I’m asked this question a lot. I’m going to give some advice and demystify the process of US admissions a bit. I also address the question “Would you be my advisor if I apply?” Since I can’t respond in detail to most emails, I hope this post answers your questions.
I won’t demystify entirely, because uniform applications are unhelpful to us reviewers. But I do want to make the application process easier to understand, both to make it easier for people like me to decipher your application, and also to level the playing field. Undergraduates at the top research institutions have the advantage of advisors who already give them this advice.
First, some caveats. My experience comes solely from two years on the admissions committee in Yale political science, one in Columbia political science, and one in Columbia sustainable development (which is essentially an applied economics PhD in science, environment and health topics). Since most professors in the world are more senior and experienced than me, you would be wise to get second opinions.
If you are applying to economics, you absolutely must heed the following:
- Greg Mankiw’s advice for aspiring economists and why you need math
- Susan Athey’s advice to economics applicants
Nearly all of Athey and Mankiw’s advice applies equally well to aspiring political scientists who want to do political economy or development work, and indeed almost any of the applied empirical fields in politics.
For advice on political science PhD applications, also see Dan Drezner and Dan Nexon, who focus a little on international relations scholars. My thoughts are are on economics and political science together, with the most relevance for those doing applied empirical work and my fields: development, comparative politics, political economy, and labor.
Please point everyone to other sources of PhD advice in the comments.
Should you do a PhD?
A first important, simple point: If your goal is to be a professional researcher and instructor, then a PhD makes sense. If not, not.
In particular, if your goal is to be influential in policy and practice, then an MA or MPA or MIA from a US or UK/European institution probably makes far more sense for you (e.g. Tufts, SIPA, Princeton, SAIS, HKS, etc).
Why? Opportunity cost. A PhD is five to seven years, and a Master’s is two. A PhD means you are sacrificing several years of valuable work experience and as much as several hundred thousand dollars in income.
I talk about choosing among Masters-level programs here.
One thing to keep in mind: a PhD program, like most organizations, don’t just teach you, they socialize you. They gradually change what you think is interesting and important, the peer group you compare yourself to, the value you place on leisure and family over career, and the kind of life you will value when you emerge. This is good for science, maybe or maybe not so good for you. Enter with open eyes.
Where should I go?
If you are set on a PhD, then you’ll want to attend an institution with full funding (which often comes in exchange for a reasonable research and teaching load). If a PhD is going to land you with tens of thousands in debt, it’s a highly questionable decision.
Your first objective is to get the best quality general research training you can. So apply to as many of the top schools as possible and then, once admitted, start to narrow down your choices based on fit and overall quality. Visit everywhere you are admitted, to be confident it’s the right place for you. I’m very serious about this..
The other reason to apply to many places is that the admissions process is not only ridiculously competitive but also somewhat random. Getting from the 100 attractive candidates down to the 30 to 50 you admit is very idiosyncratic. So even strong candidates with a good fit might not get in.
That said, schools are much more likely to admit you if you demonstrate a good fit with their faculty–something you need to help them see by researching the faculty and reading their work, and describing how you would fit in. Then explain in the letter the people you see as the best fit (see below). This is more important in politics than in economics. In my experience, in politics programs they tend to take your cover letter very seriously. In economics less so.
In the end, it is a numbers game. Applying to more programs might not change your expected probability of admission very much, but it will reduce the variance.
Ideally, however, you will want entry into the top ten schools in your field because it keeps the most doors open, especially if you want an academic job. It’s not necessarily fair, but it seems to be the way the market works. Especially in economics, which seems to me to be the most hierarchical field in social science.
As far as I can tell, PhDs outside the top 10-20 schools are unlikely to lead to careers in research universities. This varies by discipline, but in the US the top 10 to 20 schools tend to staff the top 100 to 200 US universities. For those who graduate from lower-ranked programs, many opportunities remain open at teaching universities, think tanks, international institutions, government and the like. There are a lot of fulfilling research careers, and I am willing to bet that rates of job satisfaction are pretty high.
I would love to see (and will post) numbers on this if anyone has it for political science or economics.
Greg Mankiw also has advice on choosing a grad program.
Should you study economics or political science?
Noah Smith recommends an economics PhD if you’re not sure what PhD you should do. He’s even a little more emphatic than that: “Economics is the best PhD you can possibly get.”
This is a little suspicious coming from an economist. It helps to remember that most people like to make their students in their own image (I am no exception).
I think Noah’s advice makes sense if you like economics a lot, if you want to do highly mathematical research, and if you want to be assured of a job. That is why I did an economics PhD.
