A Five-Minute Guide to Ph.D. Program Applications

I've now served on Ph.D. admissions committees for the past three four years as a professor at two different universities (and in two different kinds of departments: computer science and cognitive science). So far, I've evaluated around 600 Ph.D. program applications across both kinds of departments, which means reading 600 statements of purpose and 2,000 recommendation letters (3–4 per application).

[Update in Dec 2020: in the years since this article was first written, I've probably read a few hundred more applications and rec letters.]

This article should take you five minutes to read, which is probably the longest that anyone will spend reading your application. When I serve on admissions committees, I read around 150 applications, so I'll spend 3 to 5 minutes per app and place each into one of three piles: {good, ok, bad}. Then I may tell faculty in the relevant research sub-field to take a closer look at some of them.

Final admissions decisions are made in very different ways depending on department norms and constraints, but that's all out of your control. No matter what, though, you want to end up in the 'good' pile!

What is a Ph.D. program application?

The first step toward getting into the 'good' pile is to understand what your application really is before writing it:

A Ph.D. program application is a request for someone to invest around $500,000 and five years of mentorship time so that you can produce new knowledge via research publications. The exact economic details vary by department and institution, but the general idea remains.

When evaluating applications, I ask myself: Do I feel good about letting myself or one of my colleagues invest in this person – both in money and in time? The role of your application is to convince me to enthusiastically answer YES!!!

Concretely, this means your application needs to show me that you have either already produced research publications in your chosen sub-field, or that you have the potential to do so in the near future. Everything else is secondary.

The most important criteria: Research Density

In my experience, the property that best separates good from bad applications is research density. As the name implies, a research-dense application is one that is densely-packed with research-related content. The strongest applications are usually the most research-dense.

What this means is that the majority of your statement of purpose should be about research – not your childhood inspirations, not your personal intellectual journey, not your classes, and not your extracurricular activities. A common kind of weak statement is one that spends too much time describing the applicant's childhood; in general, avoid mentioning anything before college.

I often read weak statements that are way way way too long and not research-dense enough. Strive to write the shortest and more research-dense statement that you can.

This idea of research density also means that you need to find people who can write research-dense recommendation letters for you. For example, a letter about you doing well in a class is not compelling, since it has zero research density. A letter from a non-research-related job also has zero research density. All else equal, applicants with more research-dense letters will win. So choose your letter writers carefully.

Addendum to research density: properties of weak Ph.D. applications

The weakest Ph.D. applications are those that are the least research-dense. Some common patterns here include:

• Statement of purpose is more of an autobiography of how you developed an interest in topic X rather than talking about how you've done concrete research in topic X (in general anything that mentions what you did before college is bad)
• Statement of purpose is more about classes and class projects than formal research projects
• Statement of purpose is way too long compared to everyone else's (and may be double-spaced) ... way too long probably means that it meanders to talk about less-relevant topics
• Statement of purpose is more about your high-level observations of the research field in general terms but not about your specific projects
• Recommendation letters are not about research (e.g., letters from class instructors or from managers of non-research jobs)

How do I review Ph.D. applications?

The people who review Ph.D. program applications are professors like me, not professional admissions counselors. Some of them will be knowledgeable about your research area, and some won't.

Here's how I spend my 3 to 5 minutes evaluating each part of the app, roughly in this order:

1. C.V. / resume (30 seconds max.)
   • Is it research-dense?
   • Is it not too long? I'd say 2 pages max, maybe 3. Don't fill up your C.V. with irrelevant details.
2. Statement of purpose (1 minute max.)
   • Is it research-dense?
   • Is it not too long? My ideal is 1.5 to 2 pages.
   • How well do you understand what a Ph.D. application is? Re-read the prior section for details.
3. Scores (30 seconds max.)
   • GPA: warning sign if too low, but usually don't care. It's rare that someone with strong research credentials has a dangerously low GPA, and even if that were the case, I wouldn't care much. A high GPA from a highly-selective university is a slight positive signal, though.
   • GRE: bad if really low, but otherwise don't care.
   • TOEFL: bad if below 100, but otherwise don't care.
4. I'll reject here if the C.V. and statement aren't strong enough. Otherwise I'll move onto reading the letters ...
5. Letters of recommendation (1 minute max. per letter)
   • Is it research-dense?
   • If a letter isn't about research, I'll skip it.
   • How long is it? Two-page letters are stronger than one-pagers. You can't control this, but you can control whether you pick someone who will write you a strong letter.
   • How much effort did they put into differentiating you from the hundreds of other applicants this year?

