My wife hates presidential campaign season. Like many other political scientists (you know who you are), I carry on a running argument with election commentators on TV: “That’s not quite right.” “Not true in general.” “That is totally wrong.” “Not according to ANES data.” “Give me a break, what about the____ election?” And so on, and so forth. Like all election seasons, the 2012 campaign was rich in commentary that was at odds with or unsupported by findings from political science. In particular, the polarization narrative, which had taken a sabbatical of sorts during the 2008 campaign and its immediate aftermath, enjoyed a resurgence that had little basis in the available data, even though members of the media regularly cited snippets of data that they regarded as evidence supporting their interpretations and commentaries.
Many in the media who write and talk about elections are smart people, but they are not trained as political scientists. I began to wonder: if I could have their attention for a short time, what are the most important things I would want to tell them, insofar as the polarization narrative was concerned?1 As I thought further about the question, I had to concede that we political scientists are responsible for some of their misconceptions. We do not always explain our findings clearly, and sometimes we adopt a particular interpretation of a relationship without adequate consideration of other equally or even more plausible interpretations.
Fundamentals: The Difference between Polarization and Party Sorting
Of all the misconceptions associated with discussions of political polarization, none is more common than the confusion between party sorting and polarization. Many, if not most, discussions in the print and electronic media conflate the two processes. Simple examples illustrate the difference between them. Suppose there is an electorate consisting of 100 Democrats, 100 Independents, and 100 Republicans. These 300 voters include 100 liberals, 100 moderates, and 100 conservatives. At Time 1, the Democrats are a liberal party with a minority right wing, and the Republicans are a conservative party with a minority left wing.
Between Time 1 and Time 2, this electorate polarizes – that is, the middle disappears as all the moderates become liberals or conservatives.
This has NOT happened in the US. Figure 1 shows the trend in ideology, or more accurately, the lack of any trend since the 1970s. The ideological middle (moderates) is the same size as it was decades ago.2 Of course, specific individual issues may not tap into general ideology as closely as one might expect (Ellis and Stimson 2012), but there is little evidence of polarization on specific issues as opposed to general measures of ideology like that in Figure 1. Since 1987, the Pew Research Center has been asking Americans more than 40 questions about socio-economic attitudes, values, and policy positions. According to their 2012 report:
The way that the public thinks about poverty, opportunity, business, unions, religion, civic duty, foreign affairs, and many other subjects is, to a large extent, the same today as in 1987. The values that unified Americans 25 years ago remain areas of consensus today, while the values that evenly divide the nation remain split. On most of the questions asked in both 1987 and 2012, the number agreeing is within 5% points of the number who agreed 25 years ago. And on almost none has the basic balance of opinion tipped from agree to disagree or vice versa (2012, 17).
Rather than polarization in the distribution of public opinion, what has happened in the US is that the parties have become better sorted since the 1970s (Abramowitz and Saunders 1998; Levendusky 2009). The example at Time 2 above shows polarization without sorting: the ideological middle has vanished (polarization) at Time 2, but the parties are no better sorted at Time 2 than at Time 1 – each party still has a minority wing consisting of 20% of the party. Now consider an alternative, Time 2*: the distribution of liberals, moderates, and conservatives at Time 1 does not change, but all liberals become Democrats and all conservatives become Republicans. So everyone is now in the “correct” party.
This is pure party sorting without any ideological polarization. Some analysts call this “partisan polarization.” I prefer the term party sorting for two reasons. First, it distinguishes between the situation at Time 2 from that at Time 2*, both of which have increased partisan polarization relative to Time 1, although different processes have produced the increase in party differences. Second, in common usage, polarization tends to connote a process of individual conversion – individuals move from moderate to more extreme positions as they listen to Rush Limbaugh or watch Rachel Maddow, for example.
In contrast, sorting is more often a compositional phenomenon – rather than change their views, the categories to which people belong change. These correspondences certainly are not perfect; people could sort because they have converted on some important issue or convert on some issue because they have sorted for other reasons.3 Likewise, the two processes are not mutually exclusive. Considering our examples again, if at Time 2, conservative Democrats and liberal Republicans realize that they are hopelessly in the minority in their parties and migrate to the party that shares their views, we would have polarization and sorting.
