How defined do my research interests need to be (Statistics)?

[TenuousAtBest]

How defined should my research interests be when applying to Statistics PhD. programs? I have broad areas that I'm interested in, Computational Statistics and Statistical Learning Theory and Machine Learning, but I haven't narrowed it down any further. I have quite a bit of research experience in longitudinal data analysis and network science , but those aren't topics that I want to pursue.

 

Statistics is a graduate discipline, so how can incoming students be expected to have a deep enough knowledge of the field to know exactly what they want to research? 

[victorydance]

No one expects you to know exactly what you intend to pursue for your dissertation. However, your research interests should be narrowed enough to find some kind of gap in the literature and present some questions that you might investigate.

[stats_applicant]

No one expects you to know exactly what you intend to pursue for your dissertation. However, your research interests should be narrowed enough to find some kind of gap in the literature and present some questions that you might investigate.

I don't think they even need to be this detailed.  Talk about your broad interests and a bit about why you're not interested in pursuing what you've worked on.  You'll be judged primarily on your math preparation.

[bsharpe269]

I don't think they even need to be this detailed.  Talk about your broad interests and a bit about why you're not interested in pursuing what you've worked on.  You'll be judged primarily on your math preparation.

 

I agree that they probably don't even need to be that detailed. For example, I know my subfield will be biophysical modeling and I think that it would be cool to do this related to enzymes but I am also open to studying transcription factors, protein-protein interactions, energy landscape theory in general. I am interested in all types of models at this point and am still learning enough to know the advantages of different types. I have multiple undergrad research experiences in mathematical modeling along with over a year in a biophysical modeling lab and a couple of papers. The people who I will mentioning in my SOPs are all pretty well known in the biophysical modeing community so their research fits really well with mine even though I can't tell you exactly what I want to do.

If I were you, I would just talk broadly about your interests and then connect it to professors at the school. You don't have to say that you are interetsed in a very specific area but connect your braod area to the specific areas of professors.

 

I think that I actually know my subfield very well compared to the average applicant but I still couldn't really pick out specific gaps in the literature except realted to a couple things that I have already published on. My subfield is pretty huge so I have focussed on getting more of a general knowedge of the literature overall, rather than digging into specific areas.

[biostatdude]

How defined should my research interests be when applying to Statistics PhD. programs? I have broad areas that I'm interested in, Computational Statistics and Statistical Learning Theory and Machine Learning, but I haven't narrowed it down any further. I have quite a bit of research experience in longitudinal data analysis and network science , but those aren't topics that I want to pursue.

 

Statistics is a graduate discipline, so how can incoming students be expected to have a deep enough knowledge of the field to know exactly what they want to research? 

 

You answered your own question; since statistics is a graduate discipline, incoming students are NOT expected to have narrow research interests (unlike most other grad programs). Another reason incoming stat/biostat students are not expected to have narrow research interests is because students usually don't begin dissertation work until year 2 or 3 (if coming out of undergrad), so having narrow research interests before the program starts is not necessary. 

[cyberwulf]

No one expects you to know exactly what you intend to pursue for your dissertation. However, your research interests should be narrowed enough to find some kind of gap in the literature and present some questions that you might investigate.

 

This is (with apologies for my tone) simply nonsense. Most folks entering stat/biostat PhD programs apply just after starting their senior year of college; many have had little or no exposure to statistical research. The vast majority of faculty making admissions decisions understand that expecting these students to identify a "gap in the literature" when they clearly don't yet have the skills to even understand the literature is ridiculous.

Edited July 27, 2014 by cyberwulf

[TenuousAtBest]

This is (with apologies for my tone) simply nonsense. Most folks entering stat/biostat PhD programs apply just after starting their senior year of college; many have had little or no exposure to statistical research. The vast majority of faculty making admissions decisions understand that expecting these students to identify a "gap in the literature" when they clearly don't yet have the skills to even understand the literature is ridiculous.

 

Right! That's exactly what I thought. So many people on the forum have specific interests that I thought I was behind, though.

 

 

You answered your own question; since statistics is a graduate discipline, incoming students are NOT expected to have narrow research interests (unlike most other grad programs). Another reason incoming stat/biostat students are not expected to have narrow research interests is because students usually don't begin dissertation work until year 2 or 3 (if coming out of undergrad), so having narrow research interests before the program starts is not necessary. 

 

So broad interests all schools are looking for.  When I write my statement of purpose, should I identify a few faculty members who publish in my subfields of interest and name them? Is there any reason to try to reach out to potential advisors if most applicants don't have a specific research area in mind?

[stats_applicant]

Right! That's exactly what I thought. So many people on the forum have specific interests that I thought I was behind, though.

 

 

 

So broad interests all schools are looking for.  When I write my statement of purpose, should I identify a few faculty members who publish in my subfields of interest and name them? Is there any reason to try to reach out to potential advisors if most applicants don't have a specific research area in mind?

 

 

Spend half a paragraph mentioning a few names who may be doing work broadly related to your interests.  Don't bother contacting them.

[wine in coffee cups]

Not defined. Few to none are coming in with anything even resembling a research proposal, it's just not how things work in statistics. (Unlike other sciences, so pipe down, non-statistics people!)

