Tuesday, February 07, 2006


per·e·stroi·ka (pĕr'ĭ-stroi'ka) n.

The restructuring of the Soviet economy and bureaucracy that began in the mid 1980s.
[Russian perestroĭka : pere-, around, again (from Old Russian) + stroĭka, construction (from stroit’, to build, from Old Russian stroiti, from strojĭ, order).]

For political science geeks:

In October 2000 Mr. Perestroika issued an e-mail manifesto damning the American Political Science Association (APSA) and its flagship journal, the American Political Science Review (APSR), for excluding solid qualitative research in favor of abstruse mathematical modeling by “poor game theorists” and “failed economists.”

Perhaps the single most important point on perestroikans’ agenda is, like Mikhail Gorbachev’s reform policies from which the movement takes its name, to upset the status quo. In this case the status quo is a discipline increasingly dominated by scholars who emphasize the “science” part of their field’s name and rely on formal models such as rational-choice theory and “large-N” statistical analyses (which use large sample sizes) to identify universal or quantifiable explanations for political behavior. These quantitative types, say perestroikans, exert hegemonic tendencies, ignoring or dismissing research that they don’t consider “scientific”—for example, interpretive research by area specialists like the Rudolphs, based on fieldwork in a specific country or among a specific people, or theoretical work such as that done by Chicago international-relations realist John Mearsheimer, which relies on a few carefully chosen case studies and historical context to prove a point.

Where do I lie in this debate? Kind of straddling the fence, but verging on the point of being a complete sell out.

I hate math a much as the next law student. In fact, law school is full of kids who want job and financial security despite their unmarketable skills of term paper writing and critical thinking, but are equally afraid of math. If we could have hacked science or math, we would have been doctors or MBAs. But we can't. Our only talents are reading (and if someone says "I like to argue," as in "I am belligerent," so help me I'll bitch-slap them), cogent writing and rhetoric. Totally, totally useless. This doesnt' mean I didn't do well in statistics. I did, I just remember that it didn't come easy. And it's taken me a few years to realize that nothing does, and that doesn't negate it's utility.

All of this is rationalization for: yes, I am going to sell out and start doing math, even though I don't like it or particularly want to.

I think one of the reasons I didn't do the political science Ph.D. program four years ago was that I didn't like the direction of my department and departments around the country. I didn't like doing modeling, regressions, quantitative analysis nearly as much as I enjoyed case studies and political theory. I was all about the theory. I continued being all about the theory in law school too, concentrating in Critical Race Theory.* But in slowly getting back into political science in the last few months, skimming a few articles, looking up programs, reading up on a few academic's blogs (surprisingly few), I have come around--a little. Maybe it's because law review articles are so incredibly descriptive and proscriptive, but without anything more than textual analysis to back it up. Now, I'm all about critical theory and deconstruction and textual analysis. Those are my skills. But once in a while, I'll come across a law review article deploying statistics, and I'll find it useful, interesting, and refreshing. Sometimes they're used for evil, and I think "someone needs to counter this, point by point, fake number by fake number." I think it's that last point that has made me come around to the thought of using statistical models. That, and marketability.

I know I'll get a job 10 times faster if I do quantitative analysis in addition to qualitative analysis. As much as I love tearing apart opinions word by word, it would be pretty useful (to my career, and to the discourse) were I to do a statistical model on how federalism opinions in the past 25 years have come down on various issues of criminal law, anti-discrimination law, hate crimes laws, and environmental regulations. Would most results be on the side of the states or the federal government? What insight can be gleaned from the pattern--is it largely pro-states' rights for laws that tried to increase minority interests or regulate business interests, and largely pro-federal government for laws that regulated "moral" interests like death with dignity acts or federal drug laws like medical marijuana? See, interesting. I could also do a behavioral analysis, i.e. trying to determine whether the cases exemplify a kind of rational choice action for each justice, or whether it was some outlier personal motivation behind the vote.

I was scratching my head just a month ago wondering how the hell I would ever survive in a poli sci program if I can barely add. Then I thought, "calculators." Most statisticians use computers, I can too. But beyond this, I was wondering whether or not I even wanted to start this kind of research. Like I said, I'm a theory head. I always assumed that I would just focus on legal history, pore over texts and cases, and write very nuanced arguments that are descriptive and prescriptive. Now I think I'll try to do both--I definitely won't be an area studies person, or do some kind of political psychology case study thing (I mean, my material are cases, not individual people)--but I can imagine deploying both quantitative and qualitative analysis to attack some aggregate group like the Rehnquist Court's federalism jurisprudence, or to examine how the jurisprudence has created a disparate impact on minority interests. And then I can bolster the statistics with a qualitative analysis comparing U.S. v. Morrison (which said that providing federal remedies for gender-motivated crimes was outside of Congress' powers, despite evidence of the economic impact of such violence on interstate commerce--same argument for desegregation) to Gonzales v. Raich (which said that Congress could regulate the private, non-commercial possession and consumption of medical marijuana, just because its existence would affect the market for drugs) and say, this doesn't make sense. Two decisions with two different results, but they're both essentially non-economic activity--why the disparity? Rational choice theory can't explain the disparate results, neither can any theory of jurisprudence. Thus, the Rehnquist Court jurisprudence is not based on any coherent, discernible legal principle--it's results based, finding for the federal government only when it's convenient and impacts only a minority, instead of larger state or business interest--and here are the numbers to prove this. Numbers + Words = Crazy Delicious.

That said, I don't want to do numbers for numbers' sake. This is not a case of departmental inferiority complex. Economics has to stop feeling inferior to the hard sciences and trying to up it's "rigorousness" street cred. Political Science has to stop feeling inferior to Economics (which is such a bad prognosticator anyway, and becoming increasingly descriptive rather than proscriptive, which is both good and bad) and trying to import everything from Econ (from stats to game theory to rational choice) and remember that we used to do theory too. We used to do philosophy. Hell, Econ needs to remember this. At least in each department, scholars should feel free to go back and forth in methodology, or combine, and theorists should feel welcome--even if they can't do math. Statisticians shouldn't think of numbers in a vacuum. When we look at the achievement gap between boys and girls, blacks and whites, we should ask why, and say what we can do about it. Numbers in absence of context and without direction are awfully empty. But so are statistical interpretations without merit (affirmative action mismatches blacks to institutions that are too hard for them, which is why they fail--not the institutional bias of school and subject-matter or history of inadequate education due to de facto segregation), and proscriptive statements that are just plain idiotic (boys perform worse than girls because girls are better at reading and writing, thus we ought to change the curriculum so that boys can do book reports on comics and cater to them and call it "gender-specific pedagogy") are also pretty useless. So use numbers wisely, make arguments and interpretations carefully, and make sure your policy proscriptions are sound, or at least, not evil.

Remember: Numbers + Words = Crazy Delicious


Links to this post:

Create a Link

<< Home