is somehow reflects the esteem of the letter-writer. I think in most cases, spilling deep into the second page is probably fine. Some kinds of letters could be substantially longer. (For example, recommendation for tenure-track positions for someone you’ve worked with very closely, a student with particular circumstances that they want to you explain, a person who seems to be ideally suited for a particular job and you need to take the time to explain why you think that might be the case think). But in my opinion, a single-spaced letter spilling through a good part of the second page should work for many situations.

Here are some things to consider and idea how to navigate circumstances that may arise:

Students signing the waiver: You may not care whether the student waives their right to see the letter. But the person on the other side may well interpret the failure of a student to waive their right to see the letter as a red flag and an inability to trust the people they work with. This is a common piece of hidden curriculum for students to know what it means if they don’t sign the waiver. Be sure that all your students know how this might be interpreted on the other side. I think in many cases it would be good practice to give students a copy of the letter you send out, but I think they’ve gotta sign the waiver for their own sake.

Working efficiently: A lot of letter-writers are fans of processing student request for recommendations using an online form. For example, you could set up a google form asking students to provide their name, contact information, a copy of their CV/resume, instructions/link for completing the recommendation, the deadline, why they want this opportunity, and asking them to provide additional information or context to include. This helps avoid the back-and-forth of students sending you an email with incomplete information, then a subsequent back-and-forth. You can just respond to a request with the link, and go from there.

Ranking your the applicant to other students you’ve worked with: I think these categorical rankings should be launched into the sun, and if they miss the target, they would at least perish in the zero kelvin of outer space. About every five or ten years, I have written in a letter that this person is one of the most remarkable people I’ve ever had the privilege to work with, and then I provide extraordinary evidence to support this extraordinary claim. But otherwise, how do I compare this person to other people who I’ve worked with? I don’t. When a recommendation form asks you to check boxes to rank your trainee, some people take this more seriously than others. It’s just a gross outcome of game theory. If the system requires me to submit rankings, then what I do is give uniformly high ranks across the board, and if there is a comment section, then I indicate that I ask them to read my letter instead of looking at these rankings to let them know what I think. Also baked into these rankings is typically a comparison of “students from your institution,” which means that they also operating with biases typically against institutions that have more students with marginalized identities. So however I rank my students from Western Regional Directional University, they somehow will never get considered as highly as a student from Highly Ranked Prestigious University from the bean counters who are doing this categorically.

What if you’re writing a letter that someone else is expected to sign? I don’t really understand how or why this has evolved in our culture, but I know that in some labs, undergraduate trainees who work with graduate students and postdocs are expected to have letters written by their direct mentors, but then signed by the PI of the lab and not their actual mentor. This might be more of a thing with prestigious institutions and big-name PIs who would be more impressive on a signature line. I think this practice is bullshit, as I think the person writing the letter should be signing it, and on the receiving end, I’d rather have a letter from the person who actually knew the applicant.

What if they ask you to evaluate a person’s mental health? I’ve gotten something like this on occasion. It should happen approximately never. I tell them to eff off, and that you are not in the business of providing medical evaluations.

What if you don’t know a student that well but they really want a letter? What I do is simply tell them that I don’t think I have much to go on, and ask them if they might have others who would be a better fit. But if someone is in a bind,

What do you do if some students are hanging around you and it’s transparent they’re just grooming you as a future letter-writer? I feel like this is gross, but also see that this is how the game is played. I mean, students are often given the advice to schmooze up during office hours with the sole purpose of getting a better letter of recommendation. This transactional approach is what we need to get away from. Also, I think this dynamic favors students who have access to that piece of hidden curriculum and have the daring/gall/self-importance to use our time this way. This is a structural problem that needs a structural solution. In the meantime, I suppose you can be more explicit about how you use office student hours.

Should applicants be asked to write an initial draft or outline of the letter? I think not. For a few reasons. The big one is that people usually can’t write a better letter than you can. I mean, I’ve seen some letters that colleagues have written for me and they look way way more flattering and accurate than anything I’d dare to write about myself, and I think this is a near-universal case, and if you click through on that link, there are other reasons.

How do we know if your letters are good? I once had a colleague who had a reputation for writing bad letters of recommendation but didn’t even know it. Please don’t be that person. I think some people have chosen this route as a matter of weaponized incompetence, but I think even more are just oblivious. I’ve run some of my letters of recommendation by colleagues who I trust and they’ve given me some extremely useful feedback. Just putting it out there, you can do this too.

Is using generative AI legit? No, no it is not. Please don’t do this. I have to admit that I did this once, several months ago, when a student told me that they needed a letter to officially complete their job application for a position that they had already been working in for several months. They told me something like, “It doesn’t even have to be good and honestly nobody will even be reading it, but it needs to be in the system as received for HR to finish the process.” I didn’t feel that guilty using AI to flesh out a bunch of phrases and ideas into a letter, but then I spent so much time making it sound normal and making the former student seem like the good human that they were, I shouldn’t have even done it anyway. I don’t see myself doing this again, even if such a rare circumstance emerges.

What are some additional things that people should know about writing letters of recommendation? Are any of my opinion too off base here? Other essential element in a letter you suggest? Please leave comments!

1. I’ve been told by other authors to not be concerned by what’s in reviews. But that said, I do have a favorite review that was a failed attempt at being negative which said, “The language is so equivocal, and the author tries so hard to be fair to all perspectives, that the book ends up giving no actionable advice.” I’m glad even someone who didn’t like it still understood what I was doing. This job isn’t cookie-cutter, you’ve got to think and build relationships and adapt on the fly. This job needs less advice and more context for problem-solving.