 

I think it can be helpful if you have interests in certain areas or applications to discuss them in your statement of purpose, since faculty working in that area may have research assistantship funding and you might be given a slight preference in admissions if you're a strong applicant who could fill that slot. This was the case for me: I ended up on an RAship with my advisor coming right in because I wanted to do work related to a topic that he had a grant for, and I was entering with relevant background and an excitement for the area that I don't think many other applicants had.

 

It's not an especially common situation in our discipline, though, because many new students are fresh out of undergrad and haven't had exposure to either the methods or applications they might work on. It's probably more important to come across as mathematically prepared, intellectually curious, and an overall interesting person who someone would want to advise. The particulars of your research interests don't matter too much beyond establishing that you could fit into the department. Students also change their minds substantially, so it's hard to take lists of interests too seriously unless you actually have some experience to suggest a commitment. Off the top of my head, I can think of statistics students who shifted from genomics to demography to network models, from topic modeling to genetics, from stochastic processes to nonparametric theory, from spatial methods to algebraic statistics, from dynamic modeling to empirical processes, and from social networks to ecology.

[victorydance]

Sorry, I made an assumption (wrongly I guess) that statistics was similar to most other graduate programs where your research interests have to be relatively defined and you need to craft some researchable questions. 

[Biostat_Assistant_Prof]

Just as most others above have said, you really aren't expected to have narrow interests before beginning your graduate program, especially if you're coming straight out of undergrad. I had a general idea of a few areas of research that interested me based off my conceptual idea of what they were, but beyond that I had very little knowledge on what research in those areas actually entailed. After the first year, I've narrowed it down to a relatively specific area, but even that's kind of rare in my program... everything fell perfectly into place for me, and I was lucky. I know people going into their third year still trying to narrow what area they want their dissertation work to be in. 

 

If you want to get an idea, sure go ahead and read papers to get conceptual ideas of the different areas, but knowing what area you want to pursue comes after you take the classes and get involved with projects after you start, and the graduate programs know that. 

[cyberwulf]

Sorry, I made an assumption (wrongly I guess) that statistics was similar to most other graduate programs where your research interests have to be relatively defined and you need to craft some researchable questions. 

 

Definitely not meant as a personal attack on you, I just get frustrated seeing the same thing being parroted all over the place. Frankly, I'm a bit skeptical that stat/biostat are huge outliers on the importance of the personal statement; the PS may be more important in other fields, but I suspect its impact is exaggerated across the board.

[victorydance]

I don't know, I don't think it's just about your SOP though. I think having a pretty targeted approach to your application can help a strong applicant pull away from the pack. For example, if an applicant knows the field and recent work done, has done an honours or masters thesis on a sub-topic, has done RA work in that topic, and can show that they can put forth researchable questions in their SOP, I think that is a positive not a negative. Especially if they are tying it directly to POIs that work in the department.

 

This not may be relevant to everyone; for example, those who are switching fields or those who weren't super interested in something during their undergrad. But this is how I am approaching my applications.

 

I think there are dozens upon dozens of applicants out there with general interests and stellar academic records. There aren't many who have become rudimentary-experts in a certain topic and want to pursue it further. 

[StatsG0d]

One area of interest I had in particular was bridging the gap between econometrics and machine learning. More specifically, I would like to apply econometric techniques such as instrumental variables and quasi experiments to machine learning analysis for causal interpretation. The only problem is that I feel this may be too specific to write about and that the subject itself is presumably audacious. Any thoughts?

[cyberwulf]

One area of interest I had in particular was bridging the gap between econometrics and machine learning. More specifically, I would like to apply econometric techniques such as instrumental variables and quasi experiments to machine learning analysis for causal interpretation. The only problem is that I feel this may be too specific to write about and that the subject itself is presumably audacious. Any thoughts?

 

You do have to be a bit careful about being overly specific, for two reasons: 1) You may not have a good grasp on what exactly you're proposing, and come off as a bit clueless or naive; 2) If nobody on the faculty is interested in the project you want to work on, there might be some concern about lack of "fit". 

 

EDIT: What you've suggested actually does sound relatively reasonable, but the set of faculty working at the intersection of causal inference and ML is pretty small, and the subset familiar with the econometrics literature is even smaller.

Edited July 31, 2014 by cyberwulf

[Igotnothin]

"What's you've suggested actually does sound relatively reasonable."

 

High praise!

 

Even if the SoP isn't terribly important, I don't see the purpose of campaigning for people to whip it together in a couple hours. Can't hurt to make sure it's a well-written and thoughtful essay. I personally wouldn't feel comfortable submitting a generic half-page SoP, especially to a department I really wanted to join. "I have no exposure in statistical research, but here is my application. Enjoy!"

[cyberwulf]

Even if the SoP isn't terribly important, I don't see the purpose of campaigning for people to whip it together in a couple hours. Can't hurt to make sure it's a well-written and thoughtful essay. I personally wouldn't feel comfortable submitting a generic half-page SoP, especially to a department I really wanted to join. "I have no exposure in statistical research, but here is my application. Enjoy!"

 

You're misinterpreting what I'm saying. A shoddy SoP will obviously make you look bad; it still needs to be professional, and furthermore you never know which faculty are going to put more weight on the SoP. I'm just advocating against spending excessive amounts of time agonizing over exactly how to phrase things, which professors to name-drop, etc. 

Edited July 31, 2014 by cyberwulf

[Igotnothin]

Fair enough!

[StatsG0d]

I see your point, cyberwulf, and thanks for the advice. Perhaps my proposal is better suited for a NSF GRFP than for an SOP.