Advice for preparing to apply and actually applying to science Ph.D. programs
by Philip Guo (philip@pgbovine.net)
The Ph.D. application is best viewed as a job application rather than as a traditional college or professional school application. Candidates are applying for the job of a paid research assistant who just happens to also be able to take classes and work towards a degree. Therefore, to maximize your chances of admission, you must show how you will be a valuable asset to the school in terms of producing research results.
Obligatory introductory section
I applied to Computer Science Ph.D. programs in Fall 2005 with almost no idea of how to approach the process or what I had to do to prepare. Fortunately, with a combination of good luck and helpful advice from my friends and colleagues, I was able to make it into the Ph.D. program at Stanford University, which I began attending in Fall 2006.
This advice will probably benefit you the most if you are a sophomore or junior who intends to apply for Ph.D. programs at top research universities during your senior year, because there is still adequate time to improve your resume before application time by gaining more research experience. If you are actually in the process of applying right now, the only thing you can really control is refining your writing to cast your research experiences in the most favorable light; all of the research work leading up to your application has already been done (or is in the process of being done).
Disclaimer: At the time of writing, I have never served as a member of a Ph.D. admissions committee, so this advice is based purely on my own experiences and on conversations I had with professors and graduate students. Please seek advice from multiple reputable sources before and during your application process.
I'm gonna try to do this in the format of an FAQ. Some of these questions are real ones that I've asked or been asked; others are made up just to stick to the format. Let's see how this goes ...
What really is a Ph.D. program application?
In my opinion, an application to a Ph.D. program is really just a job application, and likewise, the admissions process is like the hiring process for a company. (The analogy is not perfect, but it's good enough to carry us through the rest of this FAQ.)
So what kind of job are you applying for?
You are applying for the job of a research assistant in a university research laboratory. As a research assistant, your school tuition will be paid for by your advisor's grants and you will be given a monthly stipend to spend on housing, food, gummi bears, etc. Your annual salary will total around $50,000 to $60,000, but of course, most of that goes into paying your tuition. In exchange, you are expected to work on projects that will lead to innovative results that can be published in peer-reviewed journals or conference proceedings. The exact day-to-day work differs depending on your field, but expect to work long hours and weekends :)
How does a Ph.D. program fundamentally differ from all other graduate school programs, such as M.D. (medical), J.D. (law), or M.B.A. (business) programs?
The fundamental difference is that you are being paid to pursue a Ph.D. whereas, for all other graduate school programs, you are paying to pursue the respective degree. You must understand this key economic difference in order to understand why Ph.D. program applications differ from med school, law school, or business school applications.
How is this fundamental difference relevant to the application process?
Because you are being paid to pursue a Ph.D., it is really a job (with the perk of getting a doctorate degree after N years), so the application process is like a job application. Simply put, the admissions committee wants to see how well you can perform as a paid research assistant. In contrast, for medical, law, business school, etc., you are paying a hefty tuition for the opportunity to learn specific skills in an area. Thus, the application processes for these schools are designed to assess whether you have the potential to be a good learner and to be a successful practitioner of the learned skills after you graduate. Thus, the admissions committees for these programs may seriously consider secondary factors such as character traits, community service, leadership potential, and charisma during an in-person interview.
What is the most important part of my application?
The most important part of your application is your prior experience in performing research and your potential for being a creative, hardworking, and productive graduate research assistant. This is pretty much the only factor in determining whether you are admitted to a Ph.D. program at a top research university (I've heard that grades and GRE scores matter a bit more at lower-ranked universities). All other secondary factors, such as your leadership roles, community service, clubs, hobbies, and other experiences that portray you as a well-rounded and balanced individual count for almost nothing. Why? Because the job description simply calls for a research assistant, not an all-around gregarious fun-loving charismatic stud.
Of course, it doesn't hurt if you really are an all-around gregarious fun-loving charismatic stud, but you still need to have strong credentials as a researcher in order to be admitted. You can't just charm your way into a reputable science Ph.D. program. You can't bullshit or game the system. The Ph.D. admissions committee consists of professors in your field who will scrutinize your application for its technical merits, not at all like the undergraduate admissions officers who are constantly on the look-out for well-rounded, diverse, charming, and emotionally healthy high school students to admit to their college.
What do you mean by research experiences being the only thing that matters?
