NIH Reference Letters for F and K Applications: Who to Ask and What to Send

What Reference Letters Do in Review

Reviewers of F and K applications are scoring a person at least as much as a project. The research plan tells them whether the science is feasible. The letters are one of the few places where someone other than you and your mentor tells them whether you're the kind of scientist who will finish it. For a predoctoral or postdoctoral fellowship, that usually shows up in the candidate and training-potential discussion. For a K award, it feeds the assessment of the candidate's commitment to a research career and likelihood of becoming independent.

Letters rarely win an application, but a lukewarm or generic one can drag down an otherwise strong file. Reviewers read a lot of these, and they're fluent in the code: "hard-working and reliable" with no examples reads as faint praise. A letter with a specific story, such as the time you rebuilt a failing assay over a weekend or taught yourself a new analysis pipeline, reads very differently.

I can't give you a number for how much weight the letters carry in a given study section. It varies by panel and by reviewer. What I can say is that nobody has ever been helped by a weak set.

Who to Ask (and Who Not To)

Most F and K funding opportunities ask for a minimum of three references and allow up to five. Check the specific NOFO for your mechanism, because requirements differ and can change between announcements. Also check whether your sponsor or mentors are expected to submit their own letters as part of the application, since those are generally separate from referee letters and shouldn't count toward your reference total.

Aim for people who have watched you do science, not people who have a title. A famous name who met you twice writes a short letter that says very little. A mid-career collaborator who supervised you through a hard project writes three paragraphs of detail. Good mixes tend to include a former mentor from an earlier training stage, a collaborator or committee member who has seen you handle setbacks, and someone who can speak to independence, such as a person who watched you design your own experiments.

A few cautions. Avoid asking someone who is likely to be conflicted on your study section. Don't pad the set with people who'll all describe the same two years of your life. And think twice about anyone you're not sure will say yes enthusiastically. A hesitant yes is usually a weak letter.

How Submission Works in eRA Commons

You don't upload these letters. Instead, you enter each referee's name and email in the application form, and each referee receives instructions to submit their letter directly through eRA Commons. Referees need an account to do this, and registering a new one takes time, so don't assume it's a five-minute job for someone who has never used the system. If you haven't sorted your own login yet, our guide to the eRA Commons and Login.gov setup covers the account side.

Letters are generally due by the application due date, not a week later. A missing letter can leave the application incomplete, and NIH policy on late or incomplete materials has been tightening. If you're unsure how your institution handles it, ask your sponsored programs office early, and read the current NOFO for the exact rule instead of relying on what a colleague remembers from a prior cycle.

You can usually see whether a letter has been received, even though you can't see its contents. That status check is the single most useful tool you have, so use it.

The Briefing Packet That Gets Specific Letters

Most referees want to write a good letter and simply don't have time to reconstruct your history. Hand them the raw material, and the letter improves. Send one email with these items attached or linked:

The prompts are the part people skip, and it's the part that matters most. Something like "You were there when the knockout line failed and I switched to the conditional allele; if you remember that, it might help show how I handle setbacks" gives a busy referee a hook. Don't write the letter for them. Most referees find that off-putting, and it can read as inauthentic. Offer reminders, not text.

A Timeline You Can Copy

Work backward from the due date. About eight weeks out, decide on your list and ask each person in person or by short email, giving them a real out. Six weeks out, send the briefing packet and confirm they can meet the date. Four weeks out, enter their details in the application so the system sends their instructions early. Two weeks out, check the status of each letter. One week out, send a polite, brief reminder to anyone whose letter hasn't been received.

That might look like overkill. It isn't. Referees are often traveling, reviewing, or writing their own grants in the same cycle, and the most common failure isn't a bad letter, it's a late one. If a referee goes quiet, have a backup name in mind well before the last week. Don't wait until the final 48 hours to discover that someone's email has been bouncing.

Also keep your mentor in the loop. They often know which colleagues write quickly and which need a firm nudge, and they can approach a slow referee in a way you can't.

Mistakes That Weaken a Letter Set

None of these are glamorous, but together they're probably the cheapest improvement available in the whole application. If you're still deciding which mechanism to apply for, the fellowship and K award review criteria guide shows what letters are expected to speak to.

Frequently Asked Questions