But if you are motivated by other questions, prefer other methods, or if your strengths are somewhere other than math, I don’t see how your path to fulfillment lies through economics.
If your interests are political economy (like a great many readers of this blog) you will be well served by economics, with these caveats:
- Economics, like any PhD, will socialize you to like what the profession likes.
- If you don’t like what the profession likes, you can either be successful in convincing them it is interesting, or struggle to be considered relevant. It’s hard to predict.
- As I have written elsewhere, it is extremely difficult to get jobs in other fields like political science, and people like me get them largely by chance.
Finally, there is the Sustainable Development PhD at Columbia. This is basically an economics PhD where people study applied sciences, health, environment, etc. The biggest mistake I see applicants make is mistaking this for a non-quantitative program. This is a hard-headed ultra-quantitative program for people who want to be on the frontier of both economics and science at the same time, and requires all the math requirements of economics to be considered (see below).
Okay, so what does it take to get into a top school?
For entry into the top 10 or 15 schools, it is exceptionally competitive. In short, focus on getting good recommendations, experience, grades and GRE scores.
After that, it’s hard to predict. As far as I can tell, most departments delegate admissions to a small committee of four to six faculty. The committee probably changes every year. Thus you never know who will be on the committee or what they care about. This adds randomness.
Some things that help, but that you may or may not be able to control at this point:
- Good grades. If you’re not at least an A- student it’s hard to make the case you are destined to teach or reach the research frontier.
- Economics applicants will want to have A’s in as many mathematics classes as possible.
- Night courses or an MA or MPA are common ways to make up for a patchy undergrad degree. That’s what I did.
- Note, though, many and perhaps most people we admit do not have an MA. The American PhD is designed for smart people to come with only undergraduate training.
- Strong letters of recommendation from professors. We like to see effusive letters from professors who know your coursework and research abilities well.
- This means that during or after your undergrad or MA you build relationships with two and ideally three faculty.
- Non-academic letters are discounted, since they can seldom speak to your ability to do what a PhD expects of you: produce great research.
- We take these letters very seriously. Professors typically specify in their letter how and how long we have known you and often give a sense of ranking relative to previous students we have recommended.
- So: Have you developed close relationships yet with professors in the field where you want a PhD? Start now. Work as an RA, take small classes, and remember that it’s better to get a great letter from someone less known than an okay letter from a well-known scholar.
- GRE scores in the 90-99 percentiles help a lot. Perhaps even worth retaking an exam for. This is especially true in economics.
- For most: quantitative methods. Economists (and sustainable development PhDs) probably need 2-3 semesters of calculus and statistics each, plus real analysis and linear algebra. Other courses (e.g. differential equations) help. Aspiring political scientists (except the theory/philosophy focused and some ethnographers) would be wise to do the same in calculus and statistics. Nine out of ten job market papers I see use quantitative theory or statistics to some extent, often inadequately. The bar is rising rapidly and those with basic math foundations have advantages. This includes the ethnographers, who often want to do multi-method work, integrating insights from game theory or run regressions. If so, 4-8 classes of methods preparation in undergrad is the minimum to be literate in half the work in your field.
- Relevant or interesting work experience. Unless you want to be an abstract theorist, 1-2 years of work experience, ideally research experience, before applying, in order to better develop your research skills, explore your interests and understanding of the literature and write a compelling research statement.
- I can’t speak for all schools, but each year I’ve served on admissions, most of the faculty on the committee discriminate against students that come straight from undergrad, at least in applied fields.
- Also see Mankiw on working before grad school.
- A compelling research statement. Basically, you should be able to articulate a concrete research question and how you would propose to answer it. More on this below.
- Outside funding. This won’t make a difference at all schools, but at many it can help. US students should apply for an NSF and foreign students may have a similar institution in their country. See my grant application advice.
Very few applicants have all of these things. Most are weak in one or two or three. So don’t stress out too much. Even so stress out enough that you do now what you can to improve your chances with the time you have.
A big piece of advice: Try to work on research projects with professors, because the best way to decide whether you want to do something is to try it out before it’s too late. Become an RA in your department, or start looking for RA jobs with professors in top departments in areas of your interest. This will also help with letters and your statement.
Thanks for the general advice, but what about you and Columbia?
If you want to know what it’s like to work with me, read this.
Most of the students I work with are interested in topics related to something under the umbrella of the political economy of development (micro- and macro-level), conflict and terrorism, political behavior (like voting or rioting or collective action), or causal inference. If your topic falls here, then I’d be a natural advisor for you, and I welcome new students. I commonly work with economics, politics and sustainable development) PhD students.
