Quantitative Psych - finding a "substantive" area of interest.

[bruin]

I have a question for any Quantitative Psychology people here.

Which "substantive" fields of psychology has your PhD research and/or career focused on, aside from pure quantitative stuff (publishing in methodology journals, etc.)?

Which "substantive" fields of psychology do you find yourself collaborating in most frequently, and what is your role in these collaborations?

For my PhD I am considering either quantitative psychology or bio-neuro-physio psychology. I'll admit, my passion is somewhat more for biopsych, but biopsych has one of the worst outlooks among the different areas of psychology, and quantitative has by far the best. And I do actually enjoy stats, I could definitely see myself as a quantitative psychologist, either in an academic role or working for, say, a standardized testing company.

The problem is that I'm not sure what my research would look like in quantitative psych. Health Psychology and Social Psych seem to be common "substantive" areas for quant psychologists to work in. I'm not interested in social psych, and health psych sounds interesting to the extent that it overlaps with biostatistics/epidemiology type stuff, but I've never actually had a class in health psych and would be a bit afraid to jump into a PhD where health psych was a major part of my specialization with so little background in it or understanding of what it is all about.

I am a very detail-oriented person so I suppose I would probably be able to get really into the measurement / test construction issues of the psychometrics part of quantpsych. But it is important for me to feel that I'm doing something important, and I'm not sure if I'd get that from simply working on building better personality scales or the like. Personality psych doesn't interest me and it seems like psychometricians in academic settings do usually study personality as their area of substantive interest. (Those in industry, obviously, work on a lot of standardized tests, which is different).

I'm also wondering to what extent behavioral genetics could be an area of significant overlap with quantitative psych. Obviously behavioral genetics is very quantitative, but I'm not sure how much overlap it would specifically have with quantitative psychology.

Any advice? Thanks in advance.

 

[spunky]

I have a question for any Quantitative Psychology people here.

there are three people on this board who very well intersect the quantitative/data analysis aspect of psychology with other substantive area, but given that i'm the only one in a formal PhD program in Quantitative Psychology/Psychometrics, i guess it's just fair that i give it the first shot (well, both that and the fact that i sleep very little.

Which "substantive" fields of psychology has your PhD research and/or career focused on, aside from pure quantitative stuff (publishing in methodology journals, etc.)?

i have to say i have consciously avoided "substantive" fields of psychology both because i don't really feel i have a good footing in other areas of psychology outside from quantitative methods and because i don't find them very interesting. still, however, i do a lot of consulting and side-projects mostly in the area of Education, so a lot of assessment and test development.

Which "substantive" fields of psychology do you find yourself collaborating in most frequently, and what is your role in these collaborations?

uhm.. EVERYTHING? the good thing about this area is that we are so few and in-between that we get called regularly to everything. from questions related to policy in education to well-being and quality of life stuff, i think i've done a little bit of everything. if you've already experienced being in a quantiatively-oriented program then you'll quickly realize you actually get more people asking you to collaborate on their projects than you can even manage.

For my PhD I am considering either quantitative psychology or bio-neuro-physio psychology. I'll admit, my passion is somewhat more for biopsych, but biopsych has one of the worst outlooks among the different areas of psychology, and quantitative has by far the best. And I do actually enjoy stats, I could definitely see myself as a quantitative psychologist, either in an academic role or working for, say, a standardized testing company.

quant psych. all the way. and leave biopsych as your side hobby 'cuz you don't wanna be one of those eternal post-docs barely making ends meet while competing with 1000 others for that one tenure-track position in some nameless community college in Anchorage. and no, i'm not joking. the job market out there for academia is NOT pretty and i could tell you so many financial horror stories from us, millenials, that what you REALLY wanna do is make sure you're super ready to take on the real world. where super ready = quant psych.

The problem is that I'm not sure what my research would look like in quantitative psych. Health Psychology and Social Psych seem to be common "substantive" areas for quant psychologists to work in. I'm not interested in social psych, and health psych sounds interesting to the extent that it overlaps with biostatistics/epidemiology type stuff, but I've never actually had a class in health psych and would be a bit afraid to jump into a PhD where health psych was a major part of my specialization with so little background in it or understanding of what it is all about.

your "substantive" area will be whichever area you want it to be. just hang around the lab you like the most and, trust me, people will come asking for your advice.

