How to Ask for Letters of Recommendation to PhD Programs

During last-year’s PhD admissions process, I was asked:

How should a PhD applicant go about asking for letters?

This post is inspired by that question.

In this post, I focus on advice for PhD applicants, though some of what I write is relevant in other contexts, e.g., to people applying for PhD fellowships or faculty positions. As with all my posts, what I write reflects my perspectives and may not be the best advice for everyone. For example, I am in the field of computer science, and practices in other disciplines might be different.

While I focus this post on advice to PhD applicants asking for letters, I also offer some suggestions for letter writers.

Terminology.

I refer to the person asking for a letter as “the applicant”; this applicant may or may not be a current student. For example, when I applied to the UC San Diego Computer Science & Engineering PhD program, I was working in industry.

I refer to the person from which the applicant asks a letter as “the letter writer”. When it is clear from context, I use the term “the letter writer” even if that person has not yet agreed to write a letter.

It can be uncomfortable to ask for letters.

First, an important acknowledgement: It can be uncomfortable to ask for letters. There is an inherent power dynamic in most letter-writing contexts; the letter writer is generally much more senior. In some cases, the applicant knows the letter writer well (e.g., the letter writer was the applicant’s undergraduate thesis advisor or their instructor in a small course). But, in other cases, the applicant does not know the letter writer well and, hence, there is less certainty in how the letter writer will respond to a request.

I doubt it is ever possible to be fully comfortable asking for letters, but I hope this post will help.

First, I hope this post demystifies some elements of the letter request + writing process. I hope that this demystification will make applicants more informed, and hence more comfortable, when asking for letters.

Second, I offer some concrete suggestions on how to ask for letters. I hope that these concrete suggestions also help remove uncertainty and, consequently, can help make the process of requesting letters more comfortable.

Third, because it can be nerve-wracking to ask letter writers for letters, I offer the following observation: Letter writers receive many requests for letters. An applicant might spend hours thinking about each letter request and each letter writer — e.g., wondering how the letter writer would interpret a letter request, wondering how the letter writer might respond, wondering how strong of a letter they would write, and so on. Such thinking can lead to a lot of stress.

But, because letter writers receive so many letter requests, the letter writers will likely not spend much time thinking about the types of things that might cause applicants stress. For example, if the letter writer agrees to write a letter, they will likely not spend time thinking about how the letter request was worded and will instead focus all their time and energy on writing the letter itself. And, if the letter writer declines (which is okay, more below), they will likely simply decline and move on, not thinking about the wording of the letter request further. Additionally, there are strategies for wording the letter request such that the applicant needs to worry less about how strong of a letter the letter writer will write (more below).

Who to ask.

When thinking about letters, one of the first challenges for many PhD applicants is the identification of candidate letter writers. I suggest keeping the following considerations in mind:

• Ask for help in identifying letter writers.

If possible, ask for help in identifying candidate letter writers.

If you have an advisor or mentor that you trust, ask them for advice. The advisor might be a school counselor, a faculty member, or a current PhD student at your institution or elsewhere.

As always, keep in mind that all advice is generally correct from the perspective of the person giving the advice, but that any advice you might receive might not be perfect for you; hence, my recommendation is to thoughtfully consider all the advice you receive, but not assume that all advice is correct for you. In general, and when possible, I suggest seeking advice from someone who has written and read many PhD application letters.

• Seek letters from people who have pursued the path that you are interested in pursuing. In the context of this blog post, that means asking for letters from people who have already received a PhD.

The above bullet is general advice based on the following observation: People who have pursued the path that you wish to pursue are generally aware of the challenges with that path and, hence, are likely to be in the best position to write a letter that comments on your ability to navigate those challenges and, consequently, your ability to succeed along that path.

To re-interpret the above advice in the context of a PhD application: The nuances and details of a PhD program are not always visible to people who have not pursued the path of a PhD. When possible, seek letters from people who have pursued a PhD and who can offer their perspective of your ability to succeed in a PhD program.

