That’s what Ella Fitzgerald used to sing, and that’s also the message that comes through in the just released The Swing Voter in American Politics, edited by William Mayer and featuring some intriguing new analyses by an array of political scientists and other campaign experts.
To get things started, Mayer asks what a “swing voter” is, anyway. After considering several different possibilities, he settles on a definition based on the familiar “thermometer scale” ratings of candidates, in which survey respondents rate their affect toward each candidate on a scale that runs from 0 (cold) to 100 (hot). To identify swing voters, Mayer simply subtracts one candidate’s rating from the other’s. So if I like one candidate a lot (say, 80) and the other candidate just a little (say, 20), my score would be 60. Depending on which candidate I like more, my score could therefore range from -100 to +100. After playing a bit with the numbers, Mayer settles on those in the -15 to +15 range as swing voters. Thus defined, swing voters constituted almost one out of every four respondents in the American National Election Study surveys, 1972-2004.
With those definitional niceties out of the way, Mayer in effect poses the musical question of whether swing voters mean a thing — that is, how influential they are in determining the outcomes of presidential elections. To address that question, he calculates the percentage of “base” (i.e., non-swing) voters who actually voted for the candidate to whom they assigned the higher thermometer rating, and finds that on average, the major party presidential candidates hold on to 96% of their base voters. “The problem for most campaigns,” he continues, “is that the base vote falls short of a majority.” That’s where the swing voters come in. In close elections (like those of 1976, 1980, 1992, and 2000), “which candidate wins will depend on how the swing vote breaks”; in each those elections, the popular vote majority went to the winner of the swing vote majority. Naturally, in landslides like the 1972, 1984, and 1996 elections, swing voters don’t mean much — they help “determine the size of the winning candidate’s victory,” but they don’t determine the winner. That’s because in those elections one candidate’s thermometer scores were so much higher than the other’s that the gap between them in their base voters was sufficient to carry the day.
There’s a ton of good stuff in this small (151-page) volume. Stay tuned for future posts on this topic as the 2008 campaign wears on.