I mean just that --- plain and simple. You are being hired to do research for a university, so the best predictor of whether you have the potential to do more good research is whether you have contributed productively to research in the past.
But what about all my wonderful extracurricular activities? My med school and business school friends keep on telling me how important they are.
There will be spots in your application to write those in, but don't bother to elaborate on them in your statement of purpose (application essay), unless they are somehow relevant to motivating you to do research. It's great to be well-rounded, and given two individuals with near-identical research credentials, I'm sure that the admissions committee would rather admit the student with better extracurriculars, but your research is really the primary factor in determining your admissions. If you've got good extracurriculars, especially those demonstrating leadership, write them down on your application, but recognize that they are really only used as a tie-breaker in close calls; your research alone will mostly determine whether you are admitted or not.
Admissions to other graduate programs such as M.D., J.D., or M.B.A. (where you have to pay to attend) is usually a more holistic process that takes secondary factors (extracurriculars, personality, etc.) seriously, because these schools want to produce graduates who are not only knowledgeable but also personable and charismatic. A business executive or doctor who is more well-rounded and emotionally healthy can interact with others better than one who is a recluse. However, Ph.D. students aren't hired for their well-rounded personalities; they are hired because they have the ability and willingness to work long hours to churn out publishable research for their advisors while they are in graduate school. Once they graduate, students who are more personable may have an easier time finding better jobs, but the research and publications for the graduate institution have already been done by that time.
How much research is 'enough'?
I think that working in 3 different research groups is ideal, although this is purely my own opinion. I only say 3 because you will need 3 letters of recommendation for your applications, and if all 3 come from professors or research scientists whom you've done research for, then your application can be quite competitive. You should get started early, preferably on your first project at the beginning of your sophomore year. It doesn't matter if your first project isn't exactly what you want to pursue for your future research career; the whole point of getting started early is to explore what you like and don't like. Doing a research-based summer internship in a government or industry laboratory is a great way to earn an extra research experience without devoting your entire school year to it. If you work at the same lab for 2 summers, then that might be enough to earn you a strong letter from your supervisor.
Of course, I know many people who've been admitted to great Ph.D. programs who have only worked in 1 or 2 research groups during their undergraduate years, so 3 is not a hard-and-fast rule. However, if you've only worked in 1 or 2 groups, you'll be expected to have done more significant work for them; in effect, to demonstrate depth rather than breadth.
What research is 'good enough'?
'Good enough' is really a subjective notion, but what the admissions committee is looking for is impact. In other words, did your work have tangible results? Or were you simply cheap labor for performing data entry and maintaining the group website?
The ideal indicator of 'good enough' is having publications in a peer-reviewed journal or conference (being first author is even more marvelous, but extremely rare for an undergraduate). Being published shows that you have made a non-trivial contribution to a research project that advances your field. However, most undergraduates (myself included) will not be able to publish by the time they apply to Ph.D. programs, so don't worry if you can't either.
What if I can't get published by the time that I apply?
This was my biggest fear when I applied to Ph.D. programs. Don't worry; most undergraduates will not be able to publish by the time they apply, so relax! If you have published in any smaller venues like the undergraduate research journal of your college, made a poster presentation, or presented your work in any other way, make a note of that. If you have submitted papers to venues and are awaiting reviews, list those papers in your application and write Submitted for publication next to them to demonstrate that you and your advisor are actively working towards publishing research that you've been involved in.
Besides having publications, how do I demonstrate that my research has impact?
You need to show tangible results of your work, whether it be a software system you've built, an experiment you've run to completion, a statistical analysis you've performed on real-world data, etc., and more importantly, demonstrate the significance of those results in advancing your particular field (even if it's in some tiny way). You will need to write persuasively about the impact of your research in your statement of purpose, and your advisor will have to corroborate your claims in his/her letter of recommendation.
But I'm just an undergraduate! I can't be expected to do research on par with all the graduate students and postdocs!
Yes, the admissions committee knows this, and does not expect you to perform at the level of a graduate student! You are not expected to be overly original or innovative, because the truth about undergraduates working in research is that most start out performing menial tasks that the graduate students don't want to do, basically the lowest of grunt work. However, the hope is that you will eventually move up to doing less grunt-like work and have a more direct impact on creating results for your research group.
The admissions committee is looking for the potential to become a productive Ph.D. student, so demonstrating your work ethic, determination, and willingness to take initiative in learning (in addition to the impact of your research, however small) are great ways to improve your chances of admissions.
