Quantitative Psychology PhD

[spunky]

@spunky 

Your comments on being able to read math reminded me of something I had read earlier this year In the example the author explained  

"It follows easily that"  = "One can now check that the next statement is true with a certain amount of essentially mechanical, though perhaps laborious, checking.  I, the author, could do it, but it would use up a large amount of space and perhaps not accomplish much, since it'd be best for you to go ahead and do the computation to clarify for yourself what's going on here.  I promise that no new ideas are involved, though of course you might need to think a little in order to find just the right combination of good ideas to apply." 

 

 

Oh, tell me about *THAT*. For my MA's thesis I worked on extending and testing some of the developments published in this article. At some point it had one of those "after finding the first, second and third derivatives of the likelihood function it easily follows that..." but finding the first, second and third derivatives took me like a week!

[green_chair]

Comment on job prospects from someone who was on the market this season: There are lots of options. If your training is broad enough, the academic job options are in Psych, Education, and other departments. Ther past two years there have been about 35 tenure track job postings. There seems to be many places that are funding new data-focused research initiatives (I saw 4 positions related to something like this this year). Add to academia the huge demand for people with a quantitative training outside of academia and it doesn't seem like there will be a job shortage any time soon. So job prospects = good and far better than any other psych field.

Edited February 3, 2015 by green_chair

[juilletmercredi]

I just want to co-sign on everything @spunky said (I love @spunky's enthusiasm for quantitative psychology, lol!).  I did my PhD in social psychology but I am currently in a quantitative psychologist postdoctoral program.  There are excellent job opportunities, both academic and non-academic, for quantitative psychologists.  There are more academic positions than there are graduates, but there are also tons of private and government agencies that desperately need people who know how to do statistics but also know how to interpret them into useful social science information for the masses.  That's where you come in!

 

I also want to add that it's quite common for programs to have a quantitative minor - so you could get a PhD in cognitive psychology and minor in quantitative psychology.  One example of such a place is UCLA (or you could do the opposite - major in quant, minor in cog psych).  Every quant psych program encourages people to develop a substantive interest as well, so you could go to a quant psych program and study primarily cog or cog neuro problems if you wanted to!  Still other programs without formal quant programs will allow you to earn a concurrent master's in applied statistics (like Penn State or Yale) or have flexible enough requirements that you could take substantial coursework in statistics and potentially earn a master's in statistics if you wanted to (like Columbia, my own program.  My former advisor there is an excellent methodologist, btw.  NYU also has a quant minor with a methodologist I know from grad school - great guy).

 

 

if we can assume this is what quant research is like, it seems it's just mathematicians who also just so happen to have a keen interest in psychology...Well… we are and we are not...There are more people on the “applied side” who have a good intuition for data analysis and can do some basic matrix algebra/college calculus and that is enough for them. For you (and basically anyone interested in Quant Psych) this is my take on why it is much more beneficial to be a “statistician/mathematician interested in Psychology” rather than “a psychologist interested in Statistics”)

 

*raises hand* that's me.  I am much more of a psychologist with a keen insight and intuition for advanced statistical methods than I am a mathematician who has an interest in psychology.  That's partially a result of my training - I went to a social psych program and not a quant psych program, because I didn't know quant psych existed in undergrad - but partially a result of my interests, too.  I've noticed this at my postdoctoral fellowship, which is in a research center focused on methodology and statistics that employs a bunch of quantitative psychologists, statisticians, computer scientists, and social/developmental psychologists (and an epidemiologist).  There are some of us who are driven to develop statistical models because of an abiding interest in the theoretical statistics itself, and they need substantive scholars to bring applied problems to them that they can use to test the statistical methods they develop.  There are others of us (including myself) who have substantive questions that we want to answer, and so we work on teams to develop the methods because we need a method to answer the questions in which we're interested. Of course, it's not so black and white - I do have some theoretical interests in statistics and methods (and I love equations and Greek letters, lol) and some of the more mathematical folks do have substantive interests in certain fields; everyone at our center does drug abuse research, for example, because it's funded by NIDA.

