Quantitative Psychology: what are the PhD programs like?

[iamthesith4382]

Here's a bit of my background. Last year I applied to Clinical Psychology programs, but I did not get in (shocker). Instead I ended up going to a Masters program in Psychological Research to further beef up my research experience and cushion a 3.3 undergraduate GPA. In this program I took an advanced statistics course, and I thoroughly enjoyed it, and I think I'm pretty good at it. The professor teaching the course is a Quantitative Psychologist. Honestly I had never even heard of this area of psychology. This is likely to having gone to a tiny undergraduate university. I was really excited when I found out about this field because my favorite areas of psychology have always been methodology and analyses (I thrived in Research Methods). Overall, I am really interested in researching measurement. In the clinical realm, I was primarily interested in the measurement of personality with emphasis on pathology.  I have a strong research background at this point, and a lot of experience with SPSS. My major questions for you guys are as follows...

1) Is a Quantitative psychology PhD program a good place for someone particularly interested in measurement of personality and psychological disorders?

2) I do not have a background in calculus. Is this a problem? Will I be out of my league? I do have four semesters of research methods and three semesters of statistics. 

3) I am not particularly interested in creating new statistical methods myself. I am more interested in tackling other people's data and looking into multi-level modeling. Is that a problem?

4) Is there anything I need to do to improve my chances for getting into a program? I was thinking about maybe learning R or SAS or both. The quantitative professor at my university offered to do an independent study with me in R. Is this worthwhile?

[St0chastic]

Hi,

Not a quant psych person so take my advice with a liberal helping of salt.  It sounds like social/personality psych might also be a good fit for your interests.  

Re: calculus--I imagine that if you want to do hardcore quant work then yes, you'll need to know calculus and probably more advanced math.  Maybe a quant psychologist can chime in about this?

R: I'm a first year grad student currently learning R.  So far I really like it!  I would highly recommend taking up your professor's offer if you are able to.

EDIT: I also recommend checking out Datacamp if you want to learn R.  I think they offer a free intro module (although of course, you have to pay a subscription fee to access the rest of the modules).  This free online textbook is also worth checking out: https://health.adelaide.edu.au/psychology/ccs/teaching/lsr/

Edited October 6, 2016 by St0chastic

[socialbrouhaha]

I can't speak to the rest of the questions but I wanted to offer my 2 cents about learning R/SAS. Do it!! If it's not going to be a terribly out of the way project for you, then I would definitely take the opportunity. I am learning (trial by fire) R right now for my job and its so amazing and powerful and it'll give you an edge in case you want to explore careers outside of academia since R is more often used for real life data than SPSS. An aside: a friend of mine went to NYU's quant program and loved it. Good luck!

[iamthesith4382]

4 minutes ago, eternallyephemeral said:

That's definitely worthwhile, any research in that area would be good. I think the masters would also help. Did you do a thesis, or was it just course based?

I completed an undergraduate honor's thesis that is up for publication, and I am currently working on my master's thesis. Based on everyone's comments it would appear learning R is just a great skill in general. My honor's thesis did some multi-level modeling that was pretty cool, but i'm pretty sure I'll just be using a repeated measures ANOVA in my master's. I'm looking at the effectiveness of coping with stress/anxiety with personality/perfectionism as a moderator. 

[spunky]

Hello! Sure, I would be happy to help. I am a 3^rd-year PhD student (graduating within a few months, actually) and I work both in a Quant Psych lab and an Educational Measurement one, so I tend to see how things are done in both types of programs. Honestly believe you may benefit by looking a little into Educational Measurement programs as well.

Lemme address each one of your questions one-by-one and I’ll post everything in the forum so other people can benefit from them. I kind of feel like Quant Psych is “psychology’s best public secret” because, although it is a very interesting and fun area to be in (with lots of opportunities for growth and employment both inside and outside academia) there is little info about this area out there and its number of PhD graduates still remains quite low (at least per stats in North American programs). The Crisis of Replicability has been shining the spotlight on us a little bit (particularly because we, as methodologists/statisticians, have been forecasting this crisis for about 30yrs or more) so let’s see how things change in the future. In any case….

My big question to you, is what can I do to improve my application beyond the basic have a good GRE, good GPA, etc.? I am not applying for programs until the Fall of 2017.

 

For this question the advice is somewhat standard. Find a lab where you can gain research experience and volunteer. Ideally, a Quant Psych lab would be the best one so you can directly look into what goes in the daily research life of people in these types of programs. Quantitative Psychology can also be very mathematical so it wouldn’t hurt if you have taken Mathematics/Statistics classes outside of Psychology. I place emphasis on outside because, in my experience, courses in research methods/statistics for social scientists are a tad bit skimpy on the theory behind the methods and you want to learn how to do these things beyond the “cookbook” level. I mean, it’s not super necessary but it’s gonna look good on your application.

