How to apply to graduate school

Posted on July 22, 2016

This is a repost of my blog's March 14th entry.

Let’s begin with a bunch of warnings. The largest warning is that this is all what I think, and I don’t work on an admissions committee and generally have no expertise in whether any of this is accurate or not. These are generalizations that I’ve observed based on my limited experiences and the those few stories from friends who have shared their experiences. (... you'll notice that people don’t talk about their graduate application processes just like they don’t talk about their SAT scores or grades. Same concept.)

Second warning: I’m not the best candidate for graduate school. I am a good candidate, and I did very well in the application process given my background and what I was applying for. But everyone’s circumstances are different and everyone’s application package is different, and it’s the complete application package that matters. For example, I don’t have a first/ second/third-authored paper, which can be considered unusual. I was applying to computational labs without substantive computational experience, and I only got a B in Intro to Machine Learning which is the field I purportedly said I wanted to work in. I still did well in the process; everyone’s package is different. I think the key thing is knowing what you want to do and then applying for that. (Plus there’s a ton of luck involved.)

Okay, warnings done! Moving into step I: what you need to be thinking about as an undergrad with regards to grad school.

Step I. What you need to be thinking about as an undergrad

There are five things that science PhD programs want from you. (I’m just going to write in the declarative. Please add the relative language and necessary disclaimers.) They are:

Research Experience

These are not in order. Research experience is the most important, because you need to know what you’re getting into. I don’t even know what I’m getting into yet, since I haven’t done a PhD, but I have much better idea than most of the people out there thinking about it. I know what it’s like because I’ve worked on research during the school year, and then I worked on it full-time during the summer. I’ve worked on things that are similar to what PhD students and lab technicians work on. This gives me a sense of what goes on in a day-to-day basis sense. I still have no idea what I’m dealing with in terms of things not working for months and months on end, but I’m confident that I’ll be able to survive it because I’ve persevered other obstacles in life before. I fully expect doing a PhD is going to be kind of terrible in many ways, and I think that the sooner people are aware of that the better they can start thinking seriously about whether they want to do it. You’re paid a stipend similar to minimum wage. You’re working constantly. You’re working on one thing so it’s very difficult when it isn’t working. You’re often working by yourself. You’re working on hard things. You’re working directly under a boss who you’re going to disagree with in many cases. You’re working with a limited group of people in your lab and you might not get along with some of them, and you have to get along with them for years. It’s going to be kind of terrible. (I’d advise people to read PhD Comics for common complaints in grad school. I thought it was really depressing the first time I came upon it, but once you’ve resolved yourself to it it’s just constant inside-jokes :)).

On the other hand, once you’ve realized all of these unhappy things will happen, you can start figuring out if the upsides are worth it. You’re working on something you really like. Your job is pretty much what you want to think about for fun. You’re learning, constantly. You can do whatever you want. You have almost complete freedom with your time. You have free coursework. You get to (have to) teach. You get to interact with leaders in your field and in the world. You get to go to conferences. You get to hang out with really smart people all day. You get to change the world in the tiniest of ways. You are constantly expanding your horizons. You’re going to end up with a nice shiny degree at the end. Did I mention you get to learn all the time about what you want to learn about?

Those are some of my reasons for wanting to go to grad school: I don’t think it matters what they are, as long as you’ve thought seriously about both the pros and cons. It takes a long time—I’ve wanted to be a professor since high school, but I did some serious waffling every time I learned a new element of what grad school actually entails. (I didn’t even know professors did research when I started out.) Then I had to do a lot of research and figure out if I liked it or not, and evaluate all of the other paths I could go down, and do lots of internal pro and con lists of what I wanted out of a job, and in the end I ended up completely sure I wanted to go to graduate school. So that’s, I think, the main reason why we need experience in research. To figure out if it’s the kind of work we like to do and would want to do for 5-7 years in the US (3-4 years in the UK).

Recommendation Letters

Next: recommendations! These are very important—one professor on the Admissions Committee at MIT told me that rec letters, grades and standardized tests are all weighted almost equally. You’re going to want a great rec letter from your research advisor. If you have another research advisor, you’re going to want one from them as well. And then you can get one from a teacher who has seen you in a lot of classes. You’re going to need 3, usually. For my PhD applications I needed exactly 3 for all of them.

What to do to get good rec letters: if you’re in class, do extra things. Do really well in class, speak up in class, work well with others, but also go visit them in office hours, be curious about the material, go above and beyond in assigned projects and see if you can get an independent project going. A lot of these should come naturally: if you’re interested in the material and willing to work hard on it these things come through. But don’t be afraid to visit in office hours and propose an independent project. Go visit in office hours so the professor knows who you are. Seriously. (The situation is complicated in larger universities where this isn’t as much of an option as it is at small liberal arts schools. I don’t actually know how rec letters work in these environments, unfortunately.)