If these are your interests, other Columbia politics faculty you should investigate include John Huber, Macartan Humphreys, Tim Frye, Vicky Murillo, Page Fortna, Kimuli Kasara, Massimo Morelli, Dan Corstange, or (methods-wise) Don Green, Andy Gelman and Shigeo Hirano. In economics and related disciplines see Eric Verhoogen, Miguel Urquiola, Supreet Kaur, Suresh Naidu, Ray Fisman, Pierre Yared, or Jonas Hjort. This is just a partial list.
Do I need to have faculty advisers picked out in advance?
Yes and no.
“Yes” because your personal statement should demonstrate that you are a good fit with the department. In your applications you should be able to point to two to five faculty who, methodologically or topic-wise, do things that are relevant to you.
- The reason you want to identify multiple faculty is that we know things you do not: who has too many students already, who takes few students because they are solitary or retiring or on long leave, who has job offers elsewhere, etc. So don’t make your application hinge on one faculty member.
- Also, make sure the people you focus on are core faculty in the department, not adjuncts or someone in the law school, since these people seldom advise PhD students.
“No” because you think you know what you want to work on right now, but that will probably change three times. You haven’t learned much about the discipline yet, and it would be odd if two years of coursework and conversations didn’t change your mind.
Also, “no” because it’s rare to have a relationship and any kind of commitment to or from a faculty member in advance. Most of us tend to let the admissions process run its course before getting involved.
As a result, I don’t recommend contacting economics and political science professors in advance.
- This is different than psychology or some of the humanities or sciences where you are expected to have a specific advisor and relationship in advance.
- The reason is that we may get 1000 applications and a small committee may make 60 offers for 20 spots. It would be almost impossible to meet and screen people personally, and the majority of faculty in the department may not be clsely involved in the admission process that year.
- Even so, we faculty can get bombarded by emails from prospective students in the months before applications are due. Different professors deal with this in different ways, and I am guessing a majority don’t respond at all. I try to respond but only to explain that I engage in depth with students mainly after the committee has made offers.
A word on personal statements
I must say a little more about personal statements, because they often miss the point, and this is unhelpful and even painful to admissions committees.
This is not an undergrad entry essay where you describe your life’s trails and tribulations. And please do not start with the story about an epiphany, such as the day that you knew you wanted to study the subject. Especially if it involves a child in a poor country. In my opinion, this is mostly irrelevant and largely cliche.
You are applying to be a professional researcher, and this is your cover letter. Personally, I like to see some of the following:
- Your fields of interest
- Who you would like to work with in the department and why this is a good fit
- Make sure they actually are there and take students–that is, that they didn’t leave last year, are actually in the department you are applying to, and are tenure-track or tenured
- Your career objective(s)
- Concrete research ideas (this seems to be more important in political science, which likes to see if you can develop an interesting research question and propose a way to answer it)
- Important highlights from your CV, including any strengths that distinguish you
- Only if necessary: Information that might help us understand any apparent weaknesses or puzzles in your application (e.g. why you studied physics but now are doing politics, or what happened in that single bad semester, or what your foreign GPA means)
I’m pretty sure other professors will have different preferences here, so don’t take this as law. I think we’d all agree, however, that we need to know whether you will fit into the department and with whom, and it’s helpful when you give your view.
I have some sympathy here, coming from Canada. Even though it’s not that foreign, is foreign enough to create some hurdles in an already hurdle-strewn field.
As far as I can tell, it’s hard to get into a US PhD program without a MA or MPA or the like from your country or the US. Like it or not, this seems to be a pattern. If other profs disagree, let me know, because I’ll update this. This can even hold for the UK, Canada, etc.
A lot of foreign recommendation letters, especially those outside Europe, say very little about how they know you, how long, where their institution ranks in the country in research, what they think of your relative quality, whether they’ve sent grad students to the US before and where, etc. This tends to be helpful information and if you can find a diplomatic way to see if your professors are aware of the US norms, the better for you.
It’s also very hard for us to remember and track how every country grades their students. I wish students would make it easier for us. If your registrar or an online site can convert your GPA, do so. At minimum, I’d suggest telling us what it means in your personal statement.
I’m not sure about this, but I’d consider putting that conversion directly into the field online where it asks your GPA. Because many schools get from 1000 applications to the 200-300 they read in depth with a big spreadsheet of GPAs, GREs, school name, and a few other pieces of info. A blank GPA field either raises or lowers the chance they look at your application, and I don’t know which. There’s no simple solution or recommendation here. But this is something I think applicants ought to know about.
Comments and other perspectives welcome. I am also happy to entertain other questions. First see my advice on the right sidebar about success and fulfillment in a PhD, including (for the idealists like me) how to still save the world.