MS in Statistics During PhD? Worth it?

[TXInstrument11]

So, I'm thinking about grabbing up a master's through the stats department at my school instead of the regular psych MS built into my PhD.

 

My thinking is, it will be good for me as a researcher because I will have more methods at my disposal, allowing for more varied experimental design and greater independence (less need for data analysis help from other faculty).

It also may also give me a competitive hiring edge later on. 

 

 

BUT the rub is that I have practically zero advanced math right now, necessitating a slow crawl through several calculus classes taught by grad students. If I magically secured an override now, it would take me Fall & Spring for Calc I and II, and then an additional Summer & Fall for Calc III and the other prereqs.

 

That means 4 semesters for only the damn prereqs!

 

Realistically, I would bone up on my trig now and then take Calc I in the Spring, so 5 additional semesters before even starting the MS.

 

 

 

Math isn't a great love for me, but I do pretty well in it - maybe even better than the average STEM student on aptitude. So, it's doable, just difficult and time-consuming. I am squeamish about adding an extra year and limiting all of my electives to stats, but maybe it's worth it. In the grand scheme of things, one year isn't that much time and if I shouldn't be penalized by my program according to the handbook. 

 

 

NOTE: my tuition is covered and, as far as I know, does not have a year cap. 

 

 

 

 

[GeoDUDE!]

You haven't taken trig and you want to get an MS in statistics?

 

What makes you think you can even handle a MS in stats let alone do one during a PhD ? 

 

What does doing an MS get you that self learning doesn't ?

 

What you call advanced math most people take in High School (if you consider trig and calc 1 and intro statistics advanced). 

 

Wouldn't you be better spent developing and working on your dissertation? 

 

These are  some questions/comment you should be answering and thinking about

Edited August 5, 2015 by GeoDUDE!

[TXInstrument11]

I aced precalculus/trig and qualified for calc. I just didn't take it because I was going to a humanities major at the time and didn't think it was worth the effort.

 

Most people do not take calc in high school. You need advanced placement or dual enrollment for it and that's not 50+% of the school population, so I'm not sure what you're talking about. 

 

An MS in stats gets me something tangible that sets me apart from other people, either as a researcher or on the job market. 

 

Also, this stats master's is designed for PhD students. I'm not signing up for it on top of my requirements. Instead, I would be subbing in additional stats classes to cover the typical psych MS and the "minor" requirements, which I was going to put into stats anyway.

Edited August 5, 2015 by TXInstrument11

[TenaciousBushLeaper]

My advice, get pdf of a typical calc book used for calc 1, and get another one of linear algebra(these pdfs are abundant online and a number of them are meant to be free). Start working through them like...right now up until the point when you have to make your decision. Once you're there evaluate how you've done. 

How far have you gotten into the books? 

Did you manage to work on problem sets every day?

How was your motivation to do this?

Also, I wouldn't recommend spending too much time on a trig prep, for most "early transcendentals" calc books you can do just fine with learning the trig as you go along, when and where you need it. 

Lastly, really evaluate your motivation to do this, you might end up doing not so well in your math courses simply because you aren't as motivated to do all of the required work, on top of the work your doing for your PhD.  

edit: 

by required work I don't just mean w/e homework is assigned (if it's even assigned). I mean that you may or may not have to do additional work, on top of that which is required to make sure you have the concepts down. 

Edited August 5, 2015 by TenaciousBushLeaper

[ExponentialDecay]

I've had this conversation in the past, and most professors said that, whilst taking stats in the math department may give you broader preparation, you end up learning a lot of stuff that isn't used in your discipline and, conversely, miss out on techniques that are not standard-fare in statistics taken broadly, but which are necessary for your field. If you're interested in statistics per se, taking the MS through the math department may be a good idea, but otherwise you may be confusing yourself and giving yourself extra work without much payback. Also, if your psych department is considerably better than your math department at your school, the stats given in the psych department may actually be stronger. As regards getting a professional edge, I don't know; maybe it would be a good idea to ask people working in your desired field for advice rather than a motley crew of current and potential students.

[GeoDUDE!]

My advice, get pdf of a typical calc book used for calc 1, and get another one of linear algebra(these pdfs are abundant online and a number of them are meant to be free). Start working through them like...right now up until the point when you have to make your decision. Once you're there evaluate how you've done. 

How far have you gotten into the books? 

