Political Science v Punditry
One funny thing about studying public policy and American politics is that people assume that you know a lot about public policy and American politics. And I do, but not at the level of detail on which most everyday political discussions take place. Academics are interested in theory, whereas most people want an answer to specific questions like "Why did that lady lose that election? What's gonna happen when the Bush tax cuts expire? Who's gonna win in such and such an election?" Academics don't ask questions at that level of particularity. We ask questions like "What's the impact of interparty polarization on the strength of Congressional leadership?" Or, "What are differences between the dynamics of policy reform at the federal and state level?" Or "What are the correlates of successful policy implementation when network governance approaches are employed?"
You get the idea. It's the difference between theorizing about the rules of football and predicting the outcome of a particular game. If I was an academic football expert, I could tell you what types of things tend to lead to victory, and I could probably tell you what types of combinations of team talent tend to lead to what kind of games under what kinds of weather conditions and so on. But that doesn't mean I'm going to be able to say much about any particular game. Now, if I had time, I could investigate a particular game and find out the relevant details and come up with some kind of reasonable prediction. But, when the average fan comes up and asks my opinion about a particular game, well, he's gonna get an opinion, but it's probably only mildly better than any other schmoe's idea.
One thing that academics political scientists do well that your average pundit does not is to explain particular events as predictable results of long term, stable patterns. So, right now, people are freaking out about Obama losing popularity and are accusing him of terrible mismanagement. But it's a fact that this is what happens to Presidents, regardless of party! People love them at first, and then they realize that shit isn't changing to their liking, and maybe they're still unemployed, and they blame the President. What this really means is that people have unreasonable expectations for what the President can do. The President's power is very circumscribed. (I think I've used this quote before, but I'll use it again. Truman, about the incoming President Eisenhower: "He'll sit here and he'll say 'Do this! Do that!' and nothing will happen. Poor Ike...it won't be like the army.") So, people expect him to wave his magic Presidential wand and make everything better, because they don't understand how government works, and then they blame him when shit goes wrong (but it's ok...they also give him credit when things go well, so maybe the unfairness evens out over time).
I'm just using this President thing as one example of how academics tend to think in large scale patterns, whereas the media likes to think in terms of discreet events. So, I guess my point is that, even though I actually have been learning stuff, I still don't really know anything.
posted by Joe @ 8:46 AM
4 Comments:
Who reconciles specific events with the established theories you work with?
Well, we don't really have established theories. We have lots of competing theories, which people try to prove or disprove using data (i.e., specific events). We don't get down to the level of, like, one specific election (except sometimes for the Presidency, which is sort of the red headed step child of academic political science). We need a large number of observations to establish general patterns.
So, for example, IN GENERAL, incumbents win. IN GENERAL, people vote for the party their parents supported. Etc, etc. That doesn't mean that all cases follow the pattern, but you need a large data set to disentangle the pattern from the random "noise."
The short answer to the question is that reconciling theory with evidence is pretty much the definition of what scientists do.
Well by "established theories" I meant those that have been widely accepted within your community of practice. Your example involving midterm election results I think would apply.
I guess what I'm wondering is how analysis of such long term trends requiring such high sample data sets can effectively provide decision support in such a insanely dynamic area of study. I mean Christ how is voter behavior in the age of Truman applicable to the age of twitter?
Let's see...if I'm interpreting your question right, you're asking how do researchers know if their findings are sensitive to time? And one might add space?
This is definitely a big issue. It's one everyone is aware of but, at least in my discipline, isn't very good at dealing with. But basically, if some researcher gets suspicious about whether or not an old finding still holds, he or she can do another study to check. If they get different results, that's a publishable paper. The buzzword for this is replication. The problem is that replication projects are sort of seen as second-class research and don't get funded as frequently.
I should say that there are also techniques of research design and statistical analysis to try to reduce those kinds of things, if a more universal conclusion is what you're after. That's called external validity.
As to the "decision support" thing, political scientists don't give a shit about that. The classic disciplines don't care about being useful. My SPEA side, however, does. There's a big divide within academia between applied research and theoretical research.
So, an applied approach to the election would probably do a bunch of polling in an area and maybe use some regression models that incorporate demographics and party-id and bla bla bla to try to predict an outcome. So, it helps by putting bounds around uncertainty....anybody promising more than that is selling snake oil.
Post a Comment
<< Home