Campaigns and elections

Political scientists’ and journalists’ narratives

Jun 7 '10

Chris Beam wrote this amusing article on the ways in which political scientists view political news differently than journalists do. Seeing this article gave me some satisfaction, as it suggests that some of our research in this area has reached the outside world.

In 1993, Gary King and I published an article, ” Why are American Presidential election campaign polls so variable when votes are so predictable?”, in which we argued (with 10 figures and no tables (except for a brief summary of data sources in the appendix)) that short-term swings in public opinion during presidential election campaigns (for example, the predictable post-convention bounce) have little if any impact on the vote. The bit about elections being so predictable was not original to us–we leaned heavily on Steven Rosenstone’s 1984 book on forecasting presidential elections. What was new in our paper was to take that finding seriously and work through its implications for campaigns.

When we wrote the article, Gary and I wanted to make a difference, to elevate public discourse. It was so frustrating to see the news media focus on the horse race, especially given that there was no evidence that these horse-race stories made any difference. We thought our article might change things, because instead of the usual strategy–criticizing the media for distorting politics with endless stories on the horse race–we were taking the opposite tack, essentially mocking the media for running story after story about campaign gaffes etc. that had no effect. If it’s really true (as we found from our analysis) that what’s most important are the so-called fundamentals (political ideology, party identification, and the economy), then the way the media could have the most influence would be to report on the fundamentals–report what’s happening in the economy and report the candidates’ positions on major issues–rather than the trivialities.

We really hoped that, if our goal was to change how campaigns were reported, we’d do better to portray the standard media practices as ineffectual, rather than as harmful. If you want someone to change, it’s better to describe him as a loser than as a bad boy.

I was frustrated for many years at how little difference our argument seemed to have made. But, if Beam’s article is any evidence, maybe our message really has been getting through!

P.S. More here. We’ve become conventional wisdom!

P.P.S. I’m certainly not claiming that the political science research that Beam alludes to is all (or even mostly) coming from us. Rather, the work that Gary and I did is part of a large stream of research in the past few decades that many many people have contributed to, and continue to contribute to. I’m most familiar with our own part in all of this, which is why that’s what I focus on.

P.P.P.S. Many of the disputes between political scientists and journalists end up being decided on the details. The short story is: political scientists have a lot of time and don’t say much, so when we do say something, we can try to be careful not to make any mistakes. In contrasts, busy journalists are under pressure to say things all the time, so sometimes they get things wrong. I’m sure that political scientists would make lots and lots of silly mistakes too if we had to publish at the rate that journalists do.