Quantitative Researcher, get a masters or PhD?

[ATLBEAR99]

I'm trying to figure out what programs to apply for this fall and could use some advice. I graduated in Spring 2022 with a bachelors degree in "quantitative social science" from Emory University. My degree is pretty much a joint major between data science and political science. For the last year, I've worked as a data analyst for a policy research team inside a large national non-profit. I really love my job, as I get to do quantitative research on policy while also getting to do some "internal consulting" with our various programmatic teams when needed. I'm really happy being in this type of applied research role and would like to stick with this type of work, but just work my way up to being more of a research director running a policy research team. Obviously that's years away from now, but that's my general goal.

Pretty much everyone I work with has a graduate degree of some sort, mostly masters, but I have noticed on the research side of public policy there are a good number of folks with PhD's too. I fully recognize that I need to get a graduate degree to move upwards, but I'm really conflicted on which degree to get for a variety of reasons. I've been mulling over getting a masters vs Phd for awhile and I feel like both are reasonable options, but I'm not sure what is more ideal.

Since my end goal is to be a research director type of role, I'm not really sure if a masters or PhD would be better for getting to that point. I really would like to reach the point of setting my own research agenda and running a small team to work on that, which I feel like a PhD gives more credence too. I don't particularly care where I work (academia, think tank, public sector, private, etc). I'm willing to work in academia, but I understand that job market is horrendous so its not a goal of mine to be a professor. I also have seen for other non-profits and some think tanks, folks with masters seeming to get pretty high up. So not sure which degree best helps me get to my career goals.

Another reason I feel conflicted is that I think a PhD would give me better training, especially considering where I already am and what I want to do. In my undergrad program, I did a pretty robust amount of coursework in stats + data science methods(full calculus sequence, linear algebra, probability & statistics, linear regression analysis, some elective coursework in NLP and data viz) and now regularly spend a big chunk of my work day in Rstudio. As a part of my job, I've also ended up learning a lot more about databases, geospatial analysis, simple machine learning models,  programming with API's, etc. And I would like to continue to expand upon that knowledge and get into more complex topics in the realm of stats and CS. And even when I look at at some of the public policy masters that also are marketed with data analytics, it seems like a lot of it would be a refresher and filling in some gaps, but not learning much new material. On the other hand, I've noticed a lot of PhD programs tend to be more flexible in the coursework you take and are more willing to let you take classes from their stats & CS departments, which I think would also be a large benefit for me. My research work will likely stay rather quantitative and computationally focused, so I would like to spend more time adding onto my current skills. 

 

Also I'm not sure how my research interests would fit well into a masters degree program. I'm interested in urban and social policy, in particular how housing policy affects neighborhood dynamics and resident access to public/private amenities. However, another huge interest of mine is how social and political dynamics affect policy-making and implementation. I want to do research on how politicians, media, and just regular folks speech effect the policy-making process. Also with advances in NLP, I think there are a ton of interesting ways to study the effect of public speech on policy. It's also admittedly a niche research area(using big data techniques to study how people talk about urban/social policy) so I'm not sure if a masters degree program would have the resources to help with me studying that. Obviously research interests can and will change, but I feel like my interests tend to stem into other domain areas enough that I feel hesitant to do a masters.

So I've given a lot of reasons for why I want to do a Phd, but there are a lot of reasons to just get a masters. I can get a masters degree way faster than a PhD, which would be really nice. Living on a PhD stipend for 5 years does not excite me at all. I grew up without much money, so I know I can handle it, but I've really enjoyed making a living wage the last year. Its been nice to be able to go out regularly and not pinch every penny. Also realistically, a lot (vast majority) of non-academic jobs don't need a PhD even if it would be nice, so it could be a waste of 3 years of my life to get a Phd. 

I know that's a lot of thoughts babbled into text, but I would love if anyone could give me some advice on what type of program to lean towards.

[2020applicant...]

