19 March 2007

More Political than Science

For those with a background or sufficient exposure to other academic disciplines, some uses in political science seem, sometimes, more embedded in politics, which in fact means political party choices, than responding to an ambitious scientific design. At some discussions among certain political scientists, I cannot refrain from imagining what a discussion among economists would be regarding whether such or such another economic policy decision can play in favour or against, for instance, Toyota or Chrysler. Some economists discuss this kind of things, but they are usually working as professionals for the corresponding company or in a close position in the sector –not as academic economists. It’s not that political scientist should not have political party and policy preferences, just as economist may have cars’ brand preferences –as anybody else. But those preferences have nothing to do with (political) “science”.

Political scientist Rein Taagepera has launched a provocative but sensible call regarding certain research methods in political science. Taagepera is well-known by the readers of this Blog as a political entrepreneur, but his more scientific background is in physics, which he takes as a model for any scientific endeavour. He organized a symposium under the title: “Why Political Science is Not Scientific Enough?”, held last year at the European Consortium for Political Research General Conference, in Budapest, just to be published now in the European Political Science journal.

Taagepera forcefully denounces the limits of what has become a standard in empirical studies in politics: linear additive regression models. He blames in part computerized mathematical programs that are presented as the only available techniques. Taagepera remarks that the profuse correlation coefficients published in political science, once printed, no one uses them again for any purpose. In contrast, physicists –he says-- publish coefficient values only if they think they will become steppingstones used by others and produce accumulative knowledge.

Taagepera makes the point that weak research methods makes political science unimpressive for outsiders, including politicians. He compares political science to politics (or, say, to public administration, public policy, electioneering or diplomacy), as biology is to medicine, or physics to geology. I quote from his introduction: “Of course, there was a time when engineers did not have to pay attention to physics, nor physicians to biology. Science becomes useful to practitioners only when it has reached a somewhat advanced stage of development. The question is: Does political science contribute to politics and society all it can, as its present stage? The answer is ‘no’, if the discipline, addicted to canned statistical programs, refuses to espouse the part of scientific thinking that cannot be abdicated to the computers.”

Other contributions to the symposium come from:
- Stephen Coleman: “Testing Theories with Qualitative and Quantitative Predictions”
- Bernard Grofman: “Toward a Science of Politics”.

My modest contribution looks at mathematical menus in other sciences, just to confirm that we, political scientists, are usually constrained to a poor and boring diet.

In my view, the underlying question of all this discussion is:
Can politics be the subject of a “normal” science?
You are invited to discuss.

See my paper here: CLICK


COMMENTS


Bogdan Enache said...

I'm a political science student at Bucharest University, Romania. In my opinion mathematical and hyper-positivist statistical methods in political science --and social sciences in general, but political science in particular-- are a dead end and a wrong way to go forward. I know that most of the profession will not agree with me --especially not Taagepera-- and that most political scientists are jealously trying to adopt mathematical models and statistical methods in the study of political interactions. I think, for my part, that a more correct and fruitful approach is to go back to a Weberian methodological approach.

A note about economists because I'm also, more recently, an economics student : indeed they don't accuse each other of political favoritism as usual as political scientists do among themselves but for an insider the profession is just as split on policy proposal and different analytical tools as political scientist despite their esoteric mathematical models.



mb - experimentador said...

I´m a bit confused. The way I read Taagepera´s comments doesn´t coincide with how Bogan reads them. Taagepera is saying that Pol Sci´s main issue is that it relies too much on canned procedures and linear additive reg models. He says: "Does political science contribute (...) all it can (...)? The answer is ‘no’, if the discipline (...) refuses to espouse the part of scientific thinking that cannot be abdicated to the computers. Therefore, the focus according to T. should be on the part of science that can't be left to computers. We don't need MORE regressions and flashy estimators that generate uninterpretable results. Rather, we need better research design, careful measurement, a clearer research agenda, an overarching goal of advancing knowledge, etc. If this is what T. says, I couldn't possibly agree more.

Now, whether these goals are better served with qual or quant research is a different question, and a question that generates (what I perceive to be) increasingly boring discussions.

2 Comments:

Anonymous Bogdan Enache said...

I'm a political science student at Bucharest University Romania. In my opinion mathematical and hyper-positivist statistical methods in political science - and social sciences in general, but political science in particular - are a dead end and a wrong way to go forward. I know that most of the profession will not agree with me - especially not Taagepera - and that most political scientists are jealously trying to adopt mathematical models and statical methods in the study of political interactions.

I think, for my part, that a more correct and fruitful approach is to go back to a Weberian methodological approach.

A note about economists because I'm also, more recently, an economics student : indeed they don't accuse each other of political favoritism as usual as political scientists do among themselves but for an insider the profession is just as split on policy proposal and different analytical tools as political scientist despite their esoteric mathematical models.

12:28 PM  
Anonymous mb - experimentador said...

I´m a bit confused. The way I read Taagepera´s comments doesn´t coincide with how Bogan reads them.

Taagepera is saying that Pol Sci´s main issue is that it relies too much on canned procedures and linear additive reg models. He says: "Does political science contribute (...) all it can(...)? The answer is ‘no’, if the discipline(...) refuses to espouse the part of scientific thinking that cannot be abdicated to the computers."

Therefore, the focus according to T. should be on the part of science that can't be left to computers. We don't need MORE regressions and flashy estimators that generate uninterpretable results. Rather, we need better research design, careful measurement, a clearer research agenda, an overarching goal of advancing knowledge, etc. If this is what T. says, I couldn't possibly agree more.

Now, whether these goals are better served with qual or quant research is a different question, and a question that generates (what I perceive to be) increasingly boring discussions.

12:13 AM  

Post a Comment

Links to this post:

Create a Link

<< Home