Campaigns don’t matter, except when they might


As a grad student, one of the most frustrating things I encountered was repeated variations on the notion of minimal campaign effects across several political science classes. It wasn’t frustrating because I disagreed with the literature — I didn’t, and still don’t — but rather because it was narrowly focused on presidential and congressional elections at the aggregate level. “Look,” the collective political science discipline seemed to say, “our models do a great job of predicting control of the White House and the House of Representatives, regardless of what happens in the campaign or who the candidates are.” And they do! But they don’t do a great job, or any job at all, at predicting other potential campaign effects (e.g., the ideological extremity of the incoming House majority), or predicting the outcomes of other types of campaigns.

Over the past couple years I’ve run into variations on this “other campaigns” questions several times in my thinking. Does it matter whether candidates identify their party affiliation in general election ads, perhaps guiding the electorate toward the modeled outcome if they do? And do incumbents have an incentive to leave their party affiliation out of their messaging? What about in low-information state-level races? (By absolute coincidence, in a quick Google Scholar search for “campaigns don’t matter,” I found this relevant AAPOR paper co-authored by a sometime soccer teammate when he was in grad school.) Since the wide open primaries of 2008 I’ve been very curious about campaign effects on primary outcomes and the effects of primary campaigns on general elections — that is, whether a tough primary batters or prepares the winner.

Tuesday’s results raise another such question for me — do campaigns matter in ballot initiatives? Matt Yglesias looks at the Prop 19 (legalizing marijuana in California) post-mortem and has a couple of relevant observations. First, it appears that the hope that Prop 19 would pull young people to the polls the way 2004’s gay marriage bans pulled conservatives didn’t pan out. Young people don’t vote in midterms traditionally, and 2010 was no exception.* Second, even without those young people showing up, Prop 19 still got 46% support. Yglesias notes some other potential drags on support, which we might file under “fundamentals,” such as the idea that libertarian activists might have been more drawn to elections for high-profile offices, but there are also some we could think of as campaign effects. For instance, as he puts it, “Prop 19 was hampered by a severe lack of elite validators.” Neither of the state’s senators supported it, nor the Democratic nominee for governor, not the sometimes moderate sitting governor. At the very least, this cost supporters the ability to argue from authority and had some effect on the campaign’s persuasive approach. But more broadly, I’m curious how different persuasive approaches might work in a campaign like this. The last six years have featured many ballot initiatives to ban gay marriage, providing a multi-case baseline from which to start; unfortunately, they’ve all succeeded, rendering comparisons based on outcome a little nuance-y, but it’s a decent place to start.

* For the record, all of the write-ups that talk about the 18-24 or 18-29 age groups being “Obama’s vote” ought to acknowledge that this year’s 18- and 19-year-olds couldn’t vote in 2008.

Share this post:

    Tags: ,