The Science Of Political Science

Back in January, I got into a debate with some work colleagues over the status of political science as a “true” scientific discipline. Though I think one or two were playing devil’s advocate, the consensus seemed to be that political science was not worthy of the second half of its name. I argued (and do believe), however, that it is indeed a valid science.

To be honest, I can’t remember my friends’ specific arguments (we may or may not have been sitting in a hot tub in the mountains and a few drinks in — thus making the debate fairly comical to begin with). I’ll be sure to let them know I’ve blogged about it so they can correct, clarify, or elaborate on their arguments in the comments. As I remember it, though, their main contention was that because political science is the study of human behavior, and because observations of human behavior cannot reliably be generalized and retested to form a consistent theory or law, the study of politics cannot be classified as a science in the same vein as physics, chemistry, or the other “natural” sciences. It’s simply too flaky.

At Davidson (where Taylor and I went for undergrad and both majored in political science), there is a huge division among the department faculty on this point. Currently, every political science major is required to take “Methods and Statistics in Political Science,” a course that is almost uniformly dreaded. “I’m majoring in a social science precisely because I suck at math”, they complain. “When will I ever use this again?” The science and math wannabes that we are, Taylor and I loved it, and ended up working with our professor one summer on a quantitative study of how congressmen shape their political messaging.

The fear of the course stems mostly from a failure by the faculty to execute any kind of positive PR, because really, it’s not that bad and not that hard. Anyone who is smart enough to get into Davidson can handle it. The first half of the course doesn’t deal with a single number, formula, or regression model, but rather shows how to take a question of politics and attack it using the scientific method. Delineate the specific question, form a hypothesis, decide what the variables are, and propose a way to run an experiment. Not hard, right? The second half is certainly more difficult because it delves into the quantitative realm using computer-aided regressions; but really, it’s nothing impossible and the professors are not only ready, but eager to help them understand and succeed.

In an article called “The Lamentable State of Science Education in Political Science” (sorry, I can’t find a free version online besides on the Davidson course page *cough* first linked reading *cough*), Kim Hill from Texas A&M University asserts that research political scientists

seek scientific explanations for observable, natural phenomena. [...] Ours is inherently a natural science - with no concern for supernatural, paranormal, or otherwise other-worldly matters. Any college graduate today who does not understand the scientific character of the social sciences is as ignorant as one who confuses astronomy and astrology or chemistry and alchemy.

She further notes that it is widely accepted that, as a relatively young discipline, political science is “essentially identical to the mature physical sciences at equivalent stages in their development.” Indeed, “it is the rare major policy decision of the U.S. government, General Motors, the Ford Foundation, or any other comparable entity that is made without notable recourse to social science argument and evidence.” Hill soberly laments that “ours is the only scientific discipline I know that essentially fails to educate its undergraduate students in its primary concerns. [...] Indeed, they are as skeptical about that matter when they graduate as when they matriculate.”

The seeds of this skepticism are fairly obvious and wholly understandable. Studying human behavior and systems is inherently unpredictable. It is seemingly impossible to control precisely for variables, and difficult to measure results or even come up with the right units of measurement. The laboratory of political science is the whole world and its recorded history; experiments can be costly, time consuming, and sometimes one-and-you’re-done opportunities (we can’t recreate the 2004 U.S. presidential election to measure “that one thing we forgot to measure”). And perhaps most interestingly, the subjects of scientific political study are… us. And not only our decisions, but often our preferences, our passions, and our opinions. These things are flimsy, indefinite, and trigger our biases. Studying politics scientifically and objectively requires a temporary suspension of our beliefs.

By no means do I argue that any study of politics is or should be scientific. It is entirely possible to study politics in other ways (and indeed, I’d argue that a considerable portion of political science courses are not really political science courses). Perhaps this is why many colleges and universities call the major “government” or “politics”. There’s nothing wrong with this at all. My argument, rather, is that almost anything can be studied scientifically. Literature. Art. Music. Basketball. History. Cooking. I’m not arguing that political science is a robust scientific discipline that has a great track record of discovering any kind of Truth, or any consistently verifiable Law of Human Behavior. But sciences are not defined by the number or importance of the truths they elicit. Science is not a conclusion. Science is an approach.

Problem. Hypothesis. Variables. Experiments. Results. Conclusion. Boom, you’re done. Your results don’t have to prove your hypothesis true — they rarely do. Your experiment could have been a complete failure — you forgot to control for an important variable; you discovered you were asking the wrong question; you realize that you cannot extrapolate anything from your data; etc. So what? Happens in chemistry labs around the world every single day. You haven’t failed to apply science. You’ve learned something. You plug the holes, change your models, and you try again.

The newest issue of Wired — guest edited by J.J. Abrams and dedicated to the idea of “mystery” — leads with an excellent piece by Brian Greene, a physicist and author of The Elegant Universe and The Fabric of the Cosmos. Greene writes (sorry, it’s not online yet…):

To the vast majority, science is solely about answers — the material that’s sandwiched between the covers of their textbooks. It’s understandable. For the most part, we teach science as if it were a technical trade: Learn these facts about cells. Memorize these equations describing motion. Balance these reactions that underlie oxidation. And then demonstrate competence by passing an exam. With this lopsided focus on the end points of research, the scientific explorations themselves receive the most minimal attention.

But science is the journey. Science is about immersing ourselves in piercing uncertainty while struggling with the deepest of mysteries. It is the ultimate adventure. [...] Einstein captured it best when he wrote, “the years of anxious searching in the dark for a truth that one feels but cannot express.” That’s what science is about”

Who would dare preclude the social sciences from this quest? What nobler or more righteous use of the scientific method than to better understand humans, how we interact with each other politically, how we construct our institutions, how we achieve efficiencies and better serve our societies’ needs… or how we manipulate these systems to exploit others and consolidate absolute power? In these times of partisan polarization, global poverty and hunger, financial uncertainty, and environmental meltdown, we need every model of analysis we can get our hands on.

And now, a word from Randy Newman:

Comic used under a Creative Commons license courtesy of xkcd.com.