I spoke with a Quantitative professor that offered to teach me R in an independent study. Is that a good idea?

 

It’s more than a good idea… I’d say you’re probably gonna be expected to know some R, SAS, STATA or some other programming environment by the time to apply. But R is very powerful and popular so I would place the bulk of my efforts on learning R. I mean, you can apply without knowing any of this but then you’re gonna be stuck with both having to learn how to program while taking classes, undergoing research, etc. You’re also not going to look as good on your application package when compared with people who already know R. At this level, SPSS is just not gonna cut it anymore so don’t forget everything you know but be prepared to rarely use SPSS ever again. I think I haven’t used SPSS in more like 2 or 3 years? Everything I do is in R. So yes, the faster you can learn R, the better.

The other software I would recommend you to become familiar with is MPLUS because that is the default now on latent variable modeling. R can do a lot of what MPLUS can, but people just use it a lot so knowing MPLUS syntax will let you communicate with other people who don’t use R. It wouldn’t hurt you to learn about other programming languages and have some idea of how to do database management (SQL) or data-visualization (Tableau), but this is really not as necessary.

The one thing that you *should* start becoming familiar with is how to code Monte Carlo simulations. Your research as a Quant Psych person happens primarily inside the computer and simulations are our bread and butter. You are gonna end up running A LOT of those so try to become familiar with the basic structure of for() and while() loops, how to optimize computer time and (if you use R) the family of apply() functions. A book I recommend first year students to get themselves started with is this one:

https://us.sagepub.com/en-us/nam/monte-carlo-simulation-and-resampling-methods-for-social-science/book241131

so if you can at least start going through some of the chapters and try to reproduce some of the examples it would place you ahead of most people by the time you begin.

I’ll admit I don’t have a calculus background beyond Calc I, but I do have a course on multivariate statistics. Do you think this will be a problem?

 

Not really, but then again it depends on what your ultimate goal is. Most people (like yourself) find this field by accident so almost everyone who applies starts off without a strong base of Mathematics/Statistics. You will learn most of what you need as you go but the downside is that (a) you will only know things at an “intuitive” level that does not necessarily generalize to the wider types of data you will encounter and (b) you may not be able to read and use the literature produced by and for Quantitative Psychologists. The primary journal of our field is Psychometrika. It is the most prestigious place to publish and what most people aim for. But, to be honest, without at least some notions of calculus, linear algebra and mathematical statistics most people can’t make it past the first two pages or so of any given article. So, would this be a problem to get in? No, I don’t think so. But it can end up becoming a problem in the future.

As a curious side-note, I did my BSc in Mathematics with some hints of Psychology and I found it somewhat peculiar that, when I was having my interview, my advisor had highlighted all the Math courses I took and basically ignored anything related to Psychology. The interview also went into that direction and I feel the reason for that is because very few people with background in Math/Stats/CompSci etc. wander into Quant Psych so whenever that happens, advisors are very happy to snatch you from the get-go. Your application does stand over other people’s if you can palpably demonstrate some sort of technical expertise (where technical means theoretical math or an ability to code).

Is there anything that applicants say/do that is specifically a deal breaker in quant programs?

 

Uhm… not that I’m aware of? Just make sure you don’t show your preference for Bayesian statistics in front of a frequentist professor (<--- HOHOHOHO I’m so clever… Am I not? Anyone? OK, I’ll let myself out now… :D). Although this may tie in with another question of yours which is…

I am not particularly interested in creating new statistical methods myself. I am more interested in tackling other people's data and looking into multi-level modeling. Is that a problem?

 

For the most part, yes. This will become a problem for you sooner or later. And the reason it will become a problem is because you’re aiming at doing the most basic implementation (i.e. data analysis) of what Quantitative Psychology has to offer. It is also a problem because, in reality, any skilled social/ clinical/personality/insert-your-area-of-choice psychologist can do the same thing. As a Quant Psych your selling point is something like “not only can I do data analysis. I can do data analysis, I can create new methods for data analysis and I can evaluate data analysis methods”. If you stop at the “I just want to do data analysis” well, that’s not gonna get you very far. And that is something that in my opinion (and from talking to other people in my area in conferences and whatnot) would be a deal-breaker if you’re trying to get into a program. I mean, think about it… from the get-go, you’re already signalling that you’re not interested in doing what most of us in the field are doing so the immediate question that pops up is “is this person even supposed to be here?”.