This is in fact the case in Congress, where we clearly observe polarization (the disappearance of the moderates) and sorting (the decline of conservative southern Democrats and liberal northeastern Republicans). But according to Poole and Rosenthal (2007, chapter 4), there is little evidence of conversion: individual-level stability is the rule in congressional voting. Party polarization has occurred because more extreme members have replaced less extreme ones at the same time that the parties have sorted.
While the two processes are not mutually exclusive, the evidence at the level of mass public opinion indicates that sorting is the dominant process in producing today’s historically high partisan conflict. Baldassarri and Gelman (2008) conclude that sorting is the primary mechanism underlying public opinion change between 1972 and 2008. Krasa and Polborn (2012) find that sorting is the dominant mechanism of change between 1976 and 2004.4
Thinking about the major factors that have contributed to party sorting, such as the realignment of the South (Shafer and Johnston 2006) and other long-term, large-scale social changes (Stonecash, Brewer, and Mariani 2003; Fiorina and Abrams 2009), it seems to me that replacement that produces better sorted parties is the more important causal mechanism underlying the electoral change observed during the past generation.
Before turning to some frequently misunderstood consequences of party sorting, I hasten to emphasize that the extent of sorting in the mass electorate is nowhere near as extreme as depicted in the examples above. Consider an updated cross-tabulation originally noted by Levendusky (2009). At the 2012 presidential nominating conventions, the delegates of both parties adopted extreme positions (relative to public opinion) on abortion. The Republican platform said essentially, “at no time, under no conditions.” The Democratic platform said essentially, “at any time, for any reason.” Now consider the positions of strong Democratic (about 20% of the population) and strong Republican (about 12% of the population) identifiers (Table 1). In the 2008 ANES, 11% of the strong Democrats queried said that abortion should never be permitted, and 26% that it should only be permitted in case of rape, incest, or a threat to a woman’s life.5 So, more than one-third of strong Democrats were closer to Mitt Romney’s position on abortion than to that of their own party. Perhaps even more surprisingly, 22% of strong Republicans said that abortion should always be available as a matter of personal choice, and another 16% in case of a clear need. So, more than a third of strong Republicans were closer to the Democratic position than that of their own party. Weak identifiers were even more out of line with their parties’ positions.
Thus, while party sorting is an important phenomenon and a major contributor to changes in electoral behavior over the course of the past several decades, the picture of regiments of voters marching in lockstep with their parties like members of Congress or the convention delegates discussed above is a serious exaggeration. Even though voters may express a clear preference for one party or the other, many still disagree with that party on one or more issues (Hillygus and Shields 2008; Pope 2012).
Many further correlations thought to reflect polarization are in fact reflections of sorting. For example, ANES data show that split-ticket voting (President-House, President-Senate) has declined since the 1970s and 1980s, suggesting to some a hardening of partisanship. But a plausible alternative hypothesis is that on average House and Senate candidates today look more similar to the presidential candidates of their parties than they did a generation ago. There are far fewer districts now where a Republican voter might be tempted to cross over for a conservative Democratic House candidate, or a Democratic voter might be tempted to cross-over for a liberal Republican House candidate.6
In short, compared to the 1970s and 1980s, if you are closer to the presidential candidate of one party in the 2000s, you are likely to be closer to the congressional candidates of that party as well. Elite party sorting would be expected to produce a decline in split-ticket voting even in the absence of voter sorting. The latter would only strengthen the expectation.
Similarly, just as party sorting would be expected to produce an increase in voting consistency across offices, so should it produce an increase in voting consistency over time. If a voter finds herself closer to the Democrats in 2008, she is more likely to find herself closer to the Democrats in 2012 than say, a voter in 1972 (George McGovern) compared to 1976 (Jimmy Carter). In general, party sorting should produce increased behavioral consistency.
Another relationship considered to be indicative of increased polarization is the finding of “affective” polarization: partisans dislike each other more than they did a generation ago (Shaw 2012). Journalists publicize findings like that of Iyengar, Sood, and Lelkes (2012) who report that people are more likely to be upset about their children marrying someone from the other party today than they were in the 1960s. Yet party sorting is a natural explanation for why partisans feel more negatively about the other party today.