Did you manage to work on problem sets every day?

How was your motivation to do this?

 

This is good advice.

 

Can you do this while being ABD?

 

Not to come of like a hardass, but again the math you have taken is nothing like linear alegebra with math majors. Plenty of people get As in math in HS. I think its a bit premature to be considering a graduate degree in math when someone hasn't taken any college math to begin with. 

[rising_star]

If what you're really interested in is research methods, is there a way to get a graduate certificate or master's specifically in methods, where you might take classes in stats, program evaluation, qualitative research, etc.? I'm thinking of something like the UGA program in qualitative research... You might also just consider taking a more quantitative-focused set of stats classes offered by sociology, political science, or economics at your institution. The only time you usually see people asking for a certain number of hours of math and stats courses are for federal jobs as a statistician (go to usajobs.gov and look some up to see what I mean). Otherwise, people just want to know that you are knowledgeable and capable, which you'd be able to demonstrate without having a separate master's in statistics.

[TXInstrument11]

This is good advice.

 

Can you do this while being ABD?

 

Not to come of like a hardass, but again the math you have taken is nothing like linear alegebra with math majors. Plenty of people get As in math in HS. I think its a bit premature to be considering a graduate degree in math when someone hasn't taken any college math to begin with. 

Rereading my post, I did come off as cocky and should have phrased some of what I said differently.

 

I was basing it off of the people I know in STEM majors that I knew in HS. They always think they're hot shit and I know I understood the math better and got better grades in high school than they did. I would be very surprised if they suddenly became math prodigies come uni.

Edited August 5, 2015 by TXInstrument11

[TXInstrument11]

If what you're really interested in is research methods, is there a way to get a graduate certificate or master's specifically in methods, where you might take classes in stats, program evaluation, qualitative research, etc.? I'm thinking of something like the UGA program in qualitative research... You might also just consider taking a more quantitative-focused set of stats classes offered by sociology, political science, or economics at your institution. The only time you usually see people asking for a certain number of hours of math and stats courses are for federal jobs as a statistician (go to usajobs.gov and look some up to see what I mean). Otherwise, people just want to know that you are knowledgeable and capable, which you'd be able to demonstrate without having a separate master's in statistics.

 

I can minor in it. I didn't know about this kind of degree being viable for statisticians. Thanks

I've had this conversation in the past, and most professors said that, whilst taking stats in the math department may give you broader preparation, you end up learning a lot of stuff that isn't used in your discipline and, conversely, miss out on techniques that are not standard-fare in statistics taken broadly, but which are necessary for your field. If you're interested in statistics per se, taking the MS through the math department may be a good idea, but otherwise you may be confusing yourself and giving yourself extra work without much payback. Also, if your psych department is considerably better than your math department at your school, the stats given in the psych department may actually be stronger. As regards getting a professional edge, I don't know; maybe it would be a good idea to ask people working in your desired field for advice rather than a motley crew of current and potential students.

 

Another grad student in the program said something similar. I initially considered taking stats through the math department because it looked like I was stuck with the bad apple for one of our first stats classes in the department, but managed to slip in when someone dropped the alternate professor's class. And yeah, you're probably right. I have contacted my adviser, but they have been busy. 

My advice, get pdf of a typical calc book used for calc 1, and get another one of linear algebra(these pdfs are abundant online and a number of them are meant to be free). Start working through them like...right now up until the point when you have to make your decision. Once you're there evaluate how you've done. 

How far have you gotten into the books? 

Did you manage to work on problem sets every day?

How was your motivation to do this?

Also, I wouldn't recommend spending too much time on a trig prep, for most "early transcendentals" calc books you can do just fine with learning the trig as you go along, when and where you need it. 

Lastly, really evaluate your motivation to do this, you might end up doing not so well in your math courses simply because you aren't as motivated to do all of the required work, on top of the work your doing for your PhD.  

edit: 

by required work I don't just mean w/e homework is assigned (if it's even assigned). I mean that you may or may not have to do additional work, on top of that which is required to make sure you have the concepts down. 

Yes, I think that would probably be the best course of action for seeing if I'm still capable of it. I was thinking about working in some hours per day/week to work entirely on math in order to do this and have researched a few free resources to learn calc.