I know of instances in which letters written by people in industry are quite strong, but strong in a way that emphasizes aspects of the applicant that are important for industry and that do not comment on aspects that are relevant for academia (more below).

If you cannot find any letter writers who have pursued a PhD themselves, that is still okay. Those evaluating your application should keep in mind that letters from people who have not themselves obtained a PhD might be written to emphasize different things. And, if the people who know you best do not have a PhD, they may still be the best letter writers for you! In such instances, I encourage you to share with them the advice for letter writers near the end of this post.

Additionally, share with them that PhD admissions committees may be looking for attributes such as the following: creativity, curiosity, vision, written and spoken communications skills, self-drive and perseverance, and the ability to explore research directions even in the face of uncertainty. Vision is a hard term to define, but for the purposes of this post, I define it as: in what ways does the applicant want to change the world with the work that they will do? That is of course a very challenging question to answer! Thus, the letter writer doesn’t need to fully answer that question. Rather, the letter writer should keep in mind that PhD programs often try to understand why an applicant wants to obtain a PhD, and what they intend to do with the knowledge and research foundations gained during the PhD process.

In writing this section, I worry that what I write here could be interpreted as suggesting that letter writers with PhD are somehow better, as people, than letter writers without PhDs. That is definitely not true! My point is that those who have completed a journey (any journey!) are in the best position to comment on someone else’s ability to also complete a similar journey. There are journeys I have not taken — I have never spent much time in industry, for example. What this means, for example, is that I doubt my own ability to write a strong letter for someone seeking an industry position. For such a position, an applicant would be better off getting a letter from someone with more industry experience than me. That said, if I take time to be informed about the industry position, I can still write a strong letter. Likewise, I hope that the text above can be useful to people in industry who might be asked to write letters for PhD programs.

How to ask.

I often suggest that people use the following text in their letter request:

Do you have time and think you know me well enough to write a strong letter of recommendation?

I like this wording because, if the letter writer says yes, one knows that (1) the letter writer has time to write the letter and (2) that the letter will be strong. I also like this wording because it gives the letter writer multiple plausible reasons to decline — they could decline because they don’t have time, because they don’t know the candidate well enough, or, for whatever reason, they don’t think that the letter will be as strong as the candidate desires.

Of course, it is also important to be polite in the letter request.

If one doesn’t know the letter writer well, I would suggest also including a PDF of a CV/resume in the letter request.

The letter request should also include the deadline(s) for the letters.

And, to the extent possible, if one has communicated with the letter writer before via email, I suggest sending the request email from the same email account. Doing so makes it easier for the letter writer to look up past email communications, if such context is useful for their letter.

In the email, the applicant can also remind the letter writer of when they met, how they know each other, and so on, if there is a chance that the letter writer might not immediately remember.

Lastly, regardless of their answer — whether they agree to write a letter or not — I encourage the applicant to respond with a polite reply.

It is okay if they decline.

There are many reasons why a letter writer might decline to write a letter that are independent of the applicant themselves, so one should not take a decline personally.

For example, some letter writers may truly not have enough time. They may know that the candidate deserves a strong letter, and they may know that it would not be possible for them to write the type of letter that an applicant deserves in the time they have available. And, letter writers are people, too — they may have personal things going on in their lives right now that they are not comfortable talking about but that can impact their availability.

In some cases (like for scholarships), they might have already agreed to write a letter for someone else, and they may not be comfortable writing two letters for the same position (because they know that, for this scholarship, the committee will expect letter writers to explicitly compare two candidates if a letter writer writes two letters, and the letter writer may not be in a position to make such a comparison in a way that is fully fair to both candidates).

In other cases, the letter writer may know that an applicant is strong and deserves a strong letter, but they don’t have sufficient past interactions to fill a strong letter, and hence the letter writer might believe that the applicant would benefit from a letter from someone else.