I am writing a response because I went through some similar dilemmas while applying to graduate school. A few points:

1. I think that a key point here is that in undergrad you took a "full calculus sequence, linear algebra, probability & statistics" and that at work you "regularly spend a big chunk of [your] work day in Rstudio." Given this background, you have more options than the average MPP applicant. You could realistically apply to statistics/data science masters degrees, certainly policy MPP masters degrees, and almost any PhD in the social sciences (though specifically for the Econ PhD, a course in measure theory/real analysis may be expected depending on the program). Having this range of options is great but intimidating. As a first step I might casually counsel the following:

Regarding the masters degree ->

- If you were to pursue the masters level, make sure that you go to a program where you can actually leverage all the work you did in undergrad. E.g. you took linear algebra and vector calc, meaning that you have the requisite knowledge to take statistics/econometrics coursework that is mathematically rigorous. Many MPPs do not offer that type of coursework. Some, like Berkeley and Chicago, allow you to waive core coursework and pursue those more mathematically rigorous alternatives, typically PhD level econometrics. If you were to pursue an MPP, make sure that you can access those types of courses and can waive the typical quant offerings.

- Many policy+data analytics masters degrees are, at the most basic level, probably seeking to get the average student up to the level of the advanced undergrad in an econ+data science double major. You are, to some extent, already at that baseline. There is, of course, some variation across programs -- E.G. Chicago's MSCAPP has some rigorous discrete math and data structures coursework* that other programs don't have -- but many programs seem to center around some sort of data wrangling course + basic applied machine learning + applied econometrics core which may not be of interest to you. This doesn't mean these programs have no value in your case, but you would need to be discerning -- 1.) don't pay more for one of these 2.) if you go with this route make sure you can waive core classes where appropriate and have access to high quality, graduate-level electives in Computer Science, Econ, and Statistics Departments. This is true of many schools, especially Chicago and Berkeley.

Regarding the difference between PhD and masters level ->

- social science PhD-level quantitative training is not, as a rule, more rigorous than at the masters level. Econ is certainly different in this respect, as it is standardized in being very mathematically advanced, but Sociology and Political Science core quant sequences typically don't cover more ground than a quantitatively-inclined MPP student with access to electives around campus, indeed many would cover less. That said, the appeal of the PhD is less the coursework and more the research. You would pursue your research agenda for ~3 of the 5 years, and be forced to learn cutting edge quantitative tools to that end. I know several social science PhDs who took at most 1 applied course in NLP, but are experts on the topic because they then spent 3 years self-studying the topic to finish their dissertation. In the masters route, you would get to take that same 1 NLP class, but you wouldn't necessarily have those next 3 years to obsess over the applications of NLP you have in your research. My view - that is generally where the quantitative gap comes up between the masters and PhD level, less so in the coursework. If you got a masters then worked on advanced quantitative problems, either in research or data science for an agency of some sort, you would be able to similarly continue developing, but these pathways are not the most typical for the average MPP student.

- "My end goal is to be a research director type of role." Becoming a research director, in academic labs, or the public/private sectors is certainly associated with having a PhD. That said, I have seen directors who have masters degrees leading quantitative research groups in industry and in government The biggest qualification seems to generally be sustained interest in the topic. People with interest in quant policy research seem to sort into PhDs, and so leaders in that space often have PhDs, but they aren't a hard req outside of the academy.

Regarding everything you actually said ->

Everything you described about yourself seems to be oriented towards research. You work in research now, you want to continue working in research, and the things you are interested in ("how housing policy affects neighborhood dynamics and resident access to public/private amenities [...] how social and political dynamics affect policy-making and implementation [...] how politicians, media, and just regular folks speech effect the policy-making process.") all sound a little bit more like Sociology/Political Science dissertation topics than things one might work on in government. You can certainly do this sort of work with a masters, but if you really love research and have a research agenda, that is traditionally what a PhD is all about. If you love the idea of using quant tools to implement policy, evaluate and design real programs, and change the social world, then one might lean a little bit away from the PhD and towards the masters, but if you love research and want to use quant tools to do research and gain the respect of other researchers via publication in journals, then there you go.

General advice ->

Go look up authors on papers from Brookings/Urban Institute/etc. that use advanced tools like machine learning, NLP. etc and see what they did, go look up directors of research orgs you would one day want to lead and see what they did, etc. Feel free to dm if you want more of my perspective, though I certainly wrote like a books worth here.

good luck!