The statistics...oh my god so many f(x)s and y1s

[peternewman89]

Just starting my first year in a PhD program and I am drowning in all of these meaningless numbers.  I'm a political theorist so I have a different approach to the discipline altogether and something of a predilection against the quantitative approach, but I tried to come in with an open mind.

 

The stats I've been getting, though have just been so mind numbing and I can't for the life of me see how any of this has anything to do, not with political theory, but with any sort of political science in general.  Its all just a mess of variables and functions and subscript and doesn't make a lick of sense. I sort of get it enough to get by but I just am left seething at the fact that this is what I'm taking my time to do.

 

It also doesn't help that we're getting empirical methodology shoved down our throats and nobody can stop talking about how we aspire to be scientists, which is not at all what I'm looking for. I'll never be considered for an NSF grant and I don't go about research in the whole question-hypothesis-method-data gathering-analysis way that most of the department (the non-theorists) say is the only way to think about anything. I want to be a thinker, not a scientist.

 

Its just an altogether frustrating enterprises thus far, though not to the point of making me think of give up; the teaching that I've done has been too rewarding for that, and I haven't even had the chance to do any theory coursework or expire my own original ideas.  Just looking for ways to persevere and rationalize all of this.

[TakeruK]

Although I am a scientist, I hope you will hear me out. I do sympathise with your frustration and I think it's common in most PhD programs (in all fields) where the student will end up learning things that don't seem to be very useful to what their interests were.

 

In the past year, there was a lot of qualitative things I had to learn when I am used to thinking about things quantitatively. For example, in most of my learning, the major ideas were a result of other laws and facts -- the reason planets orbit the Sun is a consequence of the law of gravity. But in my first year, in a geological science course, one key component was to just straight up memorize a whole bunch of terms and concepts. We had to memorise the geological timescale (e.g. "paleozoic", "mesozoic" etc.) and mineral types. The course was meant for non-geologists like me who might communicate with and work with others who do geology work. I found it very difficult to adjust my learning style to think of relationships between ideas in this manner, without equations or numbers.

 

However, the prof explained there was a good reason for all this. Although most of the students in the course were planetary scientists, and may never deal with these ideas again, there's also a good chance that we might end up working with geologists to study e.g. another planet. Then, it would be helpful to understand the basic concepts and language that is used by geologists (when they describe the Earth). The prof also explained that we were in a Earth & planetary science department, so we should be expected to know enough fundamental ideas to communicate with another person in our department that does completely different research. 

 

So while I don't know your situation exactly, I would encourage you to make the most of your opportunity to learn a different way of approaching problems in political science. You say that stats doesn't have anything to do with political science in general, but if that is the case, then why is it part of your program? Think of it as a crash course in learning just enough stats to communicate with a quantitative political science colleague (or understand a talk presented by one of them). Your own research interests may never stray so far as to actually collaborating with the empirical methods, but that does not mean you don't need to know them or that it is a good idea to stay closed off in your own subfield.

 

You also say that you have just started your first year. Will you get a chance to take the courses more directly related to political theory later on? There were some quarters in my past years of graduate school where the courses were just not useful to my research at all. But then at other times, I was super happy that my courses were complementing each other as well as my work. That's just the way things are and I think it's very important to have a broad education (at the graduate level I would say "broad" would mean "outside of your specific subfield/research area"). I am definitely glad the majority of my coursework is over though, but I am also happy that I have finished it. Grad school is a long time and you might have heard "it's a marathon, not a sprint". You will get to explore your own original idea in due time -- it makes sense to me to get the fundamental coursework out of the way first though.

 

Finally, I don't understand your statement that you're "a thinker, not a scientist". I'm not sure what you think of scientists, but I definitely spend the majority of my time as a scientists doing "thinking". I think putting up divisions between you and your colleagues (i.e. it sounds like the majority of your department) isn't a great idea. I'm a big proponent of having a broad education and multidisciplinary work. It makes sense to me that as a PhD student in an Earth AND planetary science department, I should know things about the Earth AND the planets even though I mostly work on data from stars and asteroids ("minor planets"). For you, I think since you are a political science PhD student, it makes sense for you to be expected to have working knowledge of all the methods and ideas used by all types of political scientists, including statistics. I think academia is a better place when academics from all fields recognize value in each other's work and strive to learn from one another instead of being stubborn in their own ways and solely sticking to their own interests.