How to get good rec letters in research: I’m not sure, actually. What seems to work is to show all of the qualities that you’re going to need to grad school, especially a sense of independence, drive, persistence, and initiative. Ability to work independently and generate new ideas is probably the most important thing you can demonstrate. And it’s not like you’re going to be demonstrating that off the bat (… I had so many problems with working independently in the beginning. So many problems.) but you’re all 18-year olds (… apparently I am aiming this at a new undergrad audience! Doesn’t surprise me, actually) so they don’t expect you to know what you’re doing. I will emphasize that again: professors do not expect 18-year-olds to know what they’re doing, and that’s why they want you to stay in the lab awhile, so that you can figure it out. Of course, if you do know what you’re doing, either because you’ve done some research in high school or you’re a quick study, then you’re going to be a star and start getting assigned PhD-level work when you’re an undergrad. Which puts you in a prime position for grad school, and those are the people who get in without taking a few years off to do extra research. But if you’re not amazing, you’ll get there! You’re figuring out so much stuff about yourself as an 18-year-old that no one expects you to have found exactly the field you want to be working in on your first try. Supervisors will try to get you to stay in the lab for a while (… because untrained people are useless, but they get better with time!) but it’s a mutual process of you learning what you want and figuring out what actually gets done in the lab once you understand how to make progress.

(The fact that research is an apprenticeship system is still weird to me to think about.)

But don’t try to force the rec letters—or any of this, really—since it’s all about finding out what you want to do in life. If you’re suited for research—if you like research—you’ll figure that out and it’ll come through. If you don’t like research, then doing research is an excellent way to figure that out. Get involved as soon as possible! I urge people to definitely have gotten started (gotten accepted to a lab and begun working) sophomore fall.


It is so singularly humorous to me that 4 years of work are condensed into a single page of A / B / C / D / F letters. It is insane, how much work goes into generating that page.

What people are looking for: that you took classes in the field that you say you want to be working in. They’re looking for good background knowledge, and that you did well. If you don’t have the right background knowledge (like me!) you can always learn it later, but you need to have shown that you have some ability to acquire it. A few professors told me that because I’d done well in quantitative classes, they know that I can learn other quantitative material. It’s also very important to pick classes in a specific field if you want to go into that field but don’t have research experience. In the face of lack of research experience, you’re going to have to point to classes to show your interest and aptitude.

Also, no one cares about anything but your science classes. As you might have suspected, but nice (or annoying) to have confirmed. Anything you do outside of research is for fun or to keep yourself sane and you will not write about it any of your essays. This is different than if you go to med school and you have to write about yourself as a person instead of as a potential researcher!


I very happily did not have to take a GRE subject test, which are very difficult. The GRE itself isn’t bad though. The math is actually easier than SAT math—you just have to do it fast and well. There are three sections of the GRE: math, reading, and writing. If you’re doing quantitative science, they care about your math score, then your reading score, and some care about your writing score. At the top institutions, you want to be 90th percentile and above—85th percentile seems to be fine—especially in math. You can find average GRE scores for different schools in different subjects if you search really, really hard, but I think it’s a useful exercise. Ask your professors how your scores are, and they’ll be able to tell you right away whether you need to retake them or not.

People take the average across the math and reading scores and generally use that, so your reading scores have to be decent. Generally, I’ve been told that people do best on math, well on reading, and fine on writing in applying to the MIT BCS program. I have no reverted from my high school English self, so I scored tremendously well in reading, well in writing, and fine in math as per my usual efforts.

It all comes down to practice though. This is a test that if you practice you’ll do well. I think I probably did about 10 practice tests all the way through. There are 5 sections, 4 of which are graded (but you don’t know which one) alternating math / reading, and then a writing section at the end. I did an hour a day of this for two months and only took the GRE once. The general MCAT (for med school) is much tougher and requires a lot more studying.

Statement of Purpose

This document is weighted equally with your grades, which you spent 4 years working for, so you’d better hope it matters! The reason it matters so much is that it your only place to connect all the dots of where you’ve been and where you see yourself going. You need to hit several points:

There are many resources online in how to write a good statement of purpose, and I’d spend some time poking around and figuring out what people want. My statements of purpose follow the general structure outlined above.

My statements of purpose weren’t too different from school to school, and I don’t think they should be if you are applying to one type of graduate school. (I do know people who are applying in two separate fields, and theirs are obviously going to look different. But I applied in neuroscience / psychology for all of my schools, in cognitive neuroscience / computational cognitive neuroscience / computational cognitive science, so all the topics were similar.) Your research questions you’re interested in are pretty much the same, your research experience and coursework is the same, and your final concluding statement tying it all together is pretty much the same. The one paragraph that changes a lot is your “these three labs” paragraph, and then you have to edit the first and concluding paragraph to weave in the ideas from those three labs. So it’s a bit of a hassle to translate the essays between schools, but the hard part is picking out the labs and figuring out what you want to do in the first place.

Which leads me to step II: the Application Process.

Step II: The Application Process

All right. Let’s establish the timeline first. You’re doing research starting as a sophomore. You’re doing research for as many summers as you can swing it, full-time. (During the school year you’re doing 10-20 hours per week, usually. I know people who do up to 40, but I save that for the summer.) If you have enough research experience, you apply to PhD programs your senior fall. If you don’t have enough research experience, you apply for two-year lab technician jobs and apply during your second year working. You’ve already taken your GREs the year before you applied. Applications are due in December, and it’s August. All right, timeline established: you’re now at the point where you’re applying to graduate school.

