How technical/specific should my grad school personal statement be?

December 13, 2007 6:10 PM

Grad School Filter: How technical and specific should my statement of purpose for my Computer Science PhD application be?

I'm applying to PhD programs in computer science. The prompts for the statement or purpose say include things like:

- How I became interested in doing cs research
- Current projects or experience that demonstrate I know what I'm doing
- Future plans

My question is:

Given that I don't have a huge amount of space for all this (usually between 1 and 2 thousand words), how technical and department specific should I make my essay?

I could easily spend 1000 words explaining how I became interested in the field, but I also have like 1500 in my technical current experience document. I'm trying to pull it all together and I don't know how much of each flavor to mix in.

Thanks for any and all help!

posted by christy to education (10 comments total)

The more important information will be the type of research that you want to pursue. To support that, you want to describe specific projects that give you some experience in that area. You could put some small amount about how you became interested in the first place to soften your essay and make it more engaging.
posted by demiurge at 6:26 PM on December 13, 2007

Speaking as a computer science professor...

You need enough technical detail to signal that you have depth and can understand complex material. Far more important, however, is evidence that you are familiar with what are the main problems of your desired area of work, and evidence that you are a person who can be fairly independent and can get a lot of things done without a lot of handholding.

Just as important is the ability to write and communicate clearly. When I applied to computer science grad schools, I didn't realize how much of my time would be formulating new directions of research and trying to communicate that with others. Furthermore, I've noticed that people who can write well tend to do better in terms of critical thinking.

Good luck!
posted by jasonhong at 6:26 PM on December 13, 2007

The advice I've seen suggests that the technical bits are more important, since the letter is meant to demonstrate that you have the skills and experience to do good research. I would make a couple of paragraphs' introduction describing your passion for CS and how you got that way, something punchy to draw the reader in, then spend the bulk of the letter on your experiences and qualifications. You can then reference the first part of the letter in your closing, where you describe your future plans and talk about how they will allow you to realize your specific CS passions. (for instance, if you write that you got into CS because you love optimizing search algorithms, you could refer back to that in explaining why you want to work at Google.)
posted by fermion at 6:29 PM on December 13, 2007

(jasonhong--I enjoyed your anti-phishing game paper)

I'd say future plans are the most important section. Your resume, recommendations and grades should illustrate your ability in the field, while the future work section shows how good a match you'll be to a professor and his/her interests.
posted by null terminated at 6:37 PM on December 13, 2007

I think the three major sections to the SOP are (1) What is your main research interest? (2) Why do you want to pursue that interest at this institution? (3) Why are you the right person to pursue this research? The "how did you become interested" and "future plans" parts of the prompt are, to me, largely filler text. If you can't fill 2000 good words about sections 1-3 above (which doesn't necessarily mean you're a bad candidate — lots of people coming out of undergrad don't have a whole lot to say about their past experience beyond their transcript), then dive into those questions. My balance was about 3/2/5, though I don't know if that makes sense.

Of course, if you have a particularly compelling future plans story, by all means tell it. But saying "I want to be a professor some day" isn't that gripping. Maybe you really care about energy technology, and you want to learn PhD level CS to tackle that problem as an entrepreneur or something. That's getting a little more interesting. Same goes for the "why CS?" question. I think it's a stand-in for evidence of past success in the field. If you can't describe great CS stuff you've done, then at the very least the committee wants to hear you get excited about the field in general and tell a compelling story about why it's a good path for you. But I think storytelling should take a back seat to the three main sections.

I would shy away from too much technical specificity. While it might show your technical chops, I think you're running the risk of saying something that some professor disagrees with or doesn't like for some arcane reason. Better to stay a little high level and paint a story about a research space you're really excited about. Besides, your documentation about past successes is what will really prove you know the material. Of course, you gotta know the field.

Good luck!

I'm no expert, but I did apply (successfully) to a CS PhD program two years ago. I didn't end up going to that program so I don't really have an inside scoop, but I suspect that my current department isn't that different, even though it's not CS. Also, you're welcome to email me if you have more specific questions.
posted by heresiarch at 6:39 PM on December 13, 2007

thanks very much everyone, that's all very helpful advice! they are all 'best answers' at this point, but it seems silly to mark every reply.
posted by christy at 9:16 PM on December 13, 2007

You have to remember that part of applying is trying to attract the attention of the professors in terms of how good a fit you are to their department. You could be an awesome candidate, but if the research you list as your interests isn't something they do, want to do, or have funding for, you will most likely be turned away. It's a fairly large investment on their end (considering the amount of work and money that goes into both the academic and adminstrative parts), so you have to convince them that you're worth it.

Good luck!
posted by spiderskull at 2:50 AM on December 14, 2007

You have to remember that part of applying is trying to attract the attention of the professors in terms of how good a fit you are to their department. You could be an awesome candidate, but if the research you list as your interests isn't something they do, want to do, or have funding for, you will most likely be turned away

Most grad admissions committees are not looking at your statement of purpose expecting to see a research interest that jives perfectly with the department's current projects. They want to see a spark of true interest in something that shows that you have thought this whole grad school thing through at least a little bit, plus the communication skills mentioned above. And that is if they even bother to read your SOP. I wouldn't worry about pandering to the current department research projects, just don't be totally boneheaded and talk about how much you love AI to a department with no faculty in that area at all.
posted by ch1x0r at 6:39 AM on December 14, 2007

Following up on some comments above, we have actually not accepted very good students simply because we didn't have any faculty who could plausibly advise them (based on what they wrote in their statement of purpose). This is more about finding a good match (in terms of advisor and funding) rather than anything about the student.

I'd also suggest tailoring your statement of purpose to the university you are applying to (mentioning projects that excite you), but not so much just because you want to get in (ie don't say you want to do programming languages if you really want to do robotics).
posted by jasonhong at 7:54 AM on December 14, 2007

I don't know what's in your current technical experience part of your application, but I would suggest doing something that ties your current work / research to the set of questions that is being addressed in the subfield that interests you most. And I would, if possible, further tie this work to specific faculty members at the institution that you're applying to. Mention them by name. Mention some of their recent work, and discuss how "it's a good first step to understand [foo]" or "a novel solution to problem [bar]" and then how you might extend, generalize, apply, or continue that work.

Being specific can help you, by first showing how well you know an area of CS research and how well you've understood some interesting problem in the area. And by calling that professor out by name, he or she is sure to read your application (usually after another committee member finds your app and skims it... he'll just walk it down the hall to the faculty member(s) you mention), and perhaps be convinced that he or she would like to take you on. Even if they don't, it can still help a lot to seem like you're already "part of the club" and other faculty might appreciate this, even if they're not in your sub-area.

Basically, your research statement should have something of a bibliography, if that's not obvious from the questions that are asked.

And finally: don't worry that what you put in your application essay will hem you in if you're accepted. It won't. Your research interests may change substantially during your coursework and early years as a student... So don't feel like you're sealing your fate to work with professor A on problem B by so stating in your application.
posted by zpousman at 7:55 AM on December 14, 2007