[spunky]

Just starting my first year in a PhD program and I am drowning in all of these meaningless numbers.  I'm a political theorist so I have a different approach to the discipline altogether and something of a predilection against the quantitative approach, but I tried to come in with an open mind.

 

 

It also doesn't help that we're getting empirical methodology shoved down our throats and nobody can stop talking about how we aspire to be scientists, which is not at all what I'm looking for. I'll never be considered for an NSF grant and I don't go about research in the whole question-hypothesis-method-data gathering-analysis way that most of the department (the non-theorists) say is the only way to think about anything. I want to be a thinker, not a scientist.

 

 

would switching to a more qualitative-methods-based program work, maybe? or not switching programs but just switching your approach to research? i can see that both political science and psychology (my field) have very quickly become heavily quantitative programs and tend to look down upon theory that is not backed by solid empirical research. 

 

being an absolute pragmatist, i'd advise you to, honestly, stick with it. if you resist the "question-hypothesis-method-data gathering-analysis" process, you're also closing the doors to publishing in many reputable journals. which also closes the doors to advancing your career and diminishes the chances that you'll do any "rewarding teaching" once you're done.with your PhD. remember the motto: "publish or die" (where "publish" implies question-hypothesis-method-data gathering-analysis). and if psychology works anything like political science (and i have a feeling it does, in this case), people tend to resist publishing theoretical articles by newly-minted scholars. not sure why they do that, but i'm sure they do. 

[Guest |||]

"I want to be a thinker, not a scientist."

 

At least based on this,  it might be you don't how many doors and options knowing statistics opens for you, or quite exactly what a scientist does.

[catchermiscount]

By way of predilections: I am no technician, but my work is probably more technical than average.  My program is one of the more technical ones.  So, this should probably be read with a large handful of salt.  I am also no fan of positivism, which seems to be the issue just as much as technique is.

 

It might not be obvious during the very first semester---where almost all of the focus is on gaining competence with the basic building blocks of technical work---but eventually one gets to a point that the numbers are no longer "meaningless."  This is especially evident when the numbers are the results of a very theoretically-motivated statistical model.  Think about ideal point estimation.  For those that don't know about this just yet, ideal point estimation refers to a class of methods by which one inputs congressional votes (so you have a big matrix of yeas and nays with as many rows as there are congresspeople and as many columns as there are measures voted upon), uses that data in conjunction with some spatial theory of voting, and then outputs estimates of where each congressperson falls in unidimensional ideology space.*  Those estimates mean something, though that meaning is conditioned on how we imbue the model with theory.  Ideal point estimation papers can be quite technical.  The seminal paper by Poole and Rosenthal in the 1980s made such intense computing requirements that, at the time, it wasn't replicable.  Today's bleeding edge stuff is Bayesian and uses markov chain Monte Carlo methods.  Just looking at the papers can be very intimidating, but the process is still simple:  get votes, use theory and stats together, and then interpret results.

 

So where is the science?  Does the fact that something is technical and difficult mean that it's scientific?  Maybe?   Who knows?  Who cares? Despite all the Greek letters in their paper, Poole and Rosenthal didn't really test any hypotheses in their paper:  they just measured political ideology.  Most people would probably call what they did "good science."  It was certainly deductive:  if you buy this spatial model, and if you buy these data, then you should buy the results.  Being deductive is probably a criterion for being "sciencey" for most folks, but it definitely isn't unique to data analysis.  Rawls was a pretty deductive guy.

