In his recounting of our Oglethorpe extravaganza last Wednesday, Joe Knippenberg asked whether an online edition of my comments might be in the offing. It'd be ponderous to recount my comments in full, but here is the upshot.
My comments were based on a Harry Jaffa quote that I've used here before: "This election will be determined by events which have yet to occur." That's another way of saying the campaigns matter. Events occur that we cannot predict and, of course, candidates make claims and counter claims against each other. How they do that is important in determining an election's outcome. Words and deeds shape the public's evaluation of candidates.
You might think that is an obvious observation, but it isn't to many political scientists. As I recounted in Atlanta, my only previous trip to Atlanta was for a political science conference shortly after the 2000 
election. As some may know, in the spring of presidential election years various political scientists come out with statistical models purporting to predict the coming election. These models are usually very accurate. In 2000 most of those models predicted a big win for Al Gore. Of course Gore only narrowly won the popular vote (by 0.5%) and lost the electoral vote. What went wrong? Well, some of those political scientists appeared on a panel in Atlanta to answer that question and their answer was clear: Al Gore ran the wrong campaign. Not that their models were wrong, mind you, but that Al Gore was wrong. Now there is some truth to that. These models are heavily dependent on economic data, and they suggest that if Gore had essentially run for a third Clinton term he would have been successful. But he chose to distance himself from Clinton. Perhaps that was a dumb move, although I'd point out that Gore's resurgence in the polls that year occurred after the Democratic convention when Gore unveiled his populist "people versus the powerful" (who apparently aren't people) message. But this just goes to show that campaigns matter.
This attempt to reduce politics to a science is rampant in my field (thus the name "political science"). In my new copy of the American Journal of Political Science, perhaps the second most prestigious journal in the field, there is an article regarding the influence on economics on voting. This article, statistically rigorous, ends with these words: "[This study] vindicates the enormous amount of research that has come up with positive findings on economic voting. These cumulative efforts at scientific achievement were not in vain." That's some pretty bold prose! This study must represent a giant achievement for political science. So let's look at their findings. Well, when one views the three statistical models it turns out that these three models explain 29%, 25% and 27% respectively of the variance (this is the "psuedo R squared" statistic). This is the major scientific finding, one that leaves 70% of the variance unexplained? I suspect Mr. Einstein could not get away with this. Even if a political science model can explain 90% of the variance (and that is nigh on a miracle in political science) that 10% may be all the difference in elections.
I don't mean to dump on behavioral political science (ok, I do a little). I happen to think that most people vote their interests as defined by their socio-economic status, and that status can be quantified and analyzed. But how people perceive the candidates in relation to those interests is formed by actions of the candidates. So we need to look at behavior such as Vice-Presidential selection, strength during debates, and response to news events (e.g., if there is another major terrorist attack in the world). Or, if Hillary Clinton is the Democratic nominee, how does she deal with Bill? How does John McCain deal with the age issue? How does candidate Obama deal with his relative inexperience versus McCain?
Let's look at that issue. McCain must be skillful in defining his age to mean experience and trustworthiness, while also using his "maverick" status and his personal biography to at least blunt the "change" message of Obama. Obama must try to paint McCain as part of the problem and turn his own paucity of experience (especially compared to McCain) into a virtue. This rather ethereal subject, one that defies hard quantification, is central to a McCain-Obama race.
Polls today suggest a rather close race with either Democrat versus McCain, and the electoral college race seems just as close. But even if it wasn't close, that would mean little. Most races have and ebb and flow. As John Fund points out, Michael Dukakis once held a 17 point lead over George H.W. Bush and after Labor Day 2000 Al Gore held a substantial lead over George W. Bush. Let's sit back and let the candidates be candidates and enjoy the show.
Update: Thanks for the link, Joe.
Recent Comments