I am a very detail-oriented person so I suppose I would probably be able to get really into the measurement / test construction issues of the psychometrics part of quantpsych. But it is important for me to feel that I'm doing something important, and I'm not sure if I'd get that from simply working on building better personality scales or the like. Personality psych doesn't interest me and it seems like psychometricians in academic settings do usually study personality as their area of substantive interest. (Those in industry, obviously, work on a lot of standardized tests, which is different).

measurement is a nice area to dwell into. and the paycheque is REALLY good besides, it's not just about developing and validating scales. you get to learn a lot about the way in which the environment and people influence the process of answering tests, you get to learn about the culture of a testing environment, you get to learn about latent variables and item response theories and normal ogive curves and finite mixture models and... oh god, it's so pretty! (and did I mention the paycheque is really good? )

I'm also wondering to what extent behavioral genetics could be an area of significant overlap with quantitative psych. Obviously behavioral genetics is very quantitative, but I'm not sure how much overlap it would specifically have with quantitative psychology.).

again, there will an overlap if you want for there to be an overalp. example: I really like a particular statistical concept called copula distributions (don't ask). I think there are only like maybe 3 published articles in all the vast realm of quant psych where they get used, but they are routinely used in finance and actuarial sciences. but guess what? i'm using them! so i'm creating the overlap between quant psych and financial math. there are very interesting and challenging questions related to the design of studies used in behavioural genetics, like the analysis of dyadic data in twin studies. perhaps you could become an expert in dyadic data (so you get your quant fix) and you apply the development of your new methods in your "substantive" area (behavioural genetics). the nice thing about a quant psych PhD is that it's so flexible you can probably do whatever you want with it. just make sure you hang around with people who are doing "substantive" studies related to what you're interested in... or that your advisor would approve

 

[bruin]

quant psych. all the way. and leave biopsych as your side hobby 'cuz you don't wanna be one of those eternal post-docs barely making ends meet while competing with 1000 others for that one tenure-track position in some nameless community college in Anchorage. and no, i'm not joking. the job market out there for academia is NOT pretty and i could tell you so many financial horror stories from us, millenials, that what you REALLY wanna do is make sure you're super ready to take on the real world. where super ready = quant psych.

I agree, the prospect of being an eternal post-doc is terrifying. I teach psych classes at a community college right now with my M.A., and to go through a PhD, and one or more post-docs, only to end up back at a community college (albeit as a full-timer with benefits/security/better pay) is not exactly a great thought.

I really wish there was a way to combine biopsych with quant psych, but a lot of biopsych is just animal studies and they just do relatively straightforward ANOVAs. Human studies are obviously messier so would probably require more statistical expertise...maybe something like neuroepidemiology...

uhm.. EVERYTHING? the good thing about this area is that we are so few and in-between that we get called regularly to everything. from questions related to policy in education to well-being and quality of life stuff, i think i've done a little bit of everything. if you've already experienced being in a quantiatively-oriented program then you'll quickly realize you actually get more people asking you to collaborate on their projects than you can even manage.

Yes, I've gotten that sense that virtually everything is open to you when you go the quant psych route, but that's actually what scares me. I'm incredibly indecisive. My M.A. advisor was the quantitative/stats person in her department, and had no real "substantive" interest. I had to come up with my own topic from scratch. I actually was on the verge of giving up and quitting the program when I finished all my classes and still had not struck upon a thesis topic that worked all the way through to the end without some major problem ruining it. Then I was saved by the fact that she started on a new collaboration with two Clinical people, so my thesis became a spin-off of that, and after that I finished quickly. Still, my M.A. thesis took longer than it should of and was kind of a frustrating experience.