In summary, there can be many reasons why a letter writer declines a request to write a letter, and many of those reasons are not a reflection of the applicant but are instead a reflection of the letter writer, their availability, and their knowledge of the applicant. Thus, it is very important not to take a decline personally.

When to ask.

Applicants should ask each letter writer as early as possible for a letter.

The “ask early” recommendation holds even if a letter writer has written a letter for that applicant before. It might be tempting to think that once a letter writer has written a letter for an applicant, it is easy and efficient to write a new letter. In many cases, that is not true! For example, if I knew an applicant for one year, wrote a letter, and then write a letter one year later, I will very likely write a new letter completely from scratch.

And, if an applicant is applying for a scholarship or fellowship or something else specific, the application probably has a rubric for letter writers, and hence I might write multiple letters from scratch for the same applicant, even if the letters are written only weeks apart.

Of course, if the applicant did not ask early — e.g., they did not realize that the deadline for letters was as early as it is, or for other reasons — they should still ask for letters and not self-reject their application.

What to share with letter writers if they agree to write a letter.

To help the letter writer write their letter, the applicant should send the letter writer all the materials that they intend to submit along with their application, including (likely):

• CVs/resumes.
• Academic transcripts.
• Any research statement or personal statement that the application requires.

I have seen situations in which an applicant does not wish to share their transcript — and in particular their grades — with the letter writer. What such applicants may not realize is that if they have poor grades here and there, and if a letter writer knows/sees that, they can often write their letter in a way that acknowledges the poor grades while also speaking to the applicant’s strengths. Said another way, the letter writer can write their letter in a way that supports the candidate as a whole.

Often, applicants only have limited space in their statements, and they inevitably have to omit things that they wish that the admissions committees would know. Thus, the applicant might also explicitly share with their letter writers:

• Information about themselves that they wish they could have included in their statements/materials but were not able to due to space or other reasons.

The above information could include highlighting certain activities on their CVs/resumes, describing their roles on certain projects in more detail, more information about their personal interests/motivations underlying the contents of their application statements, or many other topics.

The applicant might also share:

• Valuable to know and “unusual” personal circumstances, if relevant.

If there is additional information that might influence the letter writer’s ability to write a fully-informed letter, I encourage the applicant to share that information too. For example, if there was a personal emergency one academic term or one year that resulted in low grades, if the letter writer knows that information, then they can speak to it specifically in their letter. Of course, the applicant should only share information that they are comfortable sharing. And, if they would like the letter writer to know some information but not share that information in their letter, the applicant should let the letter writer know their preferences.

The applicant might also share:

• Information about themselves and their interests beyond the “obvious” scope of the program to which they are applying.

In my experience, while PhDs are granted in specific disciplines (e.g., computer science, philosophy, molecular biology), PhD students often benefit from leveraging their own unique, diverse backgrounds. There can be significant value in having cross-disciplinary expertise and exposure. Thus, I encourage applicants to also share information about themselves that might not seem directly related to their PhD program. Once letter writers know this additional information, they have the option to reference the information if they believe that it will complement and strengthen their letter.

Of course and as always, applicants should only share information that they are comfortable sharing.

Additionally, the following will likely be helpful for the letter writers to explicitly receive in an email:

• The application deadline(s).
• The submission URL(s) or other information how to submit the letters.
• Any specific information regarding the “thing” to which the applicant is applying.

The last bullet may not be too relevant for PhD applications. But, if a student is applying for a scholarship or another award, the web page for the award likely says something like “Recipients are selected on the basis of <criteria 1>, <criteria 2>, and <criteria 3>”. Sharing those criteria, or the relevant sentences from the web pages, will help the letter writers target their letters to the particular award.

Further, the applicant might share:

• Concrete, specific information about any past activities or projects that the letter writer might know about. For example, that past activity or project might be: a class project, a research project that they participated in, a student activity or a club that they had some role in.