 

I suppose I am worried about being able to understand calc well enough to apply to data analysis. I wasn't sure if I could self-teach some of these things later on. It does sound as if the broad education from the math department might make it harder though. I suppose it depends on how well they tailor it to education and psychology students. The head of the program is actually from education, but it's technically under the stats department.

Edited August 5, 2015 by TXInstrument11

[spunky]

I guess this can be either a hindrance or an advantage depending on how you work with it.

 

I come from a Math background and jumping into a social science/education/psych background did help me bring in some skills that I know your average student in these programs does not have. Most of the methodology courses that you take in Psych or Ed departments are much more focused on the application of statistics or how to use the methods. Very little attention is devoted to the actual theory or why they work. Whereas this probably serves the needs of your average graduate student, it does leave them at the disadvantage of not knowing how to proceed if they’re dealing with an unusual dataset with complicated dependencies (spoiler alert: those are the types of datasets where the most interesting results are found). A degree in Statistics might not prepare you to become a skilled data analyst in the social sciences but it will give you the necessary background to jump in with new ideas. But then again if you plan on working on research areas where there are standard methodologies in place, then there really isn’t much value in getting an MSc in Statistics aside from fostering your own personal knowledge.

 

An MS in stats gets me something tangible that sets me apart from other people, either as a researcher or on the job market. 

 

 

Will it make you a better researcher? Well, it depends on what you’re researching, right? It made me a good researcher because my research is very technical in nature. But if say you were interested in doing research in... oh I don’t know, standardized tests or scale development then it would benefit you much more to get some sort of degree or specialization in Psychometrics and not formal Statistics (very little to no Psychometrics are covered in a standard Statistics department). If you see yourself working in areas like neuropsychology or neurobiology, maybe a degree in Biostatistics (with emphasis in fMRI imaging which is a VERY hot topic) would serve you better. It seems to me that if your ultimate goal is to be a researcher more than a methodologist, you’d need to tailor your degree to cover the methodologies that are used in your substantive area of content. Or you can always become just a methodologist… I mean… we’re kinda short on those right now

 

Will it make you more employable in the job market? Oh, it most definitely will! But here’s the catch… will it make you more employable in jobs you are interested in? Like, big pharma companies pay well for DNA sequencing analysts… but last time I worked on something like that I wanted to gouge my eyes out of sheer boredom.

 

So all in all I think this is not a bad idea per se as long as you have a good game plan. I mean, you’re still committing a year of your life to something that you’ve already admitted to you’re not super passionate about.

[spunky]

 

I suppose I am worried about being able to understand calc well enough to apply to data analysis. I wasn't sure if I could self-teach some of these things later on. It does sound as if the broad education from the math department might make it harder though. I suppose it depends on how well they tailor it to education and psychology students. The head of the program is actually from education, but it's technically under the stats department.

 

if this is a formal Statistics degree, the calculus that you will learn won't actually be needed for data analysis itself. you'll usually need to understand calculus to understand something about the properties of estimators or how certain probability density functions become something else. when it comes to the application part of things, the calculus is usually done behind the scenes in the computer. but you'll now be able to say "aha! i know where these numbers come from!" 

[TXInstrument11]

if this is a formal Statistics degree, the calculus that you will learn won't actually be needed for data analysis itself. you'll usually need to understand calculus to understand something about the properties of estimators or how certain probability density functions become something else. when it comes to the application part of things, the calculus is usually done behind the scenes in the computer. but you'll now be able to say "aha! i know where these numbers come from!" 

 

 

I guess this can be either a hindrance or an advantage depending on how you work with it.

 

I come from a Math background and jumping into a social science/education/psych background did help me bring in some skills that I know your average student in these programs does not have. Most of the methodology courses that you take in Psych or Ed departments are much more focused on the application of statistics or how to use the methods. Very little attention is devoted to the actual theory or why they work. Whereas this probably serves the needs of your average graduate student, it does leave them at the disadvantage of not knowing how to proceed if they’re dealing with an unusual dataset with complicated dependencies (spoiler alert: those are the types of datasets where the most interesting results are found). A degree in Statistics might not prepare you to become a skilled data analyst in the social sciences but it will give you the necessary background to jump in with new ideas. But then again if you plan on working on research areas where there are standard methodologies in place, then there really isn’t much value in getting an MSc in Statistics aside from fostering your own personal knowledge.