Don't be an idealist and expect to completely push through your own wild ideas to fruition while you are still an undergraduate; you are at the absolute bottom of the research food chain, so you should accept your role and perform it with diligence so that you maximize your chances of moving up to the next higher level as a Ph.D. student.
But what about my grades and GPA?
This isn't med school or law school; Ph.D. admissions committees at top-ranked graduate schools don't care as much about your GPA, unless it's egregiously low. Basically, if you have a GPA that shows that you're a pretty good student, and you have great research experiences to write about on your application, then you're all good. It's hard to define exact numbers, but I would say that a 3.6 GPA is probably fine for applying to top-ranked schools (I originally just threw that number out there as an educated guess, but I actually got an email response confirming that 3.6 GPA is good enough as long as you demonstrate strong research experience and potential ... of course, it's only one sample point, but that's better than zero). If you feel self-conscious about your low GPA, try to explain somewhere on your application why your GPA was low ... perhaps you spent so much time and focus on your research that you didn't have enough time or effort to study for the final exams in certain classes (or something like that).
Don't be intimidated by the fact that most top schools report that the average GPA of their incoming class is 3.8 or 3.9. That doesn't mean that you can't get in with a 3.5 or 3.6. It simply means that, most of the time, students with the potential to do great research at the Ph.D. level also take their studies seriously enough to earn a high undergraduate GPA. However, they didn't get accepted primarily because of their high GPA. Having a high GPA alone can never get you accepted into a top Ph.D. program.
Who should I get to write my letters of recommendation?
You will most likely need 3 letters of recommendation for your Ph.D. applications. The best-case scenario is that all three letters come from professors for whom you have done research (a.k.a. your research advisors). If you have worked in research positions in industry or for the government, then your supervisor could also write a letter for you. If you can't get 3 research-based letters, then you will need to get letters from other sources. (However, if none of your letters come from research advisors, then your chances of getting in are quite slim.) These alternate sources might include supervisors from non-research industry positions, researchers you worked with who were not your primary advisor, or professors of classes you excelled in.
Getting a letter from a professor whose class you excelled in isn't the strongest endorsement since it doesn't directly demonstrate your research experience or potential, but it is necessary if you can't get 3 research-based letters. If you must get a letter from a professor for getting a good grade in his/her class, it's better to have that letter come from a professor who's taught you in an upper-division (possibly graduate) class rather than an introductory class. That shows you've taken the initiative to take on more challenging classes, and a letter commending your achievements in a difficult upper-division course (such as a 15-person seminar) can better demonstrate your potential for graduate-level research work than a letter talking about how you got an A+ in a 400-person freshman class.
Also, if you can get a letter from a professor for a class for which you've gone a really cool final project that demonstrated creativity, ingenuity, and/or clever insight, that's much better than a letter from a professor who taught you in a standard lecture/exam class. In fact, it's a good idea to seek out opportunities to do open-ended final projects for classes, because that often demonstrates the same kind of skills as required for being a Ph.D. student.
If I am a sophomore or junior, what can I do to prepare for my applications during my senior year?
Favor research over taking more upper-division classes, if you don't think you can manage doing both effectively. Try out different research groups if you want to, but by your junior year, find one that you can dedicate yourself to for the duration of your undergraduate years (1.5 to 2 years), and really buckle down and focus on that work. Graduate schools want to see that you can dedicate yourself to one project over several semesters instead of jumping around all the time.
Meanwhile, in the summertime, try to either get research-based internships or work in a research lab in your university rather than getting non-research internships in industry (this bit of advice is extreme, because it is often educational to get industry work experience; you might find that, like most people, you actually like working in industry better than working on research).
Take advanced graduate-level classes if you feel passionate about the subject matter, and if you can still make reasonable progress on your research at the same time. Don't fret so much about your GPA; if you don't get such good grades in those classes, you can either drop them or mention that you tried taking on more challenging classes. Taking graduate-level classes exposes you to cutting-edge research that you might actually be working on in the near future (instead of the classic well-established research that you find in undergraduate course textbooks), gives you a smaller, more intimate environment for interacting with professors and graduate students, and gives you a preview of what graduate school is like (aside from doing research).
What about if I've been working out in industry for the past couple of years but want to go back to school for a Ph.D.? What are my chances of being admitted?