Suppose a son or daughter from a Democratic household came home in the 1960s and reported that they planned to marry a Republican. A parent might well think, “What kind of Republican?” A Barry Goldwater sagebrush conservative? A George Romney midwestern Main Street Republican? A Nelson Rockefeller northeastern liberal Republican? Alternatively, suppose a son or daughter in a Republican household came home back then and announced that they planned to marry a Democrat. As a parent you might wonder “What kind of Democrat?” A western dam builder? A southern conservative? A union stalwart? An urban intellectual? In the heterogeneous parties of 1960, it was very likely that there were plenty of people in the other party who shared your views.
Contrast that with the contemporary era, after four decades of party sorting. If the daughter of Democratic parents announces her intention to marry a Republican, the parents probably will think of an evolution-denying global warming skeptic. And if their Republican son intends to marry a Democrat, their first thought might be that he is bringing an America-hating atheist into the family. As noted above, such stereotypes are gross exaggerations, but the point here is that in light of party sorting, there is more validity to them than there were to any partisan stereotypes four decades ago.
Finally, consider the rising correlation between partisanship and the vote. Here, we political scientists are largely responsible for misconceptions in the wider world. Miller (1991), Bartels (2000), and most recently Shaw (2012) show that the relationship between party ID and the vote has strengthened over time, leading some to conclude that party loyalties are “stronger than at any time since World War II” (Abramowitz 2012). Such conclusions go well beyond the data. The estimated coefficients from regressing the vote on party ID would be unbiased only if other correlates of the vote were not correlated with party ID.
Clearly that is not the case, and it is less the case today as a consequence of party sorting than it was before the sorting began. In 1964, a Democrat who strongly opposed civil rights would vote for Barry Goldwater. Likewise, a Republican who strongly supported civil rights would vote for Lyndon Johnson. Variables like racial attitudes were in conflict with party ID for some voters. Today, after decades of sorting, issues and ideology are more in line with party ID. Thus, an increasingly strong relationship between party ID and the vote may only reflect the fact that party ID now incorporates the effects of other variables that used to be less correlated with or even negatively correlated with party ID. Are party loyalties stronger today? They very well may be, but the only way to know for sure is to include those other correlates in the equations, which brings up another problem.
Kerry is not the same as Kennedy, and Bush is not the same as Nixon
As just discussed, temporal comparisons made by election commentators often implicitly assume that any change in electoral behavior indicates a change in the voters, overlooking the fact that the change may be in the alternatives between or among which voters choose. Did Republicans defect more than usual in 1964 (Goldwater) and Democrats more than usual in 1972 (McGovern) because their party identifications experienced a sudden weakening in those years? To some extent they did, according to ANES data, but the more important part of the story was that each party nominated a factional candidate who was strongly opposed by other factions of the party.
This is an easy example, not likely to be misunderstood, but the same consideration is present in every election. If an issue suddenly recedes or surges in importance, and/or there is a sharp change in electoral outcomes, it may not be the voters who are changing, but the candidates and the positions they take [Fiorina, Abrams, and Pope (2005), chapter 8]. Thus, while surveys consistently show abortion to be a second-tier issue for most Americans, some polls in 2012 showed a rise in importance during the campaign (e.g., Saad 2012). Did that reflect a change in voters’ priorities, or did it reflect the fact that out-of-the-mainstream comments by some Republican candidates made some Americans think, “Gee, these people are farther out on this issue than I had realized.”7
After presidential elections in which one party replaces another or congressional elections in which majority control shifts, one hears pronouncements that the electorate has shifted to the left or right. In part, that is true. Stimson (2004) shows that electoral reversals tend to occur at the end of a shift in voter “mood” in one ideological direction or another. But when a Democratic president takes office, mood shifts in a conservative direction, and when a Republican takes office, mood shifts in a liberal direction (Wlezien 1995). Here it appears that it is not voter preferences that are changing but the policies government produces. Loosely speaking, under Democratic administrations, centrist voters get too much of what they want, and under Republican administrations, they get too little (Stimson 2012).
As another example, consider the concept of “swing voters” (Mayer 2008, 2012). During the 2012 campaign there were frequent references to the small number of undecided voters (e.g., Epstein 2012). There are various ways of estimating the proportion of undecided voters – those who say they are undecided whom they will vote for, those who say they might change their minds, those who actually give different answers in a panel (Hillygus and Jackman 2003) – but I doubt that any method can give us a very precise estimate. The reason is that whether a voter will move between parties is to some extent endogeneous: it depends on the candidates and issues.