The concrete, specific information might include a description of the activity or project and the applicant’s role. If the applicant’s role in that project could be used to highlight a strength of the applicant, then it might be valuable to draw the letter writer’s attention to that strength explicitly. For example, if the applicant’s role in that project could be used to highlight the applicant’s creativity or their curiosity or their ability to navigate challenges, that might be something to highlight.

It is okay not to have done research before applying to PhD programs.

It is okay if an applicant to a PhD program has not already had much research experience, and it is okay if the applicant has not published research papers already. Not everyone has access to research opportunities as an undergraduate student! For example, someone may be at an institution where there are not many research options, or someone may be at a large institution with many active research projects but also many other students seeking research opportunities and hence few availabilities, or perhaps someone did not realize until later in their undergraduate career that they were interested in research.

I strongly encourage students interested in pursuing a PhD to not self-reject their own PhD applications if they have not had research experiences yet but, instead, to ask for guidance and advice from mentors — especially those within academia — on the application process.

The application statement as well as the letters of recommendation will play an important role in the PhD program’s decision. If a letter writer is from academia, they should know how to write a letter of recommendation for a PhD program even if the applicant has never done research before. And, if the letter writer is not from academia, I encourage them to read the end of the “Who to ask” section above for more information about what PhD programs might be looking for in letters.

Of course, I should acknowledge that some graduate programs may value past research experiences more than me. Still, I would consider it a mistake to not fully consider all applicants, including applicants without past research experience, and I know that it is possible to write a strong and accurate letter of recommendation to PhD programs even if an applicant has never done research in the past. It is for these reasons that I say that it is okay to not have prior research experience.

I think one of the biggest questions someone might have if they have not had prior research experience is: what, exactly, is a PhD? And, what, exactly, does one do during a PhD program? Exploring these questions is not the topic of this post, and I encourage those without past research experience to learn more about what a PhD program is, if they don’t already know. While not a substitute for talking with one’s own academic advisors and acquaintances that are already in or have been in PhD programs, the following prior posts on this blog might help hint at what is involved in a PhD program, which largely focuses around research.

• My post on the unseen efforts and challenges within the research process, and strategies for navigating those challenges
• My post on how it can sometimes be challenging to make progress with research, at least for a common definition of “progress”, and how to navigate those challenges.
• My post on the research mosaic analogy for a dissertation, and how a dissertation might be composed of multiple independent research projects that, at first, might seem disconnected.
• My post on elevator pitches, which implicitly also discusses some about what PhD research entails.
• My visual illustration on the different types of PhD dissertations.

Staying in contact.

It can be very meaningful to a letter writer to hear from the people that they have advised and/or written a letter for. If they applied to PhD programs, did they get accepted? And, if so, did they decide to pursue a PhD and if so where? If they applied for a scholarship, did they receive it?

Thus, while not strictly necessary, I encourage applicants to keep in contact with their letter writers as time progresses. The people who write your letters today might become your colleagues and peers in the years to come.

And, there are also indirect benefits to the applicant if they stay in contact with a letter writer. For example, if an applicant applies to a PhD program in one year, the following year they might apply for a graduate fellowship, in which case they may wish to reach out to past letter writers and ask for new letters.

Planning for the future.

Sometimes, one knows that they may apply to PhD programs in a few years but not immediately. For example, a graduating senior might plan to work in industry for several years and then later apply to PhD programs. In such situations, it would not be unreasonable to send a “heads up” email to potential letter writers during their senior year, letting those letter writers know their plans and their intent to apply to PhD programs in future years. Those letter writers might then take notes for themselves at that time, to make the process of remembering details and writing a letter in the future easier.

When the applicant later applies to PhD programs, they can remind the letter writer about their past interactions and send updates about their activities in recent years.