[Update in Dec 2020: I'm more busy now and we get more applications than ever ... so on my first pass I read only the letters and reject if they're not research-dense. Then if letters are good, I'll read Statement of Purpose. If all of that is good, then I'll more seriously consider the app.]

Final thoughts

If you create a research-dense application, that's the first step toward getting into the 'good' pile. You'll now be competing with tons of other applicants who also have research-dense applications, but at least you're still in the game.

The next most important thing is to choose a sub-field within the department that is most relevant to your prior experiences and to your letter writers' expertise (especially your primary letter writer who will write the strongest letter). For instance, if you've done the most significant undergrad research in sub-field X, mark down sub-field X in your app if you get to choose; that way, you will be placed in the same pool as others in sub-field X. And your app might get reviewed by faculty in X. If you choose your sub-field wrong, then you will be competing against students in another sub-field whose apps look stronger than yours because they have more prior work in that sub-field than you. (Some students did undergrad research in X but want to change to sub-field Y for their Ph.D. In those cases, I do NOT recommend listing Y. If you put Y down on your apps, then you will not look competitive at all since you don't have prior work in Y, and your letter writers are also not from Y.)

Even after you pick the best sub-field match for you, the next most important thing to work on is making your writing clear for professors who are not in your sub-field. Of the hundreds of apps I've seen, I'm not an expert in most of their research. Thus, it's crucial for you to show me why your research projects are important, what specific role you played in each one, what the main challenges were, and what the impact was – ideally a publication or steps toward one.

Finally, it's OK to cold-email professors whose research genuinely interest you – after you submit a research-dense application. Make sure to write a good email customized for them. We usually review applications in the month or two after the due date, so that's the best time to cold-email. At best, you get someone to notice you; and at worst, they'll just delete your email. (In the U.S., it is convention to refer to professors by "Professor [last name]", like "Professor Guo". This convention is likely different in other parts of the world, though.)

OK, five minutes are up; good luck!

Appendix A: Related Resources

• logistics of asking for recommendation letters.
• PhD Admissions Advice by Prof. Casey Fiesler (HCI / information school)
• Getting a Computer Science PhD in the USA: PhD Application by Prof. Shriram Krishnamurthi

Appendix B: random notes added in Dec 2020

Random tips for Statement of Purpose:

• Two pages at most (that's the convention in my field, at least).
• Don't use tiny fonts. The faculty reading your app are a few decades older than you, and our eyesights get worse with age.
• Don't use tiny margins. Stick with 1-inch or a bit smaller than 1-inch.
• Single-space, not double-space; double-space makes it look like an amateur class report.
• Don't cite a bunch of papers that aren't your own; that just takes up space. Cite only your OWN papers, and bold your name in the authors list.

Also I've noticed a trend lately of Ph.D. apps adding more supplemental questions. My advice is not to stress about those since your Statement of Purpose is far more important. Just don't write anything in supplemental questions that could be interpreted as a negative signal and you'll be OK. You won't ever get into a Ph.D. program due to supplemental questions, but it can hurt your app if what you write reflects negatively on you.