In 1984 (Reagan v. Mondale), some surveys indicated that the number of swing voters was in the single digits, but 8 years later, nearly 20% of the electorate swung against both major parties and supported Ross Perot. Did the voters change that much in 8 years? Or more likely, did a large number of them arrive at a different decision when given a different set of choices? It is unlikely that strong partisans will swing, so their proportion of the electorate may give us a rough upper limit on the number of swing voters, but below that limit, different issues and candidates can produce wide variation.
And What About Independents?
The subject of independents is another major source of confusion in election commentary. A first source of confusion is their number: is it more than a third of the eligible electorate and thus the largest “party” in the US, or is it <10%, a number so small that it justifies the modern strategy of mobilizing the base as opposed to the traditional strategy of moving to the center? The second subject of confusion is the composition of the independent category. Are independents a mass of centrists (Killian 2012) or proto-libertarians (Boaz and Kirby 2010) waiting to be mobilized and move our gridlocked political system in a positive direction?
Personally, I wish they were, but they are not. Yes, some are ideological moderates not comfortable in either a too-liberal Democratic Party or a too-conservative Republican one. Yes, some are cross-pressured, like libertarians who are closer to the Republicans on economics and to the Democrats on social issues. But others are uninterested and uninformed about politics generally. And still others are alienated from politics entirely. One thing we can say with some confidence is that those who take the independent label when asked the standard question about party identification are a heterogeneous lot.
How many independents are there? Here political scientists bear a great deal of responsibility for the rampant confusion that exists. Political scientists traditionally divide independents into “pure” independents and “leaners.” The latter are independents who reply that they are “closer to” (note, not “think of themselves as”) one party or the other. There is ample evidence that pure independents are less informed and less involved politically than leaners (Keith et al. 1992), but not so the claim that leaners are actually “closet,” “covert,” or “hidden” partisans, a claim stated with great certitude by some political scientists:
Research by political scientists on the American electorate has consistently found that the large majority of self-identified independents are “closet partisans” who think and vote much like other partisans (Abramowitz 2011, 2).…
numerous studies have shown that treating leaners as independents is ‘the greatest myth in American politics.’ … Call them IINOs, or Independents in Name Only. IINOs who say they lean toward the Republicans think and vote just like regular Republicans. IINOs who say the lean toward the Democrats think and vote just like regular Democrats (Teixeira 2012, 1).
The evidence for such assertions is remarkably thin. A careful reading of the “research” and the “numerous studies” shows that the empirical evidence is equally consistent with not one but two propositions:  leaners are covert partisans as often claimed, and  leaners use their vote intention or issue positions to answer the question about which party they are closer to.8 An analysis of ANES panel shows that in 56 of 60 comparisons, leaning Democrats and Republicans are less stable than the corresponding weak identifiers (as well as strong identifiers, of course) in their responses to the party ID item.
On average, leaners hold the same position from wave to wave <40% of the time, compared to 50% for weak identifiers. Moreover, the variance of leaners’ moves on the scale is greater (Abrams and Fiorina 2012). These findings are consistent with Brody’s (1978) and Miller’s (1991) contention that the direction in which independents lean is endogeneous.9 In addition, leaners are more likely to vote for third parties, to assert that no party represents their interests very well, and to rate the party toward which they lean lower than weak identifiers do. Moreover, in a number of respects, the policy views of leaners differ from those of weak identifiers – not greatly, to be sure, but consistently (Abrams and Fiorina 2012).
None of this shows that leaners are pure independents; they are not. But to the extent that the causal arrow is the reverse of that generally posited by many political scientists, equating leaners with weak partisans exaggerates the importance of partisanship, understates the importance of candidates and issues, and underestimates the potential change in the electorate that new candidates and issues could generate. This is a subject that cries out for more research that exploits the wealth of new data now available.
Head-to-Head Polls and Election Returns Say Less About Public Opinion than you Think
Some commentators write as if there is a direct correspondence between candidate choice in polls or elections and the underlying distribution of public opinion. In 2012, there were claims that the closeness of the polls indicated that Americans were deeply divided.10 Once again, closely-divided and deeply-divided are distinct concepts. If every voter were totally indifferent between the candidates, walked into the voting booth, flipped a coin and said “heads Romney, tails Obama,” the expected result would be a 50:50 election, but it would not indicate a polarized electorate.
