According to news reports, Arlen Specter’s switch to the Democratic Party resulted from Pennsylvania surveys that showed the Keystone State to be increasingly hostile to GOP liberals in primaries, and Republican candidates of any stripe in general “elections”:http://www.cbsnews.com/stories/2009/04/30/politics/politico/main4979700.shtml . This evidence corresponds to related research on the decline of Republican identification nationally between 2004 and 2008, by “Brady, Rivers and Harbridge”:http://www.hoover.org/publications/policyreview/35390034.html. In my own discussions of these developments with both academic and non-academic colleagues, a couple of crucial questions have arisen: How substantively important are these shifts? Are they at all permanent?
Imagine party identification to lie somewhere on a continuum between relatively permanent and enduring identities such as ethnicity or religion, on one end, and the most whimsical and unstable of opinions on the other. Party identification has commonly been considered to lie closer to the identity end of this “continuum”:http://www.amazon.com/Partisan-Hearts-Minds-Political-Identity/dp/0300092156. It is considered closer to identity because, once formed, it doesn’t change very much over the course of the lifecycle.
An individual’s party identification doesn’t change much because elections are episodic and voters rarely pause to reconsider this basic political commitment. In addition, an initial partisan commitment is likely to be reinforced by subsequent habits of news consumption that confirm rather than challenge partisan viewpoints (e.g., watching Fox News vs. MSNBC). There is more to it than this, but I only have space for a brief summary of the considerations that have supported the view that partisanship is closer to identity than opinion.
According to the partisanship-as-identity viewpoint, then, ups-and-downs in Republican or Democratic identification appear mostly as an artifact of random variation resulting from survey sampling. Once we include the appropriate margins of error around estimates of the percentage of Democrats, Republicans and independents, upward or downward trends in national partisanship largely disappear. Party identification is stable – it is identity.
But there do seem to be changes in party identification that are large enough to lie outside the customary margins of error, as I indicated in reference to the compelling evidence presented by Brady, Rivers and Harbridge. Democrats did make gains at least up until November 2008, and Republicans have lost ground. Since November, interestingly enough, the patterns seem to have changed. A “Pew study”:http://pewresearch.org/pubs/1207/republican-party-identification-slips-nationwide-pennsylvania-specter-switch shows both Republicans and Democrats losing ground to independents. Trends on the comprehensive Pollster.com site, averaging large numbers of polls, show a leveling off after November, with little real change:
Just a few days ago a “study”:http://blog.pos.org/2009/05/a-deeper-look-at-party-identification/ by a respected Republican polling firm (the one that was working for Senator Specter until he switched) suggested that as much as 20-25 percent of survey respondents have considered themselves to be identifiers of the opposite party at some point in the past. If true, this would modify the inherited wisdom about party identification in the following way: Yes, for most people (80%) party identification is stable, but there are those who seem willing to change their party identification as circumstances change. These are voters who are mostly weak party identifiers, nominal Republicans that shift over to the Democratic side, or vice versa, and who may shift back again. A small percentage (< 5%) may eventually move into the strong party identifier camp, but the vast majority of these drifters never had entrenched views to begin with, and probably don’t now, either. The question now being asked all over Washington is, How seriously should party strategists and pundits take these shifts? Certainly they are sizable enough to decide a close election. It wouldn’t have taken many of these drifters to change the outcome of the 2000, 2004 or 2008 elections. So clearly they are worth monitoring. Whether party strategists have control over these shifts, or can gain control, is another matter entirely. Campaign managers and vendors flatter themselves if they think these weak identifiers are easy to reach. The further question of whether the political parties, and especially the Republicans, ought to change their policy positions in reaction to the shift toward the Democrats has come up repeatedly over the past few months. But wholesale repositioning is surely a drastic reaction to changes in party identification that may not be the least bit permanent. Major repositioning is also overly radical in light of the closeness of the last several elections. Both Democrats and Republicans have run very close races nationally. More to the point, it is highly doubtful that the weak partisans who are shifting are switching their party identification based on detailed issue and policy considerations. More likely, they are “nature-of the-times” voters as described in the classic study, _The American Voter_, from 50 years ago. They move according to their vague sense of how things are going with the economy and the presidency. They have not shifted because they have calculated that their current party is out-of-synch on some specific policy stand. Others of them have shifted because they simply like Barack Obama. They won’t be able to articulate exactly why they dislike one party and like another, they just ‘know’ they prefer one. As one informant put it simply to _The American Voter’s_ Campbell, Converse, Miller and Stokes during the 1956 Eisenhower-Stevenson race: “I couldn’t see the Republican Party when Roosevelt was in. Now I can.” These voters are not going to be paying close attention to issues and policies, so it’s not very likely that policy shifts are going to be noticed by them. This doesn’t make them stupid, but it does mean that a good many Americans lead very busy lives of work, family and school. There isn’t a lot of time left over to make a hobby out of following politics. They may also have opinions on issues that they will readily offer up to pollsters. That does not mean that these opinions are of any relevance to their decisions in the voting booth. These are facts that are completely lost on those pundits, strategists, and academic commentators bent on projecting their own unusual interest in politics on everyone else -- or otherwise overintellectualizing the American voter.