I do NOT want a repeat of that with my dissertation, so I'm very concerned about going into a PhD with a short list of highly specific "substantive" interests, a very good idea of the specific quant methods I would need to know in order to work in each of these areas, and an idea of which quant psych departments/advisors have some knowledge in those areas.

the problem is it seems a bit hard to get a feel for the substantive areas I could work in, because while the straight-up quantitative/methodological stuff is gathered together in quant journals, the "substantive" stuff that quant psych people collaborate on is probably all spread out in the journals for those substantive fields.

What you said about specializing in dyadic data if you want to collaborate on twin studies is a great tip, but I'm wondering how I can do a more comprehensive search for this kind of information? By which I mean, how can I make a comprehensive list of statistical methods one could specialize in AND the substantive areas within Psychology (and outside of Psychology) where they are used? maybe I should just be looking at quant psych departments, then peruse their faculty's publications, and for articles that sound interesting, go look them up and see what quant methods were used. Unless there is a more efficient way than this...?

Like I said, I'm really concerned about doing this BEFORE I would enter any quant psych program, because of my experience with the M.A. thesis.

 

[spunky]

I really wish there was a way to combine biopsych with quant psych, but a lot of biopsych is just animal studies and they just do relatively straightforward ANOVAs. Human studies are obviously messier so would probably require more statistical expertise...maybe something like neuroepidemiology...

well, just because people only use ANOVAs to analyze their data doesn't mean it's the most optimal way to do it. maybe you can come up with better ways to work around this. a professor of mine who taught me linear mixed models (aka 'multilevel' models) is into social/personality psych and once told me their filed was revolutionized by this ability to separate within- and between- people variance in repeated-measures studies. that before everyone was stuck in repeated-measure ANOVA and now the field, as a whole, is swtiching towards linear mixed models. so i think you can probably bring a lot to the table in terms of the appropriateness of the analysis if you're familiar with more complex methods.

My M.A. advisor was the quantitative/stats person in her department, and had no real "substantive" interest.

this i would say is the rule rather than the exception. i've found that in Psych Depts you either get people in substantive areas (cog, bio, personality, soc, etc.) who are REALLY skilled at methods stuff OR you get the quant psych type who feels a lot more comfortable working with computer simulations and math-related stuff. quant psych is usually a vast-enough field and there are so many methods/techniques you need to master that you don't have much time left to learn about other substantive areas of psychology.

I had to come up with my own topic from scratch. I actually was on the verge of giving up and quitting the program when I finished all my classes and still had not struck upon a thesis topic that worked all the way through to the end without some major problem ruining it. Then I was saved by the fact that she started on a new collaboration with two Clinical people, so my thesis became a spin-off of that, and after that I finished quickly. Still, my M.A. thesis took longer than it should of and was kind of a frustrating experience.

sorry to hear that. was your MA also in Quant Psych?

I do NOT want a repeat of that with my dissertation, so I'm very concerned about going into a PhD with a short list of highly specific "substantive" interests, a very good idea of the specific quant methods I would need to know in order to work in each of these areas, and an idea of which quant psych departments/advisors have some knowledge in those areas.

the problem is it seems a bit hard to get a feel for the substantive areas I could work in, because while the straight-up quantitative/methodological stuff is gathered together in quant journals, the "substantive" stuff that quant psych people collaborate on is probably all spread out in the journals for those substantive fields.

What you said about specializing in dyadic data if you want to collaborate on twin studies is a great tip, but I'm wondering how I can do a more comprehensive search for this kind of information? By which I mean, how can I make a comprehensive list of statistical methods one could specialize in AND the substantive areas within Psychology (and outside of Psychology) where they are used? maybe I should just be looking at quant psych departments, then peruse their faculty's publications, and for articles that sound interesting, go look them up and see what quant methods were used. Unless there is a more efficient way than this...?