Only additional public opinion data can shed light on that question. As Pope (2012) puts it, “When scrutinized closely, it becomes clear that Americans often vote in a highly partisan way, yet they are far less likely to think in a clearly partisan fashion.” If votes were divisible, many voters might want to vote something like 66% Democratic, 34% Republican or vice-versa, but the indivisible nature of the voting act prevents such behavior.
According to some political scientists “We all know that 2012 was a partisan, polarized election …” Kondik, Geoffrey, and Sabato (2012) Do we? Partisan probably – most recent elections have been partisan – but at the time of this writing, the data that will allow us to measure how deeply divided the American electorate was in 2012 are not yet available. I suspect that when the relevant analyses are performed, they will show an American citizenry that was less intense about its candidate preferences in 2012 than in 2008, and certainly one less intense than in 2004.11
Always Remember that Most People you Talk to are Abnormal
According to Markos Moulitsas (2012), founder of Daily Kos:
Americans Elect and Unity 08 are history. No Labels an irrelevant joke. Despite repeated efforts by Beltway hacks to appeal to a mythical and nonexistent bipartisan “middle,” it’s clear there is zero appetite for such constructs from the American public.
And from the other side of the political spectrum, conservative commentator and author Stanley Kurtz (2012) concurs:
First, we need to understand that our political divisions are real and growing. They are rooted not in top-down political rhetoric but in profound and lasting social and cultural differences. For a while, analysts tended to make light of our polarization, fruitlessly predicting year after year that our culture war (still raging) was just about to end. If anything, the culture wars have expanded now to include the whole of politics.
Political journalists need to remember that most of the people they talk to professionally are abnormal, that is, that they are statistically far from the average. The political class comes from the tails of the distribution of American public opinion. Political scientists have documented this since the 1950s, when McCloskey and his associates (McCloskey, Hoffman, and O’Hara 1960) pointed out the differences between Republican national convention delegates and the Republican rank-and-file.12 While we do not have comparable data that would enable us to track the degree of divergence over time, it is likely that the process of party sorting has increased this divergence, because as noted earlier, the sorting process is much farther along at the elite level than at the voter level.
Hard though it may be for political journalists to believe, of the approximately 61 million people who voted for Mitt Romney, more than 58 million did not watch Fox News last night. (I am making the generous assumption that all of the 2.6 million nightly viewers turned out and voted Republican.) Similarly, of the 65.5 million Americans who voted for Barack Obama, nearly 64 million did not watch any MSNBC news show last night.13
In the US, there is indeed a great divide within the political class – those Americans most involved with politics, including public officials, candidates, donors, campaign activists, interest group activists, and, lest we forget, the intellectuals and commentators allied with the parties. To journalists who spend their days interacting with people like this, the evidence of polarization is overwhelming.
Yet away from the politicized haunts where the national media is headquartered, the world is different. Most Americans are not ideologues and do not hold extreme views. They feel no need to impress their friends and acquaintances with the purity of their political views and the intensity with which they hold them. And they have no monetary or career incentives to behave like jerks.
Hundreds of Millions of Dollars Spent on the Media Probably does not Make much Difference
I will write “probably” because it is always possible that research on the 2012 campaign will show the opposite. But this is most unlikely. As Diana Mutz (2012) notes, there is an “enormous chasm” between the beliefs held by journalists and the typical voter about the effects of campaign media vs. the findings of political communications scholars. The former believe that the effects of the media are major, even determinative, whereas academic research finds much smaller effects. I shall say no more about this, because given the long history of the disjunction, it is doubtful that academics could change journalists’ minds about this subject if they had a whole semester, not just a seminar. Who are they going to believe: academic researchers, or their own eyes and ears?
Finally, the American Electorate, Collectively, is not Stupid
Beyond being told over and over by self-interested big-spending campaign consultants that TV ads are powerful influences on public opinion and voting behavior, part of the reason that political commentators ascribe such great importance to the media is the implicit assumption by many of them that Americans are dumb. Even though pre-school age children quickly learn to discount commercials on their TV programs, mature adults are supposedly unable to resist the influence of political media.
To be sure, voters are generally uninformed, and many try to avoid politics until the imminent approach of election day makes it impossible. Yet the collective electorate manifests a degree of knowledge and wisdom that gives those of us who have studied that electorate for decades some cause for optimism.14 If only I could say as much about the knowledge and wisdom of the political class.