Random tips for interviews:

Back in my day (the stone age) students would get admitted without interviews, but nowadays the bar is a lot higher, so interviews seem to be a standard part of the process. My only advice is to PREPARE A LOT. It might seem like professors just want to casually chat with you about your research (“hey can we find a time to chat? lemme know, k?”), but trust me, it is 100% an interview. Take it super-seriously. They are interviewing the top few candidates on video chat and deciding who to make offers to. All else being equal, if you prepare more than your peers, then you'll have an advantage; simple as that. Here are some ways to prepare:

• First off, know every facet of your resume, application statement, and publications inside-and-out. If you can't even answer questions about what you worked on, then you're toast.
• Learn about the broader implications of your research beyond the specific implementation tasks you did on it. Professors understand that many undergrads participate only in limited ways in research, but even if you didn't get to direct the high-level course of your project, you should still know the context surrounding it. That shows you have potential as a researcher, not just as a lab technician.
• Read through a few recent papers from each professor you're interviewing with, especially those that relate to your own work. Professors will usually have lots of papers on their CV, so pick some of the most recent ones (since it's the freshest on their minds) and also try to pick those where they are the last author rather than a middle author (since that likely indicates it's their own lab's papers rather than an external collaboration). Exact authorship conventions vary by field, so use your judgment here. Basically you want to pick papers that they really care about, not those from secondary outside collaborations where they didn't do much work on it. Remember, last-author papers are often the ones from their own lab that they care about the most.
• Search the web for what other types of things this professor has written or been interviewed about online. Also, explore their academic home page to see what articles or blog posts they've written or podcasts they've been on, separate from their main research publications. This could give you a more well-rounded sense of what they care about. (Note that if their home page looks really really really out-of-date, then that might be a sign that they have not been too active in research lately, so maybe they are not up-to-date on the latest work.)
• Even better, if you can watch recent talks or other videos by the professors you're interviewing with, that will both tell you about their research (good!) and, more critically, give you a sense of how their vibe is (e.g., their tone of voice, mannerisms, and general style) so you won't be caught by surprise when you speak with them.
• Always always always write a polite and enthusiastic follow-up email after each interview, summarizing what you discussed and expressing your personal interest in their lab and Ph.D. program (if you really are genuinely interested).

These above bullet points do take work, but that's exactly how you will stand out amongst the other applicants that they are interviewing! There's no shortcuts here :) Trust me, professors will be able to tell who has done their homework sincerely and who hasn't.

Appendix C: random notes added in Nov 2023

Super-messy notes jotted down in a hurry, but hopefully somewhat helpful ...

Here's a simplified (but somewhat realistic) model of Ph.D. admissions:

At every department you apply to, you are competing for a very small number of 'spots' allocated by advisors in your research area who are taking on new students in the coming year. Usually every advisor may have anywhere from 0 to 4 spots (super-rough approximation), and there may be up to ~3 relevant advisors in each department. So you're competing for around ~10 spots against others who are applying in your research area. (Exact numbers vary, of course, but this is one rough estimate.)

This is very different from, say, undergrad college admissions or any other graduate program admissions (such as masters programs, medical school programs, law school programs, etc.). In those programs, you are competing for one of a 'large-ish' number of spots in the entering class; there is no 'advisor' or anything to be matched with. For instance, if you're applying for undergrad admissions at UCSD, you're competing for one of, say, 5,000 spots in next year's class out of maybe 20,000 applicants (I'm making up these numbers). Or if you're applying to Harvard Law School, you're competing for maybe 100 spots out of maybe 3,000 applicants this year.

In contrast, when you apply to a Ph.D. program, you are not competing for, say, 100 Ph.D. openings in the department out of 1,000 or whatever applicants. That's the wrong way to look at it. Instead, you're probably competing for maybe 10 Ph.D. openings in your immediate research area out of however many applicants are applying in that area. Thus, overall 'admissions rates' don't mean anything, since if in your research area there are few to no openings this year, then even if you're the BEST applicant you may not get admitted.

And what determines how many openings there will be in your research area? It all depends on which faculty advisors are willing to take on new students in the coming year. This is 100% totally out of your control. You have absolutely no idea whether Professor X or Y or Z is willing to take on new students or away on sabbatial next year or going on an unpaid leave to start a company or planning to retire or move into an administrative position or whatever. So if you don't get into a particular department's Ph.D. program, it may not be because your application was bad ... it might just be that no faculty are looking for students with your specific research background, or there is someone else applying who is a better fit for their lab. Don't take it personally!