Like I said, I'm really concerned about doing this BEFORE I would enter any quant psych program, because of my experience with the M.A. thesis.

i am starting to get the feeling from this and your previous post that you don't really seem all that into quant psych but what you really want to look into is more this behavioural genetics/biopsych stuff. if that is the case, i'd say you could end up walking a very iffy path if you decide that you want to do something for its employment prospects instead of because you like it. the fact of the matter is that there exists no "list" so to speak of which statistical methods are used within which areas of psychology. that is something you learn on the go because you're working on that specific area and you learn how the experts analyze their data. so if you go straight into a PhD i Behavioural Genetics/Biopsych you're guaranteed to become familiar with the family of methods that are best suited to analyze their data... but i think as a PhD in Quant Psych trying to look around for Biopsych-specific problems you'll probably end up frustrated with trying to look for this non-existent "list" of methods. do keep in mind that as a Quant Psych person your concentration is on the statistical issues and appropriate use of methods. we use "substantive" areas mostly as examples of why our methods work or how you could use them. i'm just getting this feeling that this is a very important point in your academic career where you have to look at yourself in the mirror and ask yourself if this is REALLY the path that you wish to follow.

 

[Jake]

Which "substantive" fields of psychology has your PhD research and/or career focused on, aside from pure quantitative stuff (publishing in methodology journals, etc.)?

Which "substantive" fields of psychology do you find yourself collaborating in most frequently, and what is your role in these collaborations?

My PhD (which I'm not finished with-- I'm in my fifth year) is actually in social psychology, so I do a lot of stuff in that vein in addition to my methods work. I have a little bit in political psychology, which is an area that is sort of nice for a methodologist because there are large, standard datasets available that you can run all your favorite advanced analyses on and publish the results. I do a lot of experimental work that is probably actually closer to cognitive psych than to social psych, things dealing with visual attention and whatnot. And I am just now beginning my first project that gets into clinical psych sort of issues. So I guess my point here is just that you can pretty much work on whatever you want other than the methods work. This other work does not necessarily have to be just an applied outlet for your primary methodological interests (although for many quant-type people, I think it ends up being that way.)

Often in these collaborations, when they are not projects that I am myself leading, I will pretty much fill the role of the designated data analyst and be the author that is primarily responsible for writing the method/results sections of papers. I guess this is fine. The data analysis is fun, but the writing is not too thrilling. One plus side of collaboration as a methodologist is that people will often come to you to recruit you for their projects in this way, so you can end up working on diverse projects and pretty much always having something to do.

My best advice for you about your substantive research interests is to find a PhD program where you KNOW that you will be able to work closely with at least one faculty member who is NOT primarily a methodologist. Working closely with them can help keep you grounded in basic psych research, and you will probably end up more or less adopting their substantive interests as your own. (This happens quite naturally... the more time you spent thinking about and working on those kinds of questions, the more they will come to interest you over time.)

About behavioral genetics. My department has a behavioral genetics program so I know at least a little about what they do. From what I can tell, there is VERY LITTLE overlap between what they do and what we do in quantitative psychology. If BG is what you really want to do, then you should pursue BG, not quantitative psychology. They have their own methodologically inclined people who focus on very different methods and data analyses. So maybe you want to be a "BG methods person." I guess there are some advisors you could apply to to pursue this type of path.

Hope this helps.

 

[bruin]

was your MA also in Quant Psych?

No it wasn't, it was experimental psych. That probably makes a lot of difference. Because from what you are saying, it sounds as if in an actual quant program, having your own non-quant interests within psych is much less important than I was thinking. You can easily get by doing a mixture of your own pure quant/methods stuff and collaborations which may be all over the map in terms of which subfield of psych they cover. (hoping I understood that correctly).

So what would you say distinguishes quantitative psychology from a statistics PhD program? Is it just the people you work with?

Also if I just have a background in psychology, not math, but I took the following classes before applying to a quant psych program, would I be prepared to do well in the quant psych program? (Would I also need a background in mathematical statistics or would that be part of the program? My background so far is only in "applied" stats, not mathematical stats).

Spoiler

224. Calculus III (4)
Vectors and three-dimensional analytic geometry. Partial derivatives and Lagrange multipliers. Multiple integrals. Vector calculus, line and surface integrals. Green's Theorem, Stokes' Theorem, and the Divergence Theorem.

233. Fundamental Concepts for Advanced Mathematics (3)
Fundamentals of logic and set theory, counting principles, functions and relations, induction and recursion, introduction to probability, elementary number theory, congruences. Introduces writing proofs.