 

You can take either path, but it's more lucrative to take the path in which you have more theoretical knowledge and are a statistician with a keen interest in psychology who can develop methods.  It's because there are fewer of you, and there are more opportunities to do data science in the non-academic world.  But...you should also follow your interests!

Edited February 16, 2015 by juilletmercredi

[TenaciousBushLeaper]

Thank you for your insights and comments! I recently mentioned quantitative psychology to my current cog neuro advisor and he essentially said it's a dying field,  and that the quantitative side of psychology is already incorporated in all the psychological disciplines. Although, to be fair, any time I bring a program / potential area of interest to him which isn't the program at my current university, specifically his lab, he sort of attempts to down play it. (I think he's trying to nudge me into applying to grad school at my current university, as a student in his lab lol ) 

I honestly just can't wait until I'm able to apply to quant programs, my days in undergrad go as follows, "why do I have to write so much? why am I taking graduate courses? I just want to do my probability problem sets".  The programs I'm hoping to apply to are UNC - Chapel Hill, UIUC, University of Minnesota, UCLA, Michigan State University, University of Michigan, & USC. Hopefully I can get into at least one of those. 

[spunky]

I just want to co-sign on everything @spunky said (I love @spunky's enthusiasm for quantitative psychology, lol!).  

 

Well, if I don’t get crazy excited about my field… who’s gonna do it?! XD

 

 

 

*raises hand* that's me.  I am much more of a psychologist with a keen insight and intuition for advanced statistical methods than I am a mathematician who has an interest in psychology.  That's partially a result of my training - I went to a social psych program and not a quant psych program, because I didn't know quant psych existed in undergrad - but partially a result of my interests, too.  I've noticed this at my postdoctoral fellowship, which is in a research center focused on methodology and statistics that employs a bunch of quantitative psychologists, statisticians, computer scientists, and social/developmental psychologists (and an epidemiologist).  There are some of us who are driven to develop statistical models because of an abiding interest in the theoretical statistics itself, and they need substantive scholars to bring applied problems to them that they can use to test the statistical methods they develop.  There are others of us (including myself) who have substantive questions that we want to answer, and so we work on teams to develop the methods because we need a method to answer the questions in which we're interested. Of course, it's not so black and white - I do have some theoretical interests in statistics and methods (and I love equations and Greek letters, lol) and some of the more mathematical folks do have substantive interests in certain fields; everyone at our center does drug abuse research, for example, because it's funded by NIDA.

 

You can take either path, but it's more lucrative to take the path in which you have more theoretical knowledge and are a statistician with a keen interest in psychology who can develop methods.  It's because there are fewer of you, and there are more opportunities to do data science in the non-academic world.  But...you should also follow your interests!

 

Statistics, data analysis and research methodology engages in that perfect dialectic dance where applied problems give rise to interesting theoretical questions and interesting theoretical questions provide solutions to problems that have not arisen yet… or that have appeared in other areas but haven’t quite made it into Psychology. John Tukey (that’s right, the Tukey from that ANOVA post-hoc test) invented the boxplot while analyzing the distribution of call frequencies at the Bell Lab. Harold Hotelling invented (actually more like re-discovered) Principal Components Analysis while answering a problem related to educational test scores. It is difficult (but not impossible) to find a good statistician who has never dealt with real data or who has never been inspired by it. I consider my students/clients some of my greatest sources of inspirations because seemingly innocent, simple questions can be extended into complex statistical problems if you know enough about them. Which is the main reason of why I am and will always be a champion of giving emphasis to theory before applications whenever Statistics are being taught. 

 

 

Thank you for your insights and comments! I recently mentioned quantitative psychology to my current cog neuro advisor and he essentially said it's a dying field,  and that the quantitative side of psychology is already incorporated in all the psychological disciplines. Although, to be fair, any time I bring a program / potential area of interest to him which isn't the program at my current university, specifically his lab, he sort of attempts to down play it. (I think he's trying to nudge me into applying to grad school at my current university, as a student in his lab lol ) 

 

Well… yeah. Lemme quote from this appropriately-titled article, The expanding role of quantitative methodologists in advancing psychology:

 

“A major contributor to the lack of statistical literacy in psychology researchers is the disciplinary attitude that “anyone can teach stats” (Wilcox, 2002). When psychologists without a strong quantitative methodology background teach statistics, they often lack the enthusiasm for, and knowledge of, modern statistical approaches.” (p. 85)

 

Your advisor is probably completely unaware of initiatives and grants like the $10 million dollar grant given to the Center for Open Science who are demanding Quant Psych grads, methodologists and statisticians in order to keep up with the demand for properly-trained data analysts who can oversee this new avalanche of replicability studies that are becoming so in-vogue nowadays. I could rant on and on about this stuff (and I’d love to!) but since you’ve already acknowledged that he’s biased, I’ll just let it be

[Gepetto13]

I guess the question is can someone who has taken a couple of statistics classes before and has an analytical mind/background actually get into such a program?  

[spunky]

I guess the question is can someone who has taken a couple of statistics classes before and has an analytical mind/background actually get into such a program?  

 

That's entirely possible. The majority of people who apply to Quant Psych programs are psych undergrads who took both the intro and advanced research methods sequence and somehow (usually accidentally) end up finding about this area of psychology. I think it is desirable (but not mandatory) to have a solid foundation in math/stats beyond the intro and advanced research methods courses with the benefit of hindsight, comparing my experience with that of my peers who didn't have a solid math base. But I find it difficult to believe that any program would just outright reject someone because he or she didn't go through the college calculus sequence or something like that. 

 

Just be willing to work hard, have tons of fun and you'll be OK

[Gepetto13]

That's entirely possible. The majority of people who apply to Quant Psych programs are psych undergrads who took both the intro and advanced research methods sequence and somehow (usually accidentally) end up finding about this area of psychology. I think it is desirable (but not mandatory) to have a solid foundation in math/stats beyond the intro and advanced research methods courses with the benefit of hindsight, comparing my experience with that of my peers who didn't have a solid math base. But I find it difficult to believe that any program would just outright reject someone because he or she didn't go through the college calculus sequence or something like that. 

 

Just be willing to work hard, have tons of fun and you'll be OK

Well that's good to know.  I did take Calc I and II, though that was years ago.  And, I've taken graduate level psych-stats and research methods courses.  So maybe I'd be OK….

 

But in a program like this, can you do a dissertation (or most of your research) on substantive interest areas where you apply advanced stats/methods?  Or do you have to come up with new/novel methods?

[TenaciousBushLeaper]

@ Spunky, 

I'd like to get your input on how one who's school has no quantitative psychology faculty would go about publishing a quantitative psych paper (maybe this question is too broad but let's see)

[husker123]

What are some of the best programs for quantitative psychology?

[spunky]

But in a program like this, can you do a dissertation (or most of your research) on substantive interest areas where you apply advanced stats/methods?  Or do you have to come up with new/novel methods?

[spunky]

@ Spunky, 

I'd like to get your input on how one who's school has no quantitative psychology faculty would go about publishing a quantitative psych paper (maybe this question is too broad but let's see)

[spunky]

What are some of the best programs for quantitative psychology?

[TenaciousBushLeaper]

It is a tad bit broad of a question, LOL. What’s your angle? Would you like to work on some theoretical result within statistics? Would you like to do a Monte Carlo simulation study comparing various methods? Would you like to apply some novel methodology to some interesting dataset? I think you'd need to narrow it down a little bit.

The first one (at least this is what appeals to me most), "work on some theoretical result within statistics". Although to be honesty I have no idea what this would entail however, I may have two years to get his done ! Since I might end up lab manager of my current lab which is a 2 year commitment. Anyways, point being, yes I am most probably starting at below ground zero but I have two years right right? D: lol. Any suggestions?  

[Gepetto13]

I think this is entirely advisor-dependent. For instance, Andrew Hayes is a social psychologist by training and a lot of his publications are in applied journals doing research in this area. However Hayes (alongside with Preacher) is one of the most prominent theorists of mediation analysis out there. I’m pretty sure someone like him would have no problems supervising someone who prefers to focus on other areas of psychology and apply complex methodologies to them. But you can have someone like Li Cai from UCLA who holds an MSc in Mathematical Statistics (among many other degrees) and whose interests are primarily in statistical methods and data analysis.

 

It does beg the question, however, of why you would like to specialize in Quantitative Psychology is your research interests are in other areas. Maybe doing something like juilletmercredi mentioned and get a minor in Quantitative Psychology would be much more suitable to you. 

[spunky]

The first one (at least this is what appeals to me most), "work on some theoretical result within statistics". Although to be honesty I have no idea what this would entail however, I may have two years to get his done ! Since I might end up lab manager of my current lab which is a 2 year commitment. Anyways, point being, yes I am most probably starting at below ground zero but I have two years right right? D: lol. Any suggestions?  

 

Well… uhm… to be honest, if you’re starting from “below ground zero” and don’t have an advisor to guide you through the process, I’m not sure how doable Option 1 would be for you (where Option 1 means working out some theoretical result). Usually, it does take some exposure to theoretical psychometrics or statistics to become acquainted with these types of problems and see how to go about solving them. From my experience, if something seems straightforward enough either people won’t care about it (because it’s too simple) or it’s not relevant enough. But here are some of the things in which I have (unsuccessfully) dabbled in, in case you want to take a shot at them:

 

- Tenko Raykov has been the only person I know who was able to derive the exact bias of Cronbach alpha, in the population, when the data follows a congeneric 1-factor model (different loadings, different error variances). He leaves open the problem of deriving the bias for finite samples. You could try and work on it.

 

- Ke-Hai Yuan (et.al.) derived a new definition, standard error and sampling distribution of Mardia’s multivariate kurtosis under data that is Missing Completely At Random (MCAR).There has been work (and I'm currently looking into it) to extend it to the more general case of  Missing At Random (MAR) but the standard errors/sampling distribution has been much more difficult to derive than anticipated. You could try a shot at that.

 

-  Donald Zimmerman extended the axioms of Classical Test Theory to measure-theoretic Hilbert spaces. In order to accommodate for more flexible types of norming not restricted by the inner-products, my advisor and I have attempted to extend them to Banach spaces. We haven’t been able to… but you’re welcome to try

 

If you want my opinion, I think Option 2 (a Monte Carlo simulation study) seems like a much more doable option if you’re mostly working on your own. You just need to become sufficiently proficient in statistics and R (or any other software but I’m R-biased LOL) and figure out how to program the right things there. There is A LOT of simulation work going on all the time and I feel like this is a problem that you could tackle.

 

Option 3 I am conflicted about. In general, I don’t feel like Quant Psych peepz really see analyzing data with a fancy method like anything particularly worthwhile (unless it’s a devilish design or something) because it really only says “oh look! I know how to type the right commands of code!”. A twist that I would give it to make it more interesting is maybe looking for similar problems in finance or economics or physics and bringing in those methods into Psychology. For example, in the Journal of Behavioural Statistics someone suggested the use of a method apparently used in by particle physicists to analyze likert-type data and settle, once and for all, the debate on whether you should analyze this type of data with parametric or non-parametric methods. Big Data and Machine Learning algorithms are really hot right now so you could maybe try and data-mine some big Mental Health database or something… I could see people being interested in that.

 

But in terms of what would get you published in a quant journal at this stage, I suggest Option 2. Although if you manage to solve any of the problems in Option 1 do let me know

 

I really hope someone else who is in Quant Psych would chip in here. I feel like the people who read this thread are only getting my perspective of things and I’m sure people would benefit more if more perspectives were brought into the discussion.

[spunky]

That clarifies things.  I think I'd like the training of a quant program so that, upon graduation, there's more flexibility as far as finding work.  I'm sure there are opportunities to publish substantive papers in grad school, or even do a diss that uses new/slightly modified methods in a cool study. I don't know if the minor in quant would do the same thing...

 

 

I see… yeah, that could be the case. Do keep in mind, however, that the bulk of your research would be in quantitative methods and statistical developments but I don’t see any reason as why you shouldn’t be able to pair that up with more substantive research.    

[TenaciousBushLeaper]

Well… uhm… to be honest, if you’re starting from “below ground zero” and don’t have an advisor to guide you through the process

Perhaps somehow I'll be able to coerce someone from the math department here into being an advisor lol

Thanks for all the suggestions, you really are a big help! (I feel like I should pay you, & if I wasn't dirt poor, maybe I would) 

[spunky]

Thanks for all the suggestions, you really are a big help! (I feel like I should pay you, & if I wasn't dirt poor, maybe I would) 

 

give me cookies....

 

.... OR GIVE ME DEATH!!! XD

[Gepetto13]

So then when writing a personal statement for a quant program, one should focus on researching statistical topics?  Or would it be good enough to state that experience at work has lead one to become fascinated with methodology and statistical methods for psychology research?  And to want to learn those methods and, perhaps, improve upon them?  

[spunky]

So then when writing a personal statement for a quant program, one should focus on researching statistical topics?  Or would it be good enough to state that experience at work has lead one to become fascinated with methodology and statistical methods for psychology research?  And to want to learn those methods and, perhaps, improve upon them?  

 

i think i would focus more on the statistical topics aspect of things and just tangentially mention that you also have substantive research interests and describe the ways in which you would like to apply the new methods you will learn to these other areas. that way you're conveying the idea that you're still committed to your quant knowledge as your core area of research.

 

i just fear you could face the potential risk of an advisor reading over your statement and thinking "uhm... this sounds like a social psychologist who lost her/his way. better throw it in the 'no' pile just to be sure". keep in mind that just because not many people apply to these programs does it mean they let in anybody. my program, for example, rejected all applicants for 3 consecutive years before getting in their new cohort of students even though we rarely, if ever, receive more than 10 applications per year. in the long run that's a good thing because that means advisors are seating on piles of sweet sweet unused funding waiting to be spent

 

we usually get around 5 or 6 applications from which only 1 or 2 people get an interview. on the bright side, you pretty much have the guarantee that your potential advisor will conscientiously read all of your application package for no reason other than it is not an insurmountable task. especially when you compare it with people who supervise social psych or clinical psych programs and get north of 300 applications every year.  

[TenaciousBushLeaper]

I fear this thread is getting too much attention.....more applicants will think about Quantitative Psychology programs......I WILL HAVE MORE COMPETITION....away you heathens away!  

Edited February 20, 2015 by TenaciousBushLeaper

[spunky]

.away you heathens away!  

 

 

this joke makes more sense if you've watched the last movie from "The Hobbit"

Edited February 20, 2015 by spunky

[TenaciousBushLeaper]

So a prospective POI for a quantitative program whom (I think) I have the best research fit with just emailed me saying she'd be happy to either communicate over email or talk over the phone DDDDD: 

[juilletmercredi]

I never saw the Hobbit but that macro cracked me up, lol.

 

TBL, I think you were the one wondering whether a minor in quant would do the same thing re: flexibility for finding work. It definitely can. Like I mentioned I work in a methods & stats center with a bunch of quant psychs. What I didn't go into is that quite a few of them actually have a PhD in developmental psychology with a minor in applied methods (or a master's in applied stats). One just got hired as a methodological professor in a social science department; she's spent years doing quant psych research and co-wrote The book on a widely-used technique. Another was a methodological professor for a few years before returning to the center (for personal reasons). A third just got several interviews for methodological positions and is waiting to hear back.

 

Also if you look at some of the big quant psychs in the field you'd be surprised to find out what some of them have their PhDs in. Todd Little's PhD is in developmental psychology. Patrick J. Curran's PhD is in clinical psychology (he did a postdoc in applied statistics). Kenneth Bollen actually has a PhD in sociology (he minored in math in college). And Andrew Hayes' PhD is in social and personality psychology (from his CV it looks like he started doing quant research towards the end of his PhD or shortly thereafter).