 

 

Will it make you a better researcher? Well, it depends on what you’re researching, right? It made me a good researcher because my research is very technical in nature. But if say you were interested in doing research in... oh I don’t know, standardized tests or scale development then it would benefit you much more to get some sort of degree or specialization in Psychometrics and not formal Statistics (very little to no Psychometrics are covered in a standard Statistics department). If you see yourself working in areas like neuropsychology or neurobiology, maybe a degree in Biostatistics (with emphasis in fMRI imaging which is a VERY hot topic) would serve you better. It seems to me that if your ultimate goal is to be a researcher more than a methodologist, you’d need to tailor your degree to cover the methodologies that are used in your substantive area of content. Or you can always become just a methodologist… I mean… we’re kinda short on those right now

 

Will it make you more employable in the job market? Oh, it most definitely will! But here’s the catch… will it make you more employable in jobs you are interested in? Like, big pharma companies pay well for DNA sequencing analysts… but last time I worked on something like that I wanted to gouge my eyes out of sheer boredom.

 

So all in all I think this is not a bad idea per se as long as you have a good game plan. I mean, you’re still committing a year of your life to something that you’ve already admitted to you’re not super passionate about.

Thanks for the perspective and feedback! I was hoping you or another quant person might chime in on this.

 

I have heard of poor analysis (and sometimes outright fraud) in data going unnoticed for years in psych until someone with a more technical background takes a look. Case in point - http://www.theguardian.com/science/2014/jan/19/mathematics-of-happiness-debunked-nick-brown

 

I hear things like this and I think, "Well, damn, what if I was duped by that? If somebody more qualified than me said the math on an unfamiliar technique was good, would I ask questions? Would I even know what to ask?" 

 

It's scenarios like this that make me worried about taking the average stats courses offered in psych and just sticking with that. The idea even that I could do everything ethically right, analyze the data to the best of my ability, and still royally screw up scares me. Sure, mistakes happen, but this was a bad one that was far from obvious ---- only because other psychologists did not have the technical skill to spot the BS. 

 

 

Ultimately though, I think  you all are right - I need to have a long conversation with my adviser and maybe some other people in my field on this before I leap.

Edited August 5, 2015 by TXInstrument11

[spunky]

Thanks for the perspective and feedback! I was hoping you or another quant person might chime in on this.

 

I have heard of poor analysis (and sometimes outright fraud) in data going unnoticed for years in psych until someone with a more technical background takes a look. Case in point - http://www.theguardian.com/science/2014/jan/19/mathematics-of-happiness-debunked-nick-brown

 

I hear things like this and I think, "Well, damn, what if I was duped by that? If somebody more qualified than me said the math on an unfamiliar technique was good, would I ask questions? Would I even know what to ask?" 

 

It's scenarios like this that make me worried about taking the average stats courses offered in psych and just sticking with that. The idea even that I could do everything ethically right, analyze the data to the best of my ability, and still royally screw up scares me. Sure, mistakes happen, but this was a bad one that was far from obvious ---- only because other psychologists did not have the technical skill to spot the BS. 

 

 

Ultimately though, I think  you all are right - I need to have a long conversation with my adviser and maybe some other people in my field on this before I leap.

 

 

 

i'm glad i could help

 

well, i honestly always try to believe that people don't make mistakes like this on purpose  **cough Diederik Stapel cough Michael LaCour** but sometimes it's just the lack of awareness about methodological issues combined with the lack of interest that a lot of people have when it comes to data analysis. Denny Borsboom has an awesome article with the even awesomer title The Attack of the Psychometricians (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2779444/)where he comments on the fact that, more often than not, methodologists within the social sciences and applied social scientists exist in somewhat parallel worlds and we don't really talk to each other very much. so our developments in data analysis do not inform your practice and your data analysis concerns do not always manage to make it to our computers/desks because there's just too many of them. that leaves a good chunk of what i would consider to be good, honest and bright scientists doing some pretty horrible data analysis jobs because they don't even know enough to question whether what they're doing is right. and then this goes on to be reviewed by people who are not intimately familiar with quantitative methodology/statistics and the cycle repeats itself over and over again until a few years down the line people start wondering why so many results fail to replicate

 

but, ultimately, you're the only one who can judge whether this leap is worthwhile for your or not. like if you were to tell me something like "i'm unsure about my job prospects once i graduate, i need a back-up plan that will get me a job, any job that pays more than waiting tables" then sure, absolutely, soldier on through your MSc and you'll be thankful that you have that as a back-up plan. but if you've already established a career route for you that maybe involves more research in psychology, maybe a tenure-track position, etc. then i would devote my time to get more pubs out and getting my name known as opposed to getting an MSc.

