Andrew Gelman, a political scientist and statistician who writes at The Monkey Cage, has been in a bit of a spat with TNR's Jon Chait about the electibility of various Republicans, whether the actions of independent voters are "rational," and some related issues (here's the latest entry). It's a repeat of a common argument that comes with variations: The political scientist says that presidential elections are quite predictable, based mostly on the economy, so all the time we spend talking about which candidate is better looking and whose ads are better and all the other stuff is just sound and fury, signifying pretty much nothing; the journalist responds that yes, the economy makes most of the difference, but sometimes what matters at the margins can make the difference (see the 2000 election, for instance). They're both basically right, which makes it hard to resolve the argument.
But is predicting the election's outcome all that matters? This, to be honest, was always my biggest problem with political science as it's practiced today: not that it's wrong but that it's so often boring. If you're reading this blog, chances are that you find politics fascinating. It's fascinating because it affects all our lives, because it's always changing, because the swirl of competing interests and values make it so difficult to predict (not talking about the outcome of elections here), because of the inherent conflicts, because of the characters involved, because it touches on so many things about who we are as individuals and as a nation.
Faced with all that, the political scientist often comes off like someone telling you, "Don't bother seeing that production of Macbeth; it's just a whole lot of meaningless talking and wailing and stabbing. All that matters is that he dies in the end." Well, yes, the fact that Macbeth dies in the end is important, but all the other stuff that happens along the way is compelling, too.
Speculating about what difference the talking and wailing and stabbing will have on the outcome of a race, particularly the presidential race, is kind of unavoidable. Political scientists can help those of us who comment on the day-to-day ups and downs to keep from getting ahead ourselves when doing that speculation, and show us where we're wrong when we get too emphatic about campaign cause and electoral effect. But the show still matters.