247. Introduction to Linear Algebra (3)
Matrix algebra, solution of systems of equations, determinants, vector spaces including function spaces, inner product spaces, linear transformations, eigenvalues, eigenvectors, quadratic forms, and applications. Emphasis on computational methods.

347. Linear Algebra (3)
In-depth study of linear transformations, vector spaces, inner product spaces, quadratic forms, similarity and the rational and Jordan canonical forms. Writing proofs.

361A. Introduction to Mathematical Analysis I (3)
Rigorous study of calculus and its foundations. Structure of the real number system. Sequences and series of numbers. Limits, continuity and differentiability of functions of one real variable. Writing proofs.

361B. Introduction to Mathematical Analysis II (3)
Riemann integration. Topological properties of the real number line. Sequences of functions. Metric spaces. Introduction to calculus of several variables. Writing proofs.

463. Multivariable Calculus (3)
Topology of Euclidean spaces. Partial derivatives. Derivatives as linear transformations. Inverse and implicit function theorems. Jacobians, vector calculus, Green's and Stokes' theorems. Variational problems.

About behavioral genetics. My department has a behavioral genetics program so I know at least a little about what they do. From what I can tell, there is VERY LITTLE overlap between what they do and what we do in quantitative psychology. If BG is what you really want to do, then you should pursue BG, not quantitative psychology. They have their own methodologically inclined people who focus on very different methods and data analyses. So maybe you want to be a "BG methods person." I guess there are some advisors you could apply to to pursue this type of path.

Hope this helps.

Oh yeah, I see that you're in Boulder, they definitely have a lot of BG stuff going on there.

I'm not sold on doing BG by any means. I'm more just trying to get a feel for what a quant psychologist does, and I hit on BG as I was casting around for possibilities. I was under the impression that having a substantive area of interest outside of the purely quantitative stuff was absolutely essential, but perhaps I was wrong about that.

political psych is very interesting. I attended SIPP a few summers back and it was fascinating. I can definitely see how the huge survey datasets would be ideal for a quantitative specialist.

Much of the applied stats stuff that seems really interesting to me is biostatistics--clinical trials, epidemiology, behavioral genetics. But my background is firmly in psych and I already have to spend some time beefing up my math before any stats-centric PhD program, so I don't really want to also have to spend a long time beefing up my chemistry/biology/genetics. So I'm hoping that something in quant psych grabs me, but I'm having trouble understanding exactly what quant psychologists study, perhaps because they mostly do methods stuff that is at this point way over my head.

 

[Jake]

I think the educational background that you mentioned should be just fine for a Quant Psych PhD program.

 

[spunky]

No it wasn't, it was experimental psych. That probably makes a lot of difference. Because from what you are saying, it sounds as if in an actual quant program, having your own non-quant interests within psych is much less important than I was thinking. You can easily get by doing a mixture of your own pure quant/methods stuff and collaborations which may be all over the map in terms of which subfield of psych they cover. (hoping I understood that correctly).

yup, you understood just fine. had you been in an MA in Quant Psych you'd see there really is very little room for "substantive" interests and the research in methods themselves occupies most of your time.

So what would you say distinguishes quantitative psychology from a statistics PhD program? Is it just the people you work with?

i'm actually somewhat well-prepared to reply to this because i did my BSc in Mathematics before applying to graduate school. even though Quant Psych is much more demanding in terms of the level of technical sophistication you need, it is still an applied statistics program. you get some exposure to theory but not much. in a formal Statistics program, on the other hand, you need to be very well acquainted with statistical theory and the method of proof. you rarely (if ever) engage in formal mathematical arguments in quant psych, unless you want to. in Statistics there is no other option, you HAVE to engage in them. but this changes from program to program. i have two advisors and one of them has an MSc in Statistics so she is very well-versed in theorems, proofs and whatnot. but that's specific to my program in the University of British Columbia. most psych depts have very, very skilled data analysts, but they are not necessarily trained in formal mathematics.