 

That brings us to another point:  if you don't like math, or functions, or their meaningless, or whatever, then this generalizes to your views of formal theory.  But some of the very best theorists ever did things that were very technical but also very normative:  Arrow's theorem, the work of Amartya Sen and John Harsanyi, the welfare theorems**...these things are deductive and they're technical but they're not in the vein that the OP mentioned.  But when you learn theory, you again start from annoying building blocks that seem not to mean anything:  truth tables, and set theory, and real analysis, and so on.  These are just the technical requirements for being able to engage in higher-level thinking later on.  While your average political philosophy scholar likely has little to say to your average applied empirical political scientist, they might have quite a bit to say to a formal theorist.  Heck, a few years ago we took on a new student that already had a PhD in philosophy but that wanted to do philosophically-motivated political economy.  He's a remarkable guy, but it's good evidence about how these things can work hand in glove.  

 

Note also that the best users of theory also think hard about what the theory means:  Akerlof's diatribes about the real meaning of the Arrow-Debreu general equilibrium "utopia" are really interesting reading.  The same for many of Sen's works.  And again, this isn't just specific to mathematical theory:  Schelling was a theorist that used no theory.  

 

Speaking in a language of models means that the numbers are imbued with meaning from the start.  But most technical classes are about competence, not modeling.  With enough perseverance, you get to use all these boring tools in fun ways, and that's really rewarding.  Whether it adds up to good science from your average positivist is anybody's guess, but who cares what they think, anyway.

 

----------------------

*:  Ideal point estimation is just one kind of data reduction technique, and it's used in political contexts other than the Congress (e.g. the courts).  Most of the statistical machinery comes from psychometrics, where they wanted to estimate, say, intelligence using test answers.

**:  Science envy and economics envy are everywhere, and I just contributed to the problem by focusing on economic theorists.  I could have thrown Riker or Shepsle or McKelvey or Ferejohn or some other really smart theorist from a political tradition in there.  The economic examples are more obviously normative.

[TheGnome]

OK that was a great reply. Thanks for taking the time and doing that coachrjc

[RWBG]

By way of predilections: I am no technician, but my work is probably more technical than average.  My program is one of the more technical ones.  So, this should probably be read with a large handful of salt.  I am also no fan of positivism, which seems to be the issue just as much as technique is...

Are you not teaching this semester or something?

[catchermiscount]

Are you not teaching this semester or something?

 

Aw, come on, man.  I ALREADY MADE THE ANSWER KEY FOR THE MLE HOMEWORK.  Talk about meaningless numbers:  it's the introduction to monte carlo analysis.  That would make some folks' head just about explode.

 

Are they making you teach?

 

PS:  Sorry I didn't respond to hospitality---I ended up having some family rigmarole and got off the grid for a bit.

[GopherGrad]

I will make no attempt to improve on the clinic RJC put on a few posts above. I'll simply say this:

 

I see the benefit in being able to understand statistical and game theoretic work, and to be able to think mathematically about even the most theoretic concepts we will encouter in political science. But I'm in the middle of Math Boot Camp before my first year and I would rather eat a fucking gun than take one more partial derivative.

Edited September 23, 2013 by GopherGrad

[catchermiscount]

I will make no attempt to improve on the clinic RJC put on a few posts above. I'll simply say this:

 

I see the benefit in being able to understand statistical and game theoretic work, and to be able to think mathematically about even the most theoretic concepts we will encouter in political science. But I'm in the middle of Math Boot Camp before my first year and I would rather eat a fucking gun than take one more partial derivative.

 

It doesn't get any better.  I'm knee deep in a Hessian matrix of arbitrary size that has me completely existential.  The fun normative bits supplement, but do not replace, digging in the mines.

[RWBG]

I will make no attempt to improve on the clinic RJC put on a few posts above. I'll simply say this:

 

I see the benefit in being able to understand statistical and game theoretic work, and to be able to think mathematically about even the most theoretic concepts we will encouter in political science. But I'm in the middle of Math Boot Camp before my first year and I would rather eat a fucking gun than take one more partial derivative.

This reminds me: I need to write up a new quiz for students to get more practice with partial derivatives.