Polarized: Making Sense of a Divided America

Polarized: Making Sense of a Divided America

by James E. Campbell


View All Available Formats & Editions
Choose Expedited Shipping at checkout for guaranteed delivery by Thursday, January 24

Product Details

ISBN-13: 9780691180861
Publisher: Princeton University Press
Publication date: 03/27/2018
Edition description: Reprint
Pages: 336
Sales rank: 682,027
Product dimensions: 6.00(w) x 9.20(h) x 1.00(d)

About the Author

James E. Campbell is UB Distinguished Professor of Political Science at the University at Buffalo, State University of New York.

Read an Excerpt



Let's start at the beginning. Gentlemen, this is a football.

— Vince Lombardi, legendary NFL coach

Despite an avalanche of studies in recent years, many important questions about polarization in American politics remain unsettled. Before delving into them, as Coach Lombardi ever so gently suggests in the epigraph, it is a good idea to begin with the basics, to establish what we know (or think we know) and what we do not know about the subject. This taking stock and clarifying the concepts and the questions of polarization is the principal purpose of this chapter.

It might seem unnecessary to state the case for why the subject of polarization is important, but it should be stated nevertheless. It is our football. The level of polarization is important because it affects every aspect of political life, from discussions of political issues in informal settings, to the conduct of elections, to interactions at the highest levels of government over national public policies. As noted in the Introduction, polarization is the condition of substantial and intense conflict over political perspectives arrayed along a single dimension — generally along ideological lines. Rather than political differences being aligned differently on a variety of unconnected issues on which opinions are weakly held, highly polarized conflict is aligned across different issues with strongly held and divergent views. In highly polarized politics, there are distinct sides who see the political world in diametrically opposite ways. Whether in discussions over a few beers in a pub or on roll-call votes cast in the halls of Congress, polarized conflict is not easily resolved or accommodated.

Some degree of polarization is a fact of life in politics. Politics, especially free and democratic politics, involves conflict over differing views. But the extent of polarization can vary, and this is important. The level of polarization in the public and between its political parties establishes the "degree-of-difficulty" that the nation and its political system confronts in governing. A less polarized public and party system makes the political process less contentious and its results more broadly satisfactory. Greater polarization translates into more difficult and combative politics. Considering the gravity of the stakes involved, anyone who cares about or who might be affected by American politics should have an interest in understanding its polarization. If you are not in this group, you are excused.


To move further down the road of understanding polarization in modern American politics, we need to know where we are starting from — what do we already know about polarization? While we know a good deal about it, five observations and findings are most important. The first two are observations about the character or nature of polarization. These are relevant to understanding the magnitude of both polarization in general (first-order polarization) as well as polarization between the political parties (second-order polarization). The three remaining "knowns" are basic empirical findings about party polarization on which there seems to be general agreement.

A Matter of Degree

Polarization is not an all-or-nothing condition. It is a matter of degree. Polarization is about antagonistic political perspectives confronting one another (a single dimension of conflict). Political differences are always present to some degree, but they are never all-consuming. Although the complete absence of polarization and the complete dominance of polarization are hypothetical possibilities, they are not realistic possibilities.

The idea that polarization is a matter of degree may be best appreciated by reviewing what the extreme conditions of no polarization and complete polarization would require. There are only two conceivable ways in which the politics of a free society could be free of polarization. The first of these is unanimity. If everyone were of one mind on all important matters of public policy, politics would be entirely unpolarized. This is, of course, well beyond the realm of possibility. The only other way to avoid polarization entirely is if differences on important matters were distributed in such a random way that there were no stable alliances formed across different matters in dispute — in effect, perfect pluralism. Allies on one issue were as likely as not to be opponents on the next issue. This, too, is beyond the realm of possibility. Common strands of thinking on issues build relationships — friends and foes.

At the other hypothetical extreme, politics would be completely polarized when two sides of equal size were as far apart as possible in their perspectives. Between two homogeneous and diametrically opposed perspectives would be a vast and vacant chasm of the political center. Even civil wars might not reach this extreme.