OK, so what actionable tips can I give to you, given this reality? Here is what I tell everyone:

• Tip #1: Apply broadly to as many departments as you are willing to attend if you get in. If you're super-picky about where you want to do a Ph.D., then you may have a hard time getting in anywhere. Typically I recommend students apply to 10-20 departments. It's the same amount of work for your recommendation letter writers since they'll just re-send the same letter to all the departments, so you aren't creating more work for them.
  • Sub-tip #1a: Don't just apply to the few 'top-ranked' departments in your field. Broaden your search to a larger variety of Ph.D.-granting universities in your field. Look at who has been publishing in your field in recent years and find out where they work, then research those departments. Some departments aren't traditionally 'top-ranked,' but they have some of the most renowned researchers in a particular sub-field.
  • Sub-tip #1b: If you're unsure about whether to apply somewhere but feel like there might be a good fit, then go for it. You'll never know if some faculty there finds your app interesting and you end up working well together. (Of course, you can't apply EVERYWHERE to every school, but I'd say in general for applicants in my field it's normal to apply to 10-20 Ph.D. programs. Anything over 20 feels like too much to me.)
  • Sub-tip #1c: If you don't apply in time, you won't be considered as part of the formal admissions process. So if you're on the fence, it's better to apply in time rather than wait until after the deadline, in which case it's bureaucratically very hard to get you considered as an applicant. Don't be late!
• Tip #2: Make sure you have a super clearly-defined primary research area that shows your greatest strengths, since you will be competing directly against other applicants IN YOUR RESEARCH AREA. If you appear uncertain about your research area or are waffling between a few, then your application will look weaker than other applicants who have a clearly-defined area. This may be a bummer for those of you interested in interdisciplinary work, but even if so you still need a primary area where faculty can look at your app. For instance, if you're somewhat-interested in A.I. but have no directly A.I. research experience, then if you put down A.I. in your application, you'll be competing against applicants where A.I. is their primary area, and you won't do well!
• Applying to multiple departments within the same school is totally fine, since each department does Ph.D. admissions independently without consulting one another; it's not like they will "look down on you" for applying to multiple departments.
• It's very important that your letter writers turn in their letters on time since if a department starts reviewing applications and your application is missing a letter or two, then you may get skipped over since your app looks incomplete in their system. That really sucks, and it's not your fault at all; but some professors are super-late in submitting their letters since they're too overloaded with too many letter requests and other work duties. (It gets even more confusing since sometimes 'late' letters are OK if the department doesn't start reviewing apps until a few weeks later, like after winter break or something, but you can't know ahead of time when each department starts reviewing apps, ergh.) Don't be shy about reminding your letter writers repeatedly (but politely!). Seriously, don't be shy about it. It's your future career on the line; take necessary action to secure it. Good luck! (See logistics of asking for recommendation letters.)

Appendix whatever: added in Dec 2023

When you apply to Ph.D. programs, make sure you have a personal website that can be found on search engines. Ideally it would show up first for a web search for your name; but if not, then at least for "<your name> <school>" like if someone searched for "Philip Guo UCSD" they should find me.

You can keep your website super-simple and basic; don't stress about it! Your website only needs to be a single webpage with your name and current school on it (so that search engines can find it). Then add a link to your CV PDF, email address so people can easily contact you, PDFs of your publications if you have any (no worries if you don't), and a summary of your relevant research (and maybe industry) experiences. If professors are interested in your application, they will immediately search for your name on the web and see what they find. If they don't find anything, then that could leave a slightly-negative impression (especially in technology-oriented fields like computing ... maybe it's not as big of a deal in some other fields).

Note that just having a LinkedIn page is not enough since professors don't really use LinkedIn, unlike industry hiring managers. For industry jobs, LinkedIn is important; but for Ph.D. applications ideallly you'd have your own personal website.