 

the good thing about the internet today is that you can maybe enroll in a MOOC or watch Khan Academy or something like that and get a feel of how things are. if you see yourself getting interested then maybe you can progress more into taking Statistics but if you don't like it then at least you're not in a program wasting years of your life for something that's eating away your soul. 

[Wrellie]

Which psychology discipline are you in at UIUC this fall? My advice would be to actually go to classes first before you even start thinking about this. You have people like Fritz Drasgow in I/O who can teach you more about stats in psychology than you could ever hope to learn in your entire lifetime. I met one of the incoming I/O first years at UIUC on the recruiting trail last season who is extremely impressive in her stats knowledge, and it was obvious just talking with her briefly that she knew her *!&*. These are the kind of people in your program. My advice is focus right now on what you went to UIUC for in the first place, and if you are not I/O and want stats courses, take them with the I/O professors or whoever people like Fritz Drasgow thinks is best for you.

[ginagirl]

the good thing about the internet today is that you can maybe enroll in a MOOC or watch Khan Academy or something like that and get a feel of how things are. if you see yourself getting interested then maybe you can progress more into taking Statistics but if you don't like it then at least you're not in a program wasting years of your life for something that's eating away your soul. 

I second this- definitely poke around on Coursera and see if any of the stats/data science courses offered by Hopkins (and others) interest you and fit what you're hoping to learn. If you want to take one to try it out, that might be a good idea before jumping into a masters program full-on. These topics won't be entirely the same as what would be covered in a Stats MS since the a lot of the Coursera courses now are "data science"-heavy, but would still be a good starting point. I'd stay the statistical inference course, R programming, and regression models would be good places to start.

https://www.coursera.org/course/statinference

https://www.coursera.org/course/rprog

https://www.coursera.org/course/regmods

[1HeavyDiaper]

I am surprised at some of the skepticism about getting a MS in Stats here from others, but I would recommend it 110% even if you don't plan on being a researcher. I received my Masters in Quantitative Methods, and worked alongside students in a traditional Stats MS degree program and we took roughly the same classes (at a top tier institution in the Southwest). Now, I did not have to complete 2 years of prerequisites like you are indicating you might have to, but regardless I would say it is still probably worth the investment of time (and it sounds like your program is funded anyways). I can tell you that having a generic MS in Psych will do absolutely nothing for you that the PhD in psych can't already do. Having the MS in Stats will open many doors occupationally and will give you a leg up on future internship/fellowship/job applicants because it is one more tool in your toolbox. I received the same advice a few years ago, coincidentally when I was already in the stats program, and I can tell you that it is a huge benefit to have stats experience. The majority of others in the field you're going into will not understand stats past ANOVAs and regressions, and you would be well-served by gaining this experience.

Edited August 11, 2015 by 1HeavyDiaper

[TXInstrument11]

 

Wrellie

Hm. Good tip! Thank you. I will have to look into him. My first year stats courses don't have him, but perhaps I will have classes with him later. I am not I/O though.

 

 

Well, good to hear one person who agrees with my premise at least, though it will be an uphill battle if I decide to do this. I am also intrigued by what stats could do for me professionally. I am a person who likes to have multiple backup plans and I really like the idea of having certifications that could allow me to move out of academics if I had to. For research as well, you can never understand enough about stats.  

Edited August 19, 2015 by TXInstrument11
the new quote system sucks

[Heavensmaiden]

On 8/5/2015, 5:37:06, ExponentialDecay said:

I've had this conversation in the past, and most professors said that, whilst taking stats in the math department may give you broader preparation, you end up learning a lot of stuff that isn't used in your discipline and, conversely, miss out on techniques that are not standard-fare in statistics taken broadly, but which are necessary for your field. If you're interested in statistics per se, taking the MS through the math department may be a good idea, but otherwise you may be confusing yourself and giving yourself extra work without much payback. Also, if your psych department is considerably better than your math department at your school, the stats given in the psych department may actually be stronger. As regards getting a professional edge, I don't know; maybe it would be a good idea to ask people working in your desired field for advice rather than a motley crew of current and potential students.

This, i had the same question. The above is what I heard.