Kaufman, Petrocik, and Shaw (2008) provide an earlier and more ambitious contribution along the same lines.↩
The difference between the GSS and the lower ANES series is a reflection of the additional response option in the ANES item, “haven’t thought much about this.”↩
Layman and Carsey (2002) report individual-level conversion among strong partisans in the 1992–96 ANES panel.↩
Somewhat counter-intuitively however, they find that sorting and “radicalization” (polarization) are about equal in 2008.↩
2012 data are not available at the time of this writing, but the distribution of responses on this item (and abortion items generally) shows little variance over time.↩
Noting that self-identified liberals increasingly vote for Democratic congressional candidates and conservatives for Republicans, New York Times columnist Charles Blow (2010) writes that “We have retreated to our respective political corners and armed ourselves in an ideological standoff over the very meaning of America.” In fact, liberal and conservative voters may not have changed at all. Compared to a couple of decades ago, in how many House districts today does a liberal voter have a liberal Republican candidate to vote for, and in how many districts does a conservative voter have a conservative Democratic candidate to vote for?↩
In particular, comments relating to rape by Todd Akin in Missouri and Richard Murdock in Indiana.↩
This was the contention of Shively (1980) in response to Petrocik (1974) who first claimed that leaners were hidden partisans.↩
Based on 1970s and 1980s evidence, Miller (1991) concluded that the “how strong” and “closer to” probes were contaminated by short term electoral forces and argued that only the three categories produced by the root survey item should be used. More recently, Norpoth and Velez (2012) conclude “The partisanship of leaners is not a long-term loyalty rooted in a partisan identity acquired from the previous generation. To be sure, leaners vote in fairly large numbers for the presidential ticket, but this may be due to common factors shaping both their party leanings and the partisan vote.”↩
“Obama began his 2012 presidential victory speech praising the unity of our nation. Unfortunately, exit polls reveal a devastatingly polarized electorate propelled him to victory.” Elkins (2012) bases this conclusion entirely on vote splits in demographic groups, with no supporting data on their policy references, which (Pew Research Center 2012) reports as not widening in recent years.↩
The fact that 2.5 million fewer people appear to have voted in 2012 than in 2008 despite 6 million more eligible voters and massive turnout drives by both parties is consistent with this expectation, unless the drop-off in turnout was predominantly due to nonpolitical factors like Hurricane Sandy, which does not appear to be the case.↩
On some issues Democratic delegates were actually closer to the Republican rank-and-file then Republican delegates were.↩
Based on November cable ratings, which are the latest available at the time of this writing.↩
Besides myself I would include Page and Shapiro (1992), Stimson (2004) and other scholars of a certain vintage.↩
Abramowitz, Alan I. 2011. “Setting the Record Straight: Correcting Myths About Independent Voters.” http://www.centerforpolitics.org/crystalball/articles/aia2011070702/
Abramowitz, Alan I. and Kyle L. Saunders. 1998. “Ideological Realignment in the U.S. Electorate.” Journal of Politics 60 (3): 634–652. [Crossref]
Abrams, Samuel J., and Morris P. Fiorina. 2012. “Are Leaning Independents Just Deluded or Dishonest Weak Partisans?” Paper presented at the Conference on Revisiting Party ID, Rome, Italy.
Baldassarri, Delia, and Andrew Gelman. 2008: “Partisans without Constraint: Political Polarization and Trends in American Public Opinion.” American Journal of Sociology 114 (2): 408–446. [Crossref] [Web of Science]
Bartels, Larry. 2000. “Partisanship and Voting Behavior, 1952–1996.” American Journal of Political Science 44 (1): 35–50.
Blow, Charles M. 2010. “The Great American Cleaving.” http://www.nytimes.com/2010/11/06/opinion/06blow.html?ref=charlesmblow
Boaz, David and David Kirby. 1010. “The Libertarian Vote in the Age of Obama.” Policy Analysis No. 658. Washington, DC: Cato Institute.
Brody, Richard A. 1978. “Change and Stability in the Components of Partisan Identification.” Paper Prepared for the NES Conference on Party Identification. Tallahassee, Fl.
