NOTE: This is the third item in The Democratic Strategist’s Rountable Discussion on swing and base voters. It’s by Brookings Institution senior fellow and TDS Co-Editor Bill Galston.
It’s hard to discuss “swing voters” without a precise definition of the term. The best I’ve seen so far is that of Northeastern University political scientist William Mayer: A swing voter, he says, is one who could go either way, one not so solidly committed as to make persuasion all but futile. Relative to committed voters, swing voters have mixed or balanced attitudes about the major-party candidates, as measured by the difference between the two on the American National Election Studies (ANES) so-called “feeling thermometer.” If the maximum theoretical difference is 100, corresponding to total approval of one candidate and disapproval of the other, voters who see a gap of 15 points or less constitute the pool of potential “persuadables.” Since 1972, swing voters, so defined, have averaged 23 percent of the total in ANES preelection surveys.
Mayer does not underscore a crucial point that emerges from his data: the ideological distance between the major-party candidates strongly affects the percentage of swing voters. In 1976, the Republicans nominated a moderate after a fierce intra-party battle, while the Democrats nominated a newcomer widely regarded as their most conservative candidate in decades. As a consequence, swing voters amounted to a full 34 percent of the electorate, by far the highest in any election in the past three decades. In 2004, by contrast, the Democrats nominated a Massachusetts liberal to run against a Republican who had governed as a movement conservative. In this context, swing voters amounted to only 13 percent of the total, by far the lowest in the past three decades.
This relationship has important consequences for 2008: the Democrats’ choice between Hillary Clinton and Barack Obama will shape, not only how swing voters will vote, but also how many swing voters there will be. Consider a recent Greenberg Quinlan Rosner/Public Opinion Strategies survey for NPR, which compares the two Democrats against McCain. In a Clinton/McCain contest, McCain receives 9 percent of the Democratic vote, while Clinton receives 5 percent of the Republican vote. In an Obama/McCain matchup, by contrast, McCain gets 18 percent of the Democratic vote and Obama, 13 percent of Republicans. Roughly speaking, a general election between Obama and McCain roughly doubles the number of partisans who can contemplate crossing party lines. All other things being equal, a less polarizing contest, when voters can see advantages and disadvantages in both candidates, will expand the pool of voters open to persuasion.
The net effect of this expansion depends on circumstances. The NPR survey indicates than McCain receives only 48 percent of Independents when facing Obama, versus 58 percent against Clinton. But because Clinton does a better job of consolidating the base, McCain ends up with same share of the total vote—48 percent—against each of the Democratic contenders.
A snapshot cannot foreshadow the dynamics of the general election, of course. The partisans of Sen. Obama can argue, plausibly enough, that wavering Democrats will “come home” as the differences between him and Sen. McCain are cast in higher relief. The partisans of Sen. Clinton can retort that this same process of differentiation will reduce Obama’s appeal to Republicans, and that in addition, as Independents learn more about the gap between his unifying rhetoric and his traditionally liberal position on the issues, some of them will switch as well. To pile uncertainty on uncertainty, the relationship between Independents and swing voters is loose at best: as Mayer shows, many Independents are covert partisans, and they make up only 41 percent of the total pool of swing voters. At this point, we have far more variables than equations, making prediction impossible.
What we can say is this: if a McCain/Obama contest did no more than increase the share of swing voters from its 2004 low to its post-1968 average, an additional 10 percent of the electorate would be in play after the parties’ conventions, shifting both campaigns away from all-out mobilization and toward persuasion. By contrast, if Sen. Clinton turned out to be as polarizing as her detractors maintain, the tone of the 2008 election could bear more than a passing resemblance to 2004 . . . which is not to say that the outcome would necessarily be the same.