If an applicant knows that their university email address will disappear, they might send the initial email — the email about applying to PhD programs in future years — to the letter writer from their university email address and send a carbon-copy (CC) to their personal email address. That way, they can reply to the same thread from their personal email address years later.

A note to letter writers.

I now turn to some generic advice for letter writers — not just for the writing of letters for PhD applications, but in general

• Read and write to the rubric. For many applications — e.g., for scholarships and fellowships, there are implicit rubrics. For example, I recently looked at the 2024 Alfred P. Sloan Foundation Fellowship selection criteria page, which says “Fellows are selected on the basis of their independent research accomplishments, creativity, and potential to become leaders in the scientific community through their contributions to their field.” This text suggests that the selection committee will evaluate “independent research accomplishments”, “creativity”, and “potential to become leaders in the scientific community”. Thus, if I were to write a letter for the Sloan Foundation Fellowship, I would make sure to cover at least those three elements in my letter.
• Decline politely. If you must decline for any reason, remember that it likely took effort and courage for the applicant to ask for a letter. Thus, decline politely. And, if there is any additional advice you might give the applicant, do so — e.g., advice on other scholarships or PhD programs that they might consider (if you know of any other options), and so on.
• Make it okay for the applicant to send reminders. I try to remind applicants that it is okay to send me reminders as the deadline approaches. While I try not to need reminders, I know that applicants can get nervous as deadlines approach. Thus, in my reply email — if I accept — I make it very clear that it is okay (and in fact appreciated) for them to send me reminders as the deadline approaches.
• Respect confidentiality. It is important for letter writers to keep in mind that they may know confidential/private information about the applicant. Err on the side of being conservative and only mention or refer to confidential information with permission. When asking for that permission, make it clear that it is okay for the applicant to not grant that permission.
• Read the materials that the applicant provides. A letter writer should read the materials that the applicant shares with them. That way, the letter writer can ensure that their letter aligns with the stated interests and directions of the applicant, for example. As another example, by reading the CV, the letter writer might spot items that they can highlight in their letters even if the applicants did not highlight those items in their submission materials.
• Ask for additional information and make it okay to “brag” and not be humble. I encourage asking the applicant for any additional information that might be useful for the evaluator (e.g., the PhD admissions committee, the fellowship selection committee) to know but that the applicant might not have had space to include in their submission materials. Additionally, I suggest making it acceptable for the applicant to brag about themselves as they share this additional material. For example, one might explicitly ask the applicant to not be humble in the information that they share with the letter writer.
• Be descriptive and give concrete examples. Rather than use generic adjectives, like creative, curious, self-driven, and so on without context, provide concrete supporting examples, when possible.
• Be aware of biases. Numerous resources online point to common biases in letters of reference with, for example, letter writers using different terms for women and men applicants. Here are links to some resources to familiarize oneself with and to review before writing letters, though I encourage letter writers to look for resources beyond just those that I link to here. These online resources include the Avoid Implicit Gender Bias in Recommendation Letters page from Columbia, the Avoiding Bias in Recommendation Letters page from Georgetown, and the Avoiding Gender Bias in Reference Writing page from the University of Arizona.

Local resources within the University of Washington Allen School.

For undergraduate students in the University of Washington Paul G. Allen School of Computer Science & Engineering, if you are early in your undergraduate career and are interested in exploring research, I encourage you to consider the course CSE 390 R. This is a web page for CSE 390 R that should persist over time: https://courses.cs.washington.edu/courses/cse390r/.

Acknowledgements

Thank you to Leilani Battle and Maya Cakmak for all their insights and discussions about an earlier version of this post. I learned so much about the PhD process from my own advisor, Mihir Bellare. After obtaining a faculty position, I continued to develop my own philosophy on advising and the PhD process through the advising my own PhD students. Much of my thoughts on advising and the PhD process, including letters, have also been shaped through the co-advising of PhD students with UW Security Lab co-director Franziska Roesner. Thank you to all the students and postdocs that I have advised, past and present!