Elkins, Emily. 2012. “National Exit Polls Reveal an Extraordinarily Divided Electorate.” http://reason.com/blog/2012/11/09/exit-polls-indicate-an-extraordinarily-d
Ellis, Christopher, and James A. Stimson. 2012. Ideology in America. New York, NY: Cambridge University Press.
Epstein, Reid J. 2012. “The Disappearing Undecided Voter.” http://www.politico.com/news/stories/0812/79504.html
Fiorina, Morris P., and Samuel J. Abrams. 2009. Disconnect: The Breakdown of Representation in American Politics. Norman, OK: University of Oklahoma Press.
Fiorina, Morris P., Samuel J. Abrams, and Jeremy C. Pope. 2005. Culture War: The Myth of a Polarized America. 2nd ed. New York, NY: Pearson.
Hillygus, D. Sunshine, and Simon Jackman. 2003. “Voter Decision-Making in Election 2000: Campaign Effects, Partisan Activation, and the Clinton Legacy.” American Journal of Political Science 47 (4): 583. [Crossref]
Hillygus, D. Sunshine, and Todd G. Shields. 2008. The Persuadable Voter. Princeton, NJ: Princeton University Press.
Kaufmann, Karen M., John R. Petrocik, and Daron R. Shaw. 2008. Unconventional Wisdom. New York, NY: Oxford University Press.
Keith, Bruce E., David B. Magleby, Candice J. Nelson, Elizabeth Orr, Mark C. Westlye, and Raymond E. Wolfinger. 1992. The Myth of the Independent Voter. Berkeley, CA: University of California Press.
Killian, Linda. 2012. The Swing Vote: The Untapped Power of Independents. New York, NY: St. Martin’s Press.
Kondik, Kyle, Geoffrey Skelley, and Larry J. Sabato. 2012. “Closing the Books on 2012.” https://mail.google.com/mail/ca/u/0/#search/sabato/13bb7feb8d3ea6ae
Krasa, Stefan, and Mattias Polborn. 2012. “Policy Divergence and Voter Polarization in a Structural Model of Elections.” Unpublished paper. August 18.
Kurtz, Stanley. 2012. “Two Tickets: Two Americas.” August 11. http://www.nationalreview.com/corner/313751/two-tickets-two-americas-stanley-kurtz#
Layman, Geoffrey and Thomas M. Carsey. 2002. “Party Polarization and Party Structuring of Policy Attitudes: A Comparison of Three NES Panel Studies.” Political Behavior 24: 199–236. [Crossref]
Levendusky, Matthew. 2009. The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans. Chicago, IL: University of Chicago Press.
Mayer, William G, editor. 2008. The Swing Voter in American Politics. Washington, DC: Brookings Press.
Mayer, William G. 2012. “The Disappearing – but Still Imporant – Swing Voter.” The Forum 10 (3): Article 2. [Crossref]
McCloskey, Herbert, Paul Hoffman, and Rosemary O’Hara. 1960. “Issue Conflict and Consensus Among Party Leaders and Followers.” American Political Science Review 54 (x): 406–427. [Crossref]
Miller, Warren E. 1991. “Party Identification, Realignment, and Party Voting: Back to the Basics.” American Political Science Review 85 (2): 557–568. [Crossref]
Moulitsas, Markos. 2012. “No Middle Ground Found.” http://thehill.com/opinion/columnists/markos-moulitas/229931-no-middle-ground-found
Mutz, D. C. 2012. “The Great Divide: Campaign Media in the American Mind.” Daedalus 141: 83–97. [Web of Science]
Norpoth, Helmut and Velez, Yamil. 2012. “Independent Leaners: Ideals, Myths, and Reality.” The Forum 10 (3): Article 7. [Crossref]
Page, Benjamin, I., and Robert Y. Shapiro. 1992. The Rational Public. Chicago, IL: University of Chicago Press.
Petrocik, John R. 1974. “An Analysis of Intransitivities in the Index of Party Identification.” Political Methodology 1 (Summer): 31–47.
Pew Research Center. 2012. “Partisan Polarization Surges in Bush, Obama Years.” http://www.people-press.org/2012/06/04/ partisan-polarization-surges-in-bush-obama-years/
Poole, Keith T., and Howard Rosenthal. 2007. Ideology & Congress. 2nd ed. New Brunswick, NJ: Transaction Publishers.