If you’re mostly focused on data analysis over research on statistic and theoretical psychometrics then I would encourage you to apply into a more substantive program (social/clinical/personality/etc.) and just either do a minor in Quant Psych or take as many statistics/methods courses as you can. The fact of the matter is that a Quant Psych PhD program looks more like a watered-down Statistics PhD program (with a few exceptions, Ohio State comes to mind) than a Psychology program. You’ll find out soon enough that most of your research happens inside the digital bowels of a computer and not so much going out in the field and talking to real people. I mean, you do some of that but that’s definitely not what your training as a Quant Psych will do for you.

Is a Quantitative psychology PhD program a good place for someone particularly interested in measurement of personality and psychological disorders?

 

You can do that but if what you are really looking for is the measurement aspect of things and not necessarily the statistical aspect, a program in Educational Measurement might be a better fit for you. In my assessment, Quantitative Psychology programs are more programs in Statistics with some Psychometrics thrown into them, whereas Educational Measurement programs are more programs in Psychometrics with some Statistics thrown into the mix. I do find that Educational Measurement programs tackle some interesting aspects of scale construction and development (like how to create norms, psychometrically-sound ways to score tests, etc.) that do not necessarily make it into Quant Psych. And the reason behind this is Item Response Theory, IRT. Educational Measurement programs have been, for the most part, the bastion of IRT because the sample sizes you need to run these models accurately can easily go into the 1000s. And, at least form my experience, your standard Psychology research sample size is somewhere in the low 100s. Plus Educational Measurement programs place heavy emphasis on what happens outside the context of data analysis (The Standards for Educational and Psychological Testing will become your new go-to book for everything) whereas I  feel like Quant Psych programs place a lot more emphasis on the data analysis part itself.

 

 

Now, again, this is just a wide scope, generalization of how Quant Psych and Educational Measurement works. I’m sure if you look hard enough, you may find the one program with that one faculty member that does exactly what you want to do. But from the type of questions that you’re asking, I’m wondering whether Quant Psych is actually the right fit for you and if you may be better off in another program and just being very studious with your methods. Or perhaps an Educational Measurement program, have you looked into those? I feel the faculty in those programs is a little bit more diverse as far as research interests go. Another thing I would recommend you to do is to grab maybe some of the high impact journals in the field and have a look at what kind of research they publish. When you have the time, look into these 3 journals: Psychometrika, Multivariate Behavioural Research and the British Journal of Mathematical and Statistical Psychology. Go through the most recent issues, see the kind of stuff that gets published there and ask yourself: “is this the kind of research I would like to do for the rest of my professional career?” if your answer is “yes”, then Quant Psych is definitely your field. If your answer is “no” then… well, I think looking at other options might be worthwhile.

Edited October 7, 2016 by spunky

[The_Old_Wise_One]

6 hours ago, spunky said:

Hello! Sure, I would be happy to help. I am a 3^rd-year PhD student (graduating within a few months, actually) and I work both in a Quant Psych lab and an Educational Measurement one, so I tend to see how things are done in both types of programs. Honestly believe you may benefit by looking a little into Educational Measurement programs as well.

Lemme address each one of your questions one-by-one and I’ll post everything in the forum so other people can benefit from them. I kind of feel like Quant Psych is “psychology’s best public secret” because, although it is a very interesting and fun area to be in (with lots of opportunities for growth and employment both inside and outside academia) there is little info about this area out there and its number of PhD graduates still remains quite low (at least per stats in North American programs). The Crisis of Replicability has been shining the spotlight on us a little bit (particularly because we, as methodologists/statisticians, have been forecasting this crisis for about 30yrs or more) so let’s see how things change in the future. In any case….

My big question to you, is what can I do to improve my application beyond the basic have a good GRE, good GPA, etc.? I am not applying for programs until the Fall of 2017.

 

For this question the advice is somewhat standard. Find a lab where you can gain research experience and volunteer. Ideally, a Quant Psych lab would be the best one so you can directly look into what goes in the daily research life of people in these types of programs. Quantitative Psychology can also be very mathematical so it wouldn’t hurt if you have taken Mathematics/Statistics classes outside of Psychology. I place emphasis on outside because, in my experience, courses in research methods/statistics for social scientists are a tad bit skimpy on the theory behind the methods and you want to learn how to do these things beyond the “cookbook” level. I mean, it’s not super necessary but it’s gonna look good on your application.

I spoke with a Quantitative professor that offered to teach me R in an independent study. Is that a good idea?

 

It’s more than a good idea… I’d say you’re probably gonna be expected to know some R, SAS, STATA or some other programming environment by the time to apply. But R is very powerful and popular so I would place the bulk of my efforts on learning R. I mean, you can apply without knowing any of this but then you’re gonna be stuck with both having to learn how to program while taking classes, undergoing research, etc. You’re also not going to look as good on your application package when compared with people who already know R. At this level, SPSS is just not gonna cut it anymore so don’t forget everything you know but be prepared to rarely use SPSS ever again. I think I haven’t used SPSS in more like 2 or 3 years? Everything I do is in R. So yes, the faster you can learn R, the better.

The other software I would recommend you to become familiar with is MPLUS because that is the default now on latent variable modeling. R can do a lot of what MPLUS can, but people just use it a lot so knowing MPLUS syntax will let you communicate with other people who don’t use R. It wouldn’t hurt you to learn about other programming languages and have some idea of how to do database management (SQL) or data-visualization (Tableau), but this is really not as necessary.

The one thing that you *should* start becoming familiar with is how to code Monte Carlo simulations. Your research as a Quant Psych person happens primarily inside the computer and simulations are our bread and butter. You are gonna end up running A LOT of those so try to become familiar with the basic structure of for() and while() loops, how to optimize computer time and (if you use R) the family of apply() functions. A book I recommend first year students to get themselves started with is this one:

https://us.sagepub.com/en-us/nam/monte-carlo-simulation-and-resampling-methods-for-social-science/book241131

so if you can at least start going through some of the chapters and try to reproduce some of the examples it would place you ahead of most people by the time you begin.

I’ll admit I don’t have a calculus background beyond Calc I, but I do have a course on multivariate statistics. Do you think this will be a problem?

 

Not really, but then again it depends on what your ultimate goal is. Most people (like yourself) find this field by accident so almost everyone who applies starts off without a strong base of Mathematics/Statistics. You will learn most of what you need as you go but the downside is that (a) you will only know things at an “intuitive” level that does not necessarily generalize to the wider types of data you will encounter and (b) you may not be able to read and use the literature produced by and for Quantitative Psychologists. The primary journal of our field is Psychometrika. It is the most prestigious place to publish and what most people aim for. But, to be honest, without at least some notions of calculus, linear algebra and mathematical statistics most people can’t make it past the first two pages or so of any given article. So, would this be a problem to get in? No, I don’t think so. But it can end up becoming a problem in the future.

As a curious side-note, I did my BSc in Mathematics with some hints of Psychology and I found it somewhat peculiar that, when I was having my interview, my advisor had highlighted all the Math courses I took and basically ignored anything related to Psychology. The interview also went into that direction and I feel the reason for that is because very few people with background in Math/Stats/CompSci etc. wander into Quant Psych so whenever that happens, advisors are very happy to snatch you from the get-go. Your application does stand over other people’s if you can palpably demonstrate some sort of technical expertise (where technical means theoretical math or an ability to code).

Is there anything that applicants say/do that is specifically a deal breaker in quant programs?

 

Uhm… not that I’m aware of? Just make sure you don’t show your preference for Bayesian statistics in front of a frequentist professor (<--- HOHOHOHO I’m so clever… Am I not? Anyone? OK, I’ll let myself out now… :D). Although this may tie in with another question of yours which is…

I am not particularly interested in creating new statistical methods myself. I am more interested in tackling other people's data and looking into multi-level modeling. Is that a problem?

 

For the most part, yes. This will become a problem for you sooner or later. And the reason it will become a problem is because you’re aiming at doing the most basic implementation (i.e. data analysis) of what Quantitative Psychology has to offer. It is also a problem because, in reality, any skilled social/ clinical/personality/insert-your-area-of-choice psychologist can do the same thing. As a Quant Psych your selling point is something like “not only can I do data analysis. I can do data analysis, I can create new methods for data analysis and I can evaluate data analysis methods”. If you stop at the “I just want to do data analysis” well, that’s not gonna get you very far. And that is something that in my opinion (and from talking to other people in my area in conferences and whatnot) would be a deal-breaker if you’re trying to get into a program. I mean, think about it… from the get-go, you’re already signalling that you’re not interested in doing what most of us in the field are doing so the immediate question that pops up is “is this person even supposed to be here?”.

If you’re mostly focused on data analysis over research on statistic and theoretical psychometrics then I would encourage you to apply into a more substantive program (social/clinical/personality/etc.) and just either do a minor in Quant Psych or take as many statistics/methods courses as you can. The fact of the matter is that a Quant Psych PhD program looks more like a watered-down Statistics PhD program (with a few exceptions, Ohio State comes to mind) than a Psychology program. You’ll find out soon enough that most of your research happens inside the digital bowels of a computer and not so much going out in the field and talking to real people. I mean, you do some of that but that’s definitely not what your training as a Quant Psych will do for you.

Is a Quantitative psychology PhD program a good place for someone particularly interested in measurement of personality and psychological disorders?

 

You can do that but if what you are really looking for is the measurement aspect of things and not necessarily the statistical aspect, a program in Educational Measurement might be a better fit for you. In my assessment, Quantitative Psychology programs are more programs in Statistics with some Psychometrics thrown into them, whereas Educational Measurement programs are more programs in Psychometrics with some Statistics thrown into the mix. I do find that Educational Measurement programs tackle some interesting aspects of scale construction and development (like how to create norms, psychometrically-sound ways to score tests, etc.) that do not necessarily make it into Quant Psych. And the reason behind this is Item Response Theory, IRT. Educational Measurement programs have been, for the most part, the bastion of IRT because the sample sizes you need to run these models accurately can easily go into the 1000s. And, at least form my experience, your standard Psychology research sample size is somewhere in the low 100s. Plus Educational Measurement programs place heavy emphasis on what happens outside the context of data analysis (The Standards for Educational and Psychological Testing will become your new go-to book for everything) whereas I  feel like Quant Psych programs place a lot more emphasis on the data analysis part itself.

 

 

Now, again, this is just a wide scope, generalization of how Quant Psych and Educational Measurement works. I’m sure if you look hard enough, you may find the one program with that one faculty member that does exactly what you want to do. But from the type of questions that you’re asking, I’m wondering whether Quant Psych is actually the right fit for you and if you may be better off in another program and just being very studious with your methods. Or perhaps an Educational Measurement program, have you looked into those? I feel the faculty in those programs is a little bit more diverse as far as research interests go. Another thing I would recommend you to do is to grab maybe some of the high impact journals in the field and have a look at what kind of research they publish. When you have the time, look into these 3 journals: Psychometrika, Multivariate Behavioural Research and the British Journal of Mathematical and Statistical Psychology. Go through the most recent issues, see the kind of stuff that gets published there and ask yourself: “is this the kind of research I would like to do for the rest of my professional career?” if your answer is “yes”, then Quant Psych is definitely your field. If your answer is “no” then… well, I think looking at other options might be worthwhile.

This is very well thought out advice. Quantitative psychology is much like a statistics degree, and to really excel in it you will have to know advanced mathematics (i.e. calculus and linear algebra) as well as have a solid foundation in probability theory. The research consists almost purely of generating new mathematical models to describe cognition, learning, measurement, etc., and it takes quite a lot of time and effort to learn this sort of stuff if you have no prior experience. I am not saying it can't be done, just that you really have to devote all your time to catching up with people who already have the background. It is doable if you love it, as I know someone who was pursuing a developmental psych degree before he switched to quant psych (he had to take calc, linear algebra, and probability theory during one semester just to meet pre-reqs for other classes).

As for actually doing the research – you will not be using SPSS in a quant psych program. You will likely be using something like R along with other similar statistical programming languages. The reason for this is that SPSS comes with pre-packaged stat tools, whereas by definition a Quantitative psychology Ph.D. will have you focus on creating new models/methods that have not ben used before. Some above users have said that you will not be creating new methods... this is not true. You will most definitely be pushed to come up with new ways of doing analysis (e.g. creating novel mathematical models), and I would be very surprised if a Ph.D. program would graduate you without you first having done something novel. "Novel methods" does not mean paradigm shifting ideas, but you will at least have to create a variation of some existing model/method and show that it works better than other existing ones. 

Essentially, I would advice against pursuing a Quantitative Psychology degree unless you want to develop new mathematical models and/or methods of analyzing psychological data. If you just like to analyze data with existing tools, pick a sub-field of psychology that collects data that interests you and keep up with the state-of-the art methods in that particular sub-field. At the end of the day, the most important factor for grad school is that you maintain your interest in your day-to-day activities.

[The_Old_Wise_One]

On 10/6/2016 at 7:49 PM, eternallyephemeral said:

1) Is a Quantitative psychology PhD program a good place for someone particularly interested in measurement of personality and psychological disorders?

Some programs are personality and measurement, but yes most programs would allow you to do this, if that's what the professor studies.

 

2) I do not have a background in calculus. Is this a problem? Will I be out of my league? I do have four semesters of research methods and three semesters of statistics. 

I would say no, it's more about learning the statistics and the methods. Calculus is not the kind of math that's relevant, but you definitely have to be strong in math/analytic skills. I would say knowing your general linear models and other kinds of modeling is way more important. And either way, people who have stats probably just have one course in the first year of undergrad: you don't have to be a stats major.

 

3) I am not particularly interested in creating new statistical methods myself. I am more interested in tackling other people's data and looking into multi-level modeling. Is that a problem?

I don't think that's what most people do in quant psych (the first thing, creating new methods). I'm doing a measurement type project for my master's (I'm in organizational psychology),and it's more about tackling other people's measures and what they use to collect their data, and seeing if it's valid/which measure is the best one to use. But you should look at what the faculty do and what the program requirements are.

 

4) Is there anything I need to do to improve my chances for getting into a program? I was thinking about maybe learning R or SAS or both. The quantitative professor at my university offered to do an independent study with me in R. Is this worthwhile?

That's definitely worthwhile, any research in that area would be good. I think the masters would also help. Did you do a thesis, or was it just course based?

Addressing some comments made by eternallyepmemeral above:

(2) Advanced mathematics is definitely required for a Quantitative Psychology Ph.D. Probability theory involves complex integrals (i.e. it requires advanced calculus knowledge), and advanced statistics courses will require knowledge of linear algebra. The statistics courses you will take in a quant psych program will assume you can understand these subjects. 

(3)  Quantitative Psychology programs will expect you to come up with novel methods of analyzing data, so it will be an issue if this is not something you want to do. In fact, the degree will be focussed almost purely on you coming up with novel methods/mathematical models for analyzing psychological data. 

 

I say this having managed a Quantitative Psychology lab and worked with many quant pscyh Ph.D. students in the past few years. But you don't need to take my word for it – a search on wikipedia gives this information: 

"Quantitative psychology is a field of scientific study that focuses on the mathematical modeling, research design and methodology, and statistical analysis of human attributes and psychological processes.^[1] Quantitative psychologists research traditional and novel methods of psychometrics, a field of study concerned with the theory and technique of psychological measurement.^[2] At a general level, quantitative psychologists help create methods for all psychologists to test their hypotheses."

and from the APA website: 

"Quantitative psychology is the study of methods and techniques for the measurement of human attributes, the statistical and mathematical modeling of psychological processes, the design of research studies and the analysis of psychological data."

 

Looking at these definitions, it is obvious that quant psych programs will be focused on you creating novel methods. I don't mean to come off as rude, but people should refrain from giving advice when they do not have knowledge of a subject matter. Posters on this sub count on the community to give credible advice. 

[TenaciousBushLeaper]

To the OP, I would encourage a healthy skepticism of  some of the comments on this thread. Ask yourself if you enjoy the type of work you'd be doing in a quant program ? How, lets say "deep" do you enjoy going into the layers of "quantyness" ?  

If you decide to apply to quant programs and are accepted, maybe you can try this type of exercise, pick a well known method, try to understand it conceptually, look at the math, try working it out, and then give it a go at implementing it in a programming language. 

For example, start small with simple linear regression (one predictor -> one response variable), that would require basic algebra at most,  then move on to multiple linear regression, for efficiency, in what's called a closed form solution you'll need to understand some matrix algebra (because you'll need to compute dot products and matrix inverses), then try other methods of estimating the coefficients (which by the way is all you really need in regression, once you have those you can easily get a lot of other things). The other methods might involve optimization methods, the "argmin" you see in methods papers at times. 

This type of exercise might not be helpful for everyone, but it was definitely helpful for me.

 

EDIT: when I say implement I don't mean start up R and doing: 

lm(y ~ x)

and also...the exercise described above in no way attempts to mimic or relate a similar idea of the complexity of the work quant phd students might have to do.

 

Edited October 8, 2016 by TenaciousBushLeaper

[eternallyephemeral]

On 10/7/2016 at 11:33 PM, The_Old_Wise_One said:

Addressing some comments made by eternallyepmemeral above:

(2) Advanced mathematics is definitely required for a Quantitative Psychology Ph.D. Probability theory involves complex integrals (i.e. it requires advanced calculus knowledge), and advanced statistics courses will require knowledge of linear algebra. The statistics courses you will take in a quant psych program will assume you can understand these subjects. 

(3)  Quantitative Psychology programs will expect you to come up with novel methods of analyzing data, so it will be an issue if this is not something you want to do. In fact, the degree will be focussed almost purely on you coming up with novel methods/mathematical models for analyzing psychological data. 

 

I say this having managed a Quantitative Psychology lab and worked with many quant pscyh Ph.D. students in the past few years. But you don't need to take my word for it – a search on wikipedia gives this information: 

"Quantitative psychology is a field of scientific study that focuses on the mathematical modeling, research design and methodology, and statistical analysis of human attributes and psychological processes.^[1] Quantitative psychologists research traditional and novel methods of psychometrics, a field of study concerned with the theory and technique of psychological measurement.^[2] At a general level, quantitative psychologists help create methods for all psychologists to test their hypotheses."

and from the APA website: 

"Quantitative psychology is the study of methods and techniques for the measurement of human attributes, the statistical and mathematical modeling of psychological processes, the design of research studies and the analysis of psychological data."

 

Looking at these definitions, it is obvious that quant psych programs will be focused on you creating novel methods. I don't mean to come off as rude, but people should refrain from giving advice when they do not have knowledge of a subject matter. Posters on this sub count on the community to give credible advice. 

Sorry about that. I thought measurement and quantitative psychology were similar, as I have come across a lot of personality psychologists that call themselves quantitative psychologists, and they do not create novel methods. I'll delete my post.

[iamthesith4382]

Perhaps I should have been more clear in my original post. I don't view myself as the next person to come up with Cohnen's d, but maybe that has more to do with confidence. I've actually done a fair amount of multilevel modeling for repeated measures data. It was the most enjoyable aspect of my research experience. In addition, I would look into apply item-response theory to personality measures and scales. I think people are getting hung up on my comment about not wanting to create my models. It was more of a misunderstanding in how I wrote it.

I agree that a strong math background would make me a better applicant. Thankfully, my Master's program is very flexible so I'm pursuing several courses in the statistics graduate program. Also, I am going to do the independent study on R.

As for looking into a measurement PhD in Education Psychology. I did investigate that area, but I get hung up on the research interests. I'm not particularly drawn to what they are researching.

[The_Old_Wise_One]

On 10/9/2016 at 10:06 AM, eternallyephemeral said:

Sorry about that. I thought measurement and quantitative psychology were similar, as I have come across a lot of personality psychologists that call themselves quantitative psychologists, and they do not create novel methods. I'll delete my post.

No hard feelings! I just wanted to clear things up. There's nothing worse than delving into a Ph.D. program that you end up hating... I know some people that have gone through this and it is rough

[eternallyephemeral]

22 hours ago, The_Old_Wise_One said:

No hard feelings! I just wanted to clear things up. There's nothing worse than delving into a Ph.D. program that you end up hating... I know some people that have gone through this and it is rough

I totally understand, no worries! It seems the OP would definitely want something more like a personality/measurement program than a quant program, so it's good to hear what Quant IS and what it ISN'T

[esotericorigin]

1. like any area in psychology, the program will be a great fit if you can find a professor whose research interests match your own. there are plenty of quant programs that are measurement focused (psychometrics) and more general quant programs that have professors who work in measurement and assessment.

2. i came into my program with no calculus (except AP in high school, but who remembers that?). they required that i take the calc sequence once there, but it wasn't a big deal that i didn't have it at entry.

3. coming up with new methods is really only one aspect of quant, but what you describe is much more applied than typical for quant research. my master's thesis (and a lot of quant work) involved testing the effect of violating an assumption on a specific type of analysis - this often uses simulated data, not real data. 

4. i came in with very little experience on SAS or R. my program typically uses SAS, though a lot of students in other areas in the department use SPSS. that kind of learning experience would definitely be helpful, plus if you could get a letter of rec from a quant person, it would be a bonus.

[PsychStats]

The advise from most people here is excellent. Just adding an extra idea, since you mention you plan to learn R and the fact that you are not sure whether you will like methodological research:

Maybe when you are learning R, ask your professor to help you set up a small simulation study. It does not have to be a hard topic, just something simple and obvious, but it can give you an idea of what research is like in quant psych (which is very, very different from applied research). Obviously not all research in quantitative psychology is based on simulations, but it is a good introduction to seeing whether this is something that you might enjoy or not. If you end up loving it, then you are probably in the right field, and you can take the math classes later to get a stronger background. If you end up finding it repetitive and dry, then you know that this is not your area.

[MissData]

Hi All, I'd just like to add a few things. I think generally the question has been answered, but if there are more questions from Quant Psych PhD hopefuls I'm happy to answer. I'm a 3rd year PhD student in a Quantitative PhD program. I think each program is a little different, but here are some answers based on my experience.

1) Is a Quantitative psychology PhD program a good place for someone particularly interested in measurement of personality and psychological disorders?

I would say no. This seem like a very substantive interest. I agree with the recommendation that you check our some educational measurement program. You might also consider applying to a personality or clinical program that also has a quantitative program. That way you could do something like minor in quantitative psychology, take a factor analysis or item response theory class, and learn how to do scale development. The big different here is implementing a method compared to developing a method. If you spend your time thinking about how we could be doing scale development better (i.e. the recommendation to people building scales need to improve) that is inline with quantitative psychology. If you want to build a scale, that is more in line with substantive work.

2) I do not have a background in calculus. Is this a problem? Will I be out of my league? I do have four semesters of research methods and three semesters of statistics. 

I think this depends on the program. I would say very few students are admitted to our program without advanced calculus and matrix algebra. We have a number of students who double majored in psychology and either math or statistics. Many of our students also do a Masters in Applied Statistics while they are doing their PhD in Psychology. Calculus is very much required for understanding probability and statistics, and the ability to do calculus will likely come into play when doing research (particularly if you're doing Bayesian methods).

3) I am not particularly interested in creating new statistical methods myself. I am more interested in tackling other people's data and looking into multi-level modeling. Is that a problem?

Yes. To me the thing that almost all of our students have in common is a critical eye on statistical methods. They were asking questions about why we do things the way that we do from the beginning (like in introductory statistics classes in undergrad). If you spend your time in the class wonder "Does that always work? When doesn't it work? Are there methods that would work in a larger variety of circumstances?" then welcome to quantitative psychology. Our goal is to come up with new ways to analyze data in psychology and new mathematical models for psychological phenomenon. If your goal is to apply those methods to a specific question, then you need to find an area of research that is interesting and apply those methods. Again though, I would consider finding a school that has a quant program, so you can learn advanced methods from the professors who are innovating those methods.

I personally bandied back and forth between quantitative psychology and social psychology. The big thing that got me was that I was constantly unhappy with the methods that people in social were using and I wanted to improve them. I spend most of my time working on improving statistical methods that people are currently using. I do spend some time consulting and helping other people do their data analysis but this is a very small proportion of what I do.

4) Is there anything I need to do to improve my chances for getting into a program? I was thinking about maybe learning R or SAS or both. The quantitative professor at my university offered to do an independent study with me in R. Is this worthwhile?

High quantitative GRE scores and some indication that you know what quantitative research is. Try reading Psychological Methods or Behavioral Research Methods. Learning to program is really important. I came into graduate school proficient in SPSS and R. I've picked up SAS, MatLab, Python, and GAUSS on the way. It's really important that you be able to think like a programmer. Take a few Coursera courses or something on R or any other syntax based language. The ability to point and click on SPSS is not going to get you into a quant program. I agree with the previous recommendation to try our some simulations. There is a nice book Monte Carlo Simulation and Resampling Methods for Social Science which uses R and could be useful to work through. I think this book should be approach to anyone who's taken introductory statistical methods in a psychology department.

[artsy16]

I'm not familiar with quant psych, but I want to chime in and say I wouldn't recommend a clinical program to the OP. Clinical psych requires practicum and a yearlong clinical internship in order to graduate. If that's NOT what the OP either 1) wants to do after graduating or 2) is passionate about enough that they will put in effort into being a good therapist during practicum and internship, then OP will not only be miserable but also be doing a disservice to the clients that OP would be seeing throughout their grad program. Experimental psych programs have professors studying personality and other psychological disorders. Those do not require practica and are more up OP's alley in terms of interest in measurement. 

[socsciguy31]

Looks like I'm late to this discussion, but I will pass on a bit of what I know in hopes that it helps. I think the advice you've gotten is good, but I find some of it to also be a touch misleading, But keep in mind, what follows is all based on my experience, which is limited in ways. First, I agree that, yes, given your interests, you may want to consider an Educational Measurement program; however, with that being said, the difference between educational measurement and quant psych programs is sometimes very small (and I will stress very). Essentially, it's the focus and the questions that differ, the methods and methodological training are, in many cases, nearly the same. If you examine programs/programs of study and talk to and interact with people from both fields, you'll quickly come to realize this. I've had the pleasure of working with both, and I will be the first to say that at times the difference seems almost trivial, in my experience. I will say, though, that quant psych programs are generally more quantitatively-oriented.

As far as your mathematical background being an issue, it's going to (heavily) depend on the program to which you apply and your area of focus. But in all, I don't think it's an issue and I know of many quant psych/measurement students with little to no math background (you can pick it up along the way! And I agree with others, take such courses outside the Psych dept, if possible). I think this thread seems to slightly overemphasize mathematical training as some sort of requisite, with people even saying advanced mathematics is absolutely required for quant psych. Quant psych programs are not stats programs, and are thus much more forgiving in this regard (try getting into a PhD stats program without at least Analysis). Simply put, if they were stats programs, they'd be stats programs (except for maybe Ohio, I really can't tell the difference between that and a stats program lol). All jokes aside, they can be like stats programs, but advanced math a hard requirement? No, it simply isn't true, and I think some are forgetting or simply overlooking the fact that there is a lot of applied work being done in quant psych departments. A lot. Look in journals such as the Journal of Personality and so on and you'll see some great quant psych people from great quant psych schools doing more applied, but great, stuff. Many programs accept both types of students, those with more applied interests and those with more theoretical interests. Focusing more on the use (not creation) of analytic techniques flies in some quant psych depts. But now I will say this, the more math/theory the better lol. But many programs have a great deal of flexibility, and if your interests are more applied then that certainly shouldn't stop you. The fact you have stated an interest in a method (i.e., HLM), that's a good start right there.

Absolutely agree that programming skills are huge, and you should jump at the opportunity to learn R. Sometimes SPSS is useful, sometimes it's not. The fact is, you'll want to know multiple software packages (and languages, too, shoot, don't stop at R, pick up Python and Java, if you can). That's it for right now, if anything else comes to mind I'll be happy to give advice. I think the advice here has been good, I just wanted to add my piece.