I really recommend people start applying in August, because this whole process TAKES FOREVER. I cannot stress enough how much time it takes. It’s also why I am really glad I waited a year before applying, because I simply did not have this much time my undergraduate senior fall. So I’d recommend you start tinkering around in August, especially if you’re not already in a research lab and still have to figure out what you want to study. I know that some people start their applications over thanksgiving break, so it’s possible, but deadlines are in early December (or in some cases November) and it’s really best to start as soon as possible.

There are several goals here: figure out what you want to research, figure out what schools you’re applying to, etc. The best way to do the above is to start the search of looking for labs you’d be happy to work in, and things fall into place from there.

My rule was: if I find three labs I want to work in at a single school, I’m applying to that school. Generally I was recommended to apply to 5-8 schools, and I applied to 8. That meant that I kept on searching for labs until I had 8 schools to apply to. It’s surprisingly hard to do this. You usually have a few labs in mind because those are the ones close to you. Then you click on their collaborators and search their topics until you find other labs. Soon enough you find networks of people and you’re reading their webpages and papers and find more people to look for, and you get a lot of tabs with different professors. It’s a fantastic way to learn about the field. It’s a fantastic way to learn about yourself as well because you figure out what research you’re interested in. This was probably the most learning-intensive and time-intensive part of the process, because you have to spend so much time figuring out what you want. Research wise, and also people wise, because you have access to who is in their lab, and also people around you are going to know the reputation of these people. ASK THE PEOPLE AROUND YOU what they think about certain professors, because if they’re in the field they know. Also, I looked at things like whether there were any women in the lab or whether there was online information about whether the professors were good teachers. I asked very carefully if professors were good mentors. A lot of this stuff is more easily done when talking to the graduate students in the lab, but the more screening you can do beforehand the better. And it’s not like all of the screening is general: some people like professors who are more hands-off, or hands-on— people have different learning styles.

So this is the most important part: picking the labs. Be very organized about it. And you have to read / skim a lot of papers, because this leads to the next step: emailing all of these labs.

You have to send an email to every lab you’re interested in working with. You basically write a miniature statement of purpose in this email—but it has to be miniature. You ask them if they are taking graduate students, you tell them why you’re interested in their work, you give them your background, then you thank them very much for their time and sign off. It’s the most important document, I think, and it’s not even in your official application package. It’s how your potential supervisors figure out if they want you or not, and if they want you they’re going to push your application through. These letters matter a lot. Ask a lot of people to read them over. (Here's a good post to check out on this!)

(It was also extremely encouraging in my case. I almost didn’t apply to a lot of labs because I was intimidated, but received positive responses at this stage which encouraged me to continue).

Then it’s probably about mid-October and you’re onto the almost-easy part: writing the statement of purpose. I say easy because by this point you’ve already got the outlines of your statement of purpose in all of these emails. You have to say why you’re interested in each professor’s work in these emails, so I basically copy-pasted parts of these emails into my statement of purposes for each school. You already know the general scope of what questions you want to answer in graduate school, because you had to pick all of these labs out and it’s easy to figure out the commonalities between them. You know why you want to go to these specific schools because you hand-picked each of these labs. At this point I felt pretty empowered, because I had so much better an idea of what I wanted than when I started the whole process.

There are quite a few miscellaneous essays you have to do as well—sometimes you’ll write a real personal statement (where you write about challenges you’ve overcome) or funding application essays, and lots of short additional research essays. And you have to do a heck of a lot of formatting, because every school wants the same information in different formats. Your CV is going to look beautiful by the end, too, because you’ve moved it around so much trying to make it perfect for the emailing stage. The key here is to be really, really organized. I had a master spreadsheet with all of the deadlines, individual spreadsheets for each school (with tabs for each professor), and a master “letter” document containing all of emails to professors. I won’t tell you how to organize it, because it drives me crazy when people do that, but just be organized. I have a friend who went through the entire process and was accepted to Brown, and told everyone and was confirmed with the lab and everything, but he forgot to confirm with the school online and so lost his place and has to reapply again this year. Insane things happen—this story scarred me, especially since he’s someone I know personally and we were all really excited for him. Watch your administration details!

Cool, so you’ve written your personal statements. Get someone to read them over. Find that one person who’s going to be willing to read it over and over and over until it’s great. For my undergrad thesis, that was Prof. Hildreth (so much love to her). For my statements of purpose, that was Ika’s mom, Prof. Kovacikova (so much love to her as well). You never know who might be willing, so send it to all of those connections you built up. I think almost every professor who I had a relationship with ended up reading some version of section of my statements. It was an incredible feeling, having these people willing to read and comment on what was essentially a formulation of my life-goals with regards to research.

Statements of purpose done! Then do all of the administrative stuff. There’s a lot of data entry—it’s all based on the CV and short essays. Most applications are due by Dec. 1st, and it’ll be a rush to those deadlines, I guarantee. Triple-check your rec letters arrived and all of your information is submitted.

And now it's time to check out the next stage in the process: the interviews!