Saad, Lydia. 2012. “Abortion is Threshold Issue for One in Six U.S. Voters.” http://www.gallup.com/poll/157886/abortion-threshold-issue-one-six-voters.aspx
Shafer, Byron E., and Richard Johnston. 2006. The End of Southern Exceptionalism. Cambridge, MA: Harvard University Press.
Shaw, Daron. 2012. “If Everyone Votes Their Party, Why Do Presidential Elections Vary So Much?” The Forum 10 (3): Article 1. [Crossref]
Shively, Philip W. 1980. “The Nature of Party Identification: A Review of Recent Developments.” In The Electorate Reconsidered, edited by John C. Pierce and John L. Sullivan, 219–236. Sage: Beverly Hills.
Stimson, James A. 2004. Tides of Consent. New York, NY: Cambridge University Press.
Stimson, James A. 2012. “On the Meaning & Measurement of Mood.” Daedalus Fall 23–34. [Web of Science]
Stonecash, Jeffrey M., Mark D. Brewer, and Mack D. Mariani. 2003. Diverging Parties: Social Change, Realignment, and Party Polarization. Boulder, CO: Westview Press.
Teixeira, Ruy. 2012. “The Great Illusion.” http://www.tnr.com/book/review/swing-vote-untapped-power-independents-linda-killian
Wlezien, Christopher. 1995. “The Public as Thermostat: Dynamics of Preferences for Spending.” American Journal of Political Science 39 (4): 981–1000. [Crossref]
About the article
Morris P. Fiorina
Morris P. Fiorina is the Wendt Family Professor of Political Science at Stanford University and a Senior Fellow at the Hoover Institution. His interests include representation, elections and public opinion, and democratic theory generally.
Published Online: 2013-02-09
Kaufman, Petrocik, and Shaw (2008) provide an earlier and more ambitious contribution along the same lines.
The difference between the GSS and the lower ANES series is a reflection of the additional response option in the ANES item, “haven’t thought much about this.”
Layman and Carsey (2002) report individual-level conversion among strong partisans in the 1992–96 ANES panel.
Somewhat counter-intuitively however, they find that sorting and “radicalization” (polarization) are about equal in 2008.
2012 data are not available at the time of this writing, but the distribution of responses on this item (and abortion items generally) shows little variance over time.
Noting that self-identified liberals increasingly vote for Democratic congressional candidates and conservatives for Republicans, New York Times columnist Charles Blow (2010) writes that “We have retreated to our respective political corners and armed ourselves in an ideological standoff over the very meaning of America.” In fact, liberal and conservative voters may not have changed at all. Compared to a couple of decades ago, in how many House districts today does a liberal voter have a liberal Republican candidate to vote for, and in how many districts does a conservative voter have a conservative Democratic candidate to vote for?
In particular, comments relating to rape by Todd Akin in Missouri and Richard Murdock in Indiana.
This was the contention of Shively (1980) in response to Petrocik (1974) who first claimed that leaners were hidden partisans.
Based on 1970s and 1980s evidence, Miller (1991) concluded that the “how strong” and “closer to” probes were contaminated by short term electoral forces and argued that only the three categories produced by the root survey item should be used. More recently, Norpoth and Velez (2012) conclude “The partisanship of leaners is not a long-term loyalty rooted in a partisan identity acquired from the previous generation. To be sure, leaners vote in fairly large numbers for the presidential ticket, but this may be due to common factors shaping both their party leanings and the partisan vote.”
“Obama began his 2012 presidential victory speech praising the unity of our nation. Unfortunately, exit polls reveal a devastatingly polarized electorate propelled him to victory.” Elkins (2012) bases this conclusion entirely on vote splits in demographic groups, with no supporting data on their policy references, which (Pew Research Center 2012) reports as not widening in recent years.
The fact that 2.5 million fewer people appear to have voted in 2012 than in 2008 despite 6 million more eligible voters and massive turnout drives by both parties is consistent with this expectation, unless the drop-off in turnout was predominantly due to nonpolitical factors like Hurricane Sandy, which does not appear to be the case.
On some issues Democratic delegates were actually closer to the Republican rank-and-file then Republican delegates were.
Based on November cable ratings, which are the latest available at the time of this writing.
Besides myself I would include Page and Shapiro (1992), Stimson (2004) and other scholars of a certain vintage.