Perhaps 1/4 of my incoming Ph.D. class had worked in industry before starting the Ph.D. program (in other words, they didn't come right out of college), so it is definitely possible to get in. The statistics might vary by department or school. My intuition is that you would stand a better chance of being admitted if you did more research-like work in industry rather than production-like work, but I have very little knowledge about this issue. You might have to tailor your application a bit differently, but the end goal is the same: to demonstrate that you have the potential to contribute to the research community during your years as a Ph.D. student (and hopefully beyond).
I've been advised that if you've only been in industry for several years (maybe 1 to 3 years), then you can still get recommendation letters from professors you've worked for or taken classes from in your undergraduate institution, and you will be considered similarly to students applying straight from college. However, if you've worked for a significant amount of time (e.g., 8 years), then you should show how your work experience prepared and motivated you to come back to pursue a Ph.D., which might be difficult if you've been outside of the research community for too long.
How much should I stress about my Ph.D. program applications?
None, or almost none. By the time you apply to Ph.D. programs, all of the hard work --- the research that you have worked on during your undergraduate years --- has already been done (or is continuing to be done). All that remains is to get 3 letters of recommendation from your professors and write a statement of purpose that summarizes your research experiences and mentions examples of research that you would like to pursue at the respective institution.
The Ph.D. program application is notoriously simple. Most of it is just routine fill-in-the-blanks which don't matter much at all (e.g., personal information, list of classes, awards and honors, GRE scores). Spend lots of time refining your statement of purpose because the admissions committee is going to read that piece of writing the most carefully.
Be honest and straightforward in your application, but write with assertiveness and confidence; there is not much you can do to game the system and improve your chances if you are not qualified with adequate research experience. Professors on the admissions committees didn't get where they are now by being naive and easily fooled.
Why do graduate schools ask me what other schools I am applying to?
You might notice that some graduate schools have a section on their application where they ask you to list what other institutions you are applying to at the same time. Some people are wary that the school might be biased if you put down certain choices (something like, "oh, this applicant is applying to 4 other top schools, so it's okay if we don't accept her ... one of the other schools will"), but I don't think that this is the case. They simply want to collect statistics about where else their applicants are considering. My advice is to simply answer that question truthfully, but if you're really uncomfortable about revealing this information, simply leave it blank. It should not penalize you.
Should I make contact with professors whose research I'm really interested in working on? Will that improve my chances of admissions?
The short answer: no. The longer answer: no most of the time, but yes only if you have some sort of connection to the professor (usually through a professor at your own school who is favorable to you). I'll explain:
Do not contact professors whom you have no visible connections with or else you will most likely not get a response since you are essentially sending them spam (unsolicited email). If you take the time to look on the websites of professors, many of them have a note on there saying something like Please do not contact me about information regarding admissions --- only the admissions committee can decide who is admitted; professors cannot admit individual students (with varying degrees of politeness). To a professor who receives tons of emails every day, there is nothing that sounds more insincere, contrived, and annoying than receiving impersonal emails from students who express an interest in his/her research, paraphrasing the contents of abstracts from his/her papers, and citing elaborate stories about their enthusiasm. There is no possible way that a professor you don't know will respond favorably to you if you send them email before you are admitted.
One way to legitimately contact a professor is if you either know them personally (which is unlikely for an undergraduate) or, more likely, your current research advisor (you ARE doing research right now, aren't you?) knows that professor. It doesn't hurt to ask whether your advisor knows professors at schools that you are applying to, and to ask whether it would be appropriate for you to contact them regarding possible research opportunities. Please be tactful, however, because you don't want to annoy your current advisor since he/she will be writing your recommendation letter :)
Another way to legitimately contact a professor is if you have met him/her at a conference or other professional venue and had some meaningful discussions about your own research (or someone else's research that you assisted with); that way, there is some solid basis to your email so that it is not totally spam. Again, please be tactful; you'll know when you are in a position to take this approach (very few undergraduates are ... I certainly wasn't).
If you have no connections, don't sweat it. I didn't contact any professors before I was admitted, and I was still okay. If you still genuinely want to find out information about the professor's research (beyond what you can find on the Internet, which is often more than adequate) or whether he/she is looking for new students in the upcoming year, the best thing to do is to email the professor's current Ph.D. students. Students are usually less annoyed at those kinds of unsolicited emails because they usually have no say in determining admissions. However, do this sparingly or else you will be the butt of the jokes around the grad student lounge water cooler (Hey, did you also get that email from that loser undergrad who wants to work with our advisor? What a chump!).
Addendum (2007-01-10): Several people have informed me that in their field of study, they are actually encouraged to contact professors before they are admitted and have had positive experiences doing so. Perhaps it depends on the field or even on the department (professors at smaller departments who receive fewer applicants might want to reach out to them more than professors from larger departments who get thousands of applicants annually). Ask your friends and academic advisors about how best to proceed.
Addendum (2007-02-22): Additionally, I've also heard that it really depends on the status of the particular department you're applying to. Perhaps professors at lesser-known schools are more likely to be receptive to receiving personalized emails than those at more well-known schools, but I don't know this from firsthand experience (I've received some email feedback mentioning this opinion, though).
Addendum (2007-06-27): I have since softened my tone on this issue, due to comments from several other graduate students. My previous hard-ass stance used to be NEVER CONTACT PROFESSORS BEFORE YOU ARE ADMITTED, but now my more liberal stance is, "only contact professors if you have some sort of connection to them, most likely via a professor from your own school."
Is the statement of purpose kinda like my awesomely creative college admissions essay?
No, not at all! Don't try to be cute and creative. The people who are going to be reading your statement of purpose (otherwise known as your application essay) are professors in your field who expect to find out about your research experiences, motivations, and reasons why you want to work towards a Ph.D.; they do not want to wade through some melodramatic narrative of how you were a child genius who suffered through lunchroom attacks by bullies but persevered to study even harder and pursue your passions for science. Your writing should be passionate and persuasive, but it should mainly focus on your research, not on your personal life. Professors don't want to read sappy bullshit essays.
How should I write my statement of purpose?
Show, don't tell. Reinforce every claim you make with supporting evidence from your own experiences. Professors can easily sniff out fluff and bullshit.
Motivation is great. Don't just list out your research projects as successive bullet points. What motivated you to undertake those projects? What did you learn from those projects that motivated you to further pursue research? Why do you want to pursue a Ph.D.?
Transitions are important. Don't just have sentences mash up against one another without any connective tissue. This is mostly a stylistic tidbit, so don't worry about it when you're doing your initial drafts, but an essay that flows well shows that you are better-organized and simply makes a more appealing impression.
Try to put a 'researchy' spin on your work, even if the research work that you did was mostly grunt work, which is understandable because most undergrads have to do a lot of grunt work (programming, entering data, making graphs, running simulations, wet lab work, etc.) as part of a research team. After all, you're at the absolute bottom of the academic food chain. What I mean by that is to not just say that you did the grunt work for a project, but try to step back and mention the 'bigger picture' of what impact your grunt work had on the rest of the project. However, don't be too facetious or you'll end up sounding like a pathological exaggerator (don't say that you co-developed a novel algorithm when all you did was code it up and test it on some simple inputs, under the guidance of the grad student who actually developed the algorithm).
Should I name names (of professors whose research interest me) in my statement of purpose? How can I go about doing this intelligently?
Yes, you should, if you have an interest in actually working with them, especially if they work in an area that relates to your prior research in some way (show, don't tell). At the end of your statement of purpose, briefly mention the professors and projects that interest you, but don't suck up too much. Aim for being concise and direct.
Before mentioning a professor whom you'd like to work with, though, try to make sure that they are willing to accept new students in the upcoming year. Don't email them directly to ask that question, because they will just get pissed and delete your email. Instead, if you're a bit brave, email their current Ph.D. students to ask whether their advisor is planning to accept new students.
Also, if you are going to name names, name several of them and not just one (unless, of course, you would only be willing to work with one professor at that school). If you only name one professor, you risk being pigeonholed and getting your application discarded if that one professor is simply not taking new students in the upcoming year.
Should I customize my statement for each school?
Yes, if you want, but not much. The only part that should be customized is the paragraph (probably the last one) where you mention which professors or projects at that school interest you. There is no need to suck up to the school by saying how it's such an esteemed institution or how it's world-famous for this kind of research, blah blah blah. If the school is famous for something, I'm sure that they already know that. Just cut out the bullshit to leave more room for real content. (It's perfectly fine to not customize, either, if you don't feel like doing so.)