Also if I just have a background in psychology, not math, but I took the following classes before applying to a quant psych program, would I be prepared to do well in the quant psych program? (Would I also need a background in mathematical statistics or would that be part of the program? My background so far is only in "applied" stats, not mathematical stats).

i would say it's an overkill but, by all means, take them all. Math/Stats students don't usually gravitate towards Quant Psych programs so they're not necessarily used to students showing up with lots of math under their belts. i've been in my dept for 4 yrs now and i've been the only student who did any formal math. now this has helped me out a lot because not only was i able to ace through my courses (mostly because i've already seen all the stuff from a much harder, rigorous perspective) but i also feel very comfortable in reading technical articles in journals like Psychometrika without feeling intimidated with all the Greek letters and calculus.

for Quant Psych i'd say Statistics, Linear Algebra and Calculus are your best friends. but they're definitely not necessary. many people get into these programs with only a BA in Psych and maybe a few advanved research methods courses.

but I'm having trouble understanding exactly what quant psychologists study, perhaps because they mostly do methods stuff that is at this point way over my head.

this is actually an issue that any student in quant psych faces whenever you start your program. it feels like a Statistics program but it's not quite Statistics... like a Psychology program but it isn't quite Psych. my take is the following: Statistics is a vast, vast area of knowledge. so people in its many sub-disciplines need to be trained to cater to the specific needs of their "substantive" peers. in quant psych you will learn a lot about Statistics but also a lot about other methods like Structural Equation Modelling and Item Response Theory which basically only exist for the social sciences. it's very hard to find courses or specializations in these areas in formal Statistics Depts. is it possible? oh, absolutely! is it common? not really. so as a Quant Psych in the making you'll make sure to become familiar with Statistics as a whole and the individual methods that are used across the various psychological disciplines.

have a look at Psychometrika. that's what lots of Quant Psych peeps do

oh, and btw, if you're an SPSS-only person and want to go into Quant Psych it's about time you ditch it and start learning how to program in R. it'll save you A LOT of time

 

[bruin]

Thanks! That was more than helpful.

 

[Lazar]

I am one of those Spunky referred to. Outside North America there really is no such thing as a quat psych PhD program or a specialised PhD program in any particular domain of psychology (indeed my PhD only has PhD on it and does not even refer to psychology anywhere). Thus I am not sure I have much to offer on a PhD program but in terms of post-PhD I can say that there is a real need for quantitatively sophisticated researchers in Educational Psychology. Ed psych is odd in that it has some of the most sophisticated quant folks and some of the least and a massive gap in between. However, governments and others are increasingly demanding quant based research and thus there is a need for people to fill those roles in Ed Psych departments around the world. Given many things in Ed Psych are not amenable to RCT there are many interesting statistical problems that need creative solutions.

EDIT: I echo the call to drop SPSS and learn R.

 

[bruin]

Thanks again for these replies, you guys are awesome.

one more question:

Math/Stats students don't usually gravitate towards Quant Psych...

Do you think this is equally true of biostatistics? I get the impression that biostatistics is kind of "hot" at the moment, so probably there are more math/stats people going into biostats than going into quant psych....does that seem correct?

 

[spunky]

I get the impression that biostatistics is kind of "hot" at the moment, so probably there are more math/stats people going into biostats than going into quant psych....does that seem correct?

it is, indeed, correct. biostatistics is pretty hot because big pharmaceutical companies pay BIG BUCKS to hire top-notch statisticians to design their clinical trials and whatnot. it's big enough that many university PhD programs (including mine) offer a concentration on biostats and clinical trials at the PhD level. and the people who survive that get snatched like crazy! besides, biostats always has some pretty kewl, interesting problems.

i dont think math/stats-types like myself gravitate easily towards social sciency stuff because it can be a real shock at first. there is absolutely no formal theory taught, most math is relegated to just hand-waving arguments, there's way too much "stuff" that has to be said and boring explanations... and if you come from an undergrad in math/stats to a graduate school program in the social sciences, you'll find out VERY quickly that you probably know more than the prof who's teaching this stuff... which is quite a disappointment.

honestly, if i had never met my 2nd advisor i would have probably quitted my program in frustration.

:: so Lazar, Jake and myself have replied to your thread. those are the 3 people i wanted you to listen (read?) so you'd get a more balanced perspective ::