Real world politics are considerably more messy than these hypothetical extremes of no polarization and complete polarization. Public opinion is complex, motivated in part by common interests, often not well formed and articulated, and routinely measured with error. With these characteristics, simple pictures of universal agreement or random differences, on the one hand, or neatly concentrated bimodal divisions, on the other, are not remotely realistic possibilities.

As a consequence, whether with reference to the public or to the political parties, there is no bright line over which normal political differences become polarized political differences. In statistical parlance, polarization is an interval condition, not a dichotomy. No one would seriously contend that adding one more contentious ideologue to an otherwise unpolarized society or legislature would "flip the switch" and convert it to a polarized body. The real question is not whether American politics are or are not polarized, but to what degree are American politics highly polarized or relatively unpolarized?

This observation may seem obvious, but debates about polarization are often conducted as if they were about the extremes. We are polarized or we are not. Those seeing politics as fairly moderate tend to downplay the extent of real conflict. Those seeing a polarized America tend to look past the fact that differences among ordinary citizens as well as differences between the political parties could be larger and even more intense.

The interval nature of polarization has many implications. It is an important reason why: (1) polarization research has been so frustratingly inconclusive; (2) data used to address questions about polarization are often inadequate or problematic; and (3) the question about the extent of polarization so often becomes conflated with the question about whether the level of polarization has changed. Because polarization is a matter of degree, one needs a benchmark or some way to determine what is a large amount of polarization and what is a small amount. We are or are not largely polarized compared to what? It is difficult to determine with much of the data whether the polarization glass is two-thirds full or only one-third full.

The benchmark difficulty can be illustrated by a hypothetical example of two polling questions. Suppose a poll reported that only 15% of Americans loved their country. Most would conclude that this was a dangerously low number. On the other hand, suppose a poll reported that 15% of Americans advocated the violent overthrow of the government. Still 15%, but most would probably conclude that this was a dangerously high number. In both cases, we would have some basis for expectations and the observed percentage would have been lower or higher than those expectations. It is not self-evident what our expectations should be for a normal level of polarization, and so some benchmark is needed to make sense of the data. If half of the nation is ideologically committed with one-quarter strongly opposing another quarter, is this a high or low level of polarization and why should we think so?

Because polarization is a matter of degree and there are no established benchmarks for determining whether various measurements indicate that polarization is great or small, assessments of the level of polarization often slide into examinations of the change in the level of polarization. The past is implicitly used as a benchmark and one not carefully validated or calibrated. If polarization increased (or decreased or stayed the same), was this a change from a condition when political conditions were highly polarized or substantially unpolarized?

There is an irony in all of this. The one thing that we know or should have known all along about polarization, that it is a matter of degree, is a major reason why we do not know more about polarization. Examinations of polarization's direct evidence in chapters 3 and 4 and its circumstantial evidence in chapter 5 offer several solutions to the benchmark problem.

Small Change, Big Difference

Attention to the degree of polarization cautions us to be careful in interpreting evidence of the level of polarization. The second point warns us to be equally careful in interpreting the significance of any change in polarization levels. Changes in aggregate political conditions (macropolitics) may appear to be small, but seemingly small changes may have huge consequences.

Some perspective on meaningful changes in polarization may be gained from other aspects of macropolitics, particularly other areas that involve more deeply held and durable political dispositions such as changes in the normal vote and in macropartisanship. The normal vote division is the vote that one would expect in the absence of conditions temporarily favoring one party or the other. It is essentially the electorate's default vote division. A sizeable change in the normal vote is an indication of a partisan realignment, a change in the balance of electoral power between the parties. A closely related concept is macropartisanship. It is an aggregate measure of party identifications. It is computed as the percentage of party identifiers (Democrats and Republicans) who identify themselves as Democrats. Like the change in the normal vote, a sizeable change in macropartisanship may be read as evidence of a party realignment.

Realignments are infrequent but important in reshaping American politics. They have huge political repercussions. The historic changes arising from realignments are based on changes in the normal vote and macropartisanship that might at first appear numerically small. The momentous realigning changes in the normal vote division between the parties and in macropartisanship historically have been around ten percentage points and often smaller. The realignment of 1896 changed the normal vote by about six percentage points. This shifted the political balance of power from being quite even between the parties to one of Republican dominance. The New Deal realignment in the 1930s was undoubtedly the greatest political change in the twentieth century, ushering in a Democratic majority that dominated American politics through the 1960s. The New Deal realignment was a shift of about ten or eleven points in the normal vote.

The current era of competitive party politics resulted from single-digit percentage changes in macropartisanship and the normal votes. The realignment establishing the current parity between the parties came about through a shift in the normal vote of only about five percentage points. Macropartisanship in this realignment (between 1980 and 1984) shifted about seven percentage points from a decided Democratic Party advantage to near parity. These realignments had enormous political repercussions, but the extent of change could easily be misread as minor. The point is that since polarization involves long-term political orientations akin to partisanship and voting habits, significant changes in polarization may easily be misread as minor or even inconsequential.

The idea that what looks like a small change can make a big difference is by no means unusual or in any way limited to politics. Examples exist everywhere. A few feet to the left or right in a last-second 47-yard field goal attempt may not sound like much, but they can be the difference between a Super Bowl win or loss. A long fly ball hit a few feet to the left or right of a foul pole in a World Series game can be the difference between a walk-off win and just another strike. Outside of sports and politics, the change of a few degrees in temperature may not sound like much, but it could make the difference between water and ice. The global warming or climate change issue is based on concerns raised by a change of less than two degrees Celsius in the earth's average surface temperature.

The most dramatic illustration of how seemingly small differences potentially can have big consequences is offered by the bonobo. In politically incorrect times, the bonobo was known as the pygmy chimpanzee. Geneticists studying the bonobo's genome have determined that the genetic blueprint of bonobos differs from that of humans by a mere 1.3% to as little as four-tenths of 1%. This numerically tiny difference obviously makes a huge difference (though the reader may have some possible exceptions in mind). Monkey business aside, while small change may not necessarily have great consequences, it can be a big deal.

The McClosky Difference

The third point in analyzing polarization is drawn from the seminal study "Issue Conflict and Consensus Among Party Leaders and Followers" by Herbert McClosky, Paul J. Hoffman, and Rosemary O'Hara. The McClosky group and those following in their footsteps found that ideological and issue differences are greater between the parties' leaders than between their followers. That is, the Democratic Party's leaders generally disagreed more with the Republican Party's leaders than the Democratic Party's followers differed with their counterparts in the Republican Party. Party leaders in the study were national convention delegates, and party followers were drawn from two national surveys of voters. Along the same lines, Philip Converse found that elites, this time congressional candidates, held more consistently liberal or conservative positions on issues than did average citizens.

The importance of the McClosky group's findings is that they mesh with other research on differences between elite and mass opinions in suggesting that opinion formation is more definite and polarization much greater among those who care more and think more about politics — in this case, the elites or leaders. This should not be surprising. It is actually the business of a party's elected officials to know about political issues and to articulate their positions on those issues. They are the most conversant with political terms and ideas and most able to express them in a sophisticated way. More sharply defined opinions lead to greater differences of opinion.

Moving down the scale of political engagement, when a question asking for their opinion pops up in a survey, one is likely to run into opinions that are less well shaped and adhered to, and somewhat fuzzier views of politics and certainly less well formed for expression. Although less refined and less clearly articulated than the views of those more deeply immersed in politics, those opinions are no less real. Differences among those at these middle rungs of the political engagement ladder are likely to be more muted and to appear even less sharply defined as expressed. At the very lowest rungs of political engagement, we should expect no real opinions at all and, therefore, the absence of polarization. If someone is clueless about politics, completely unconcerned about it, we should expect non-attitudes — responses with no real cognitive content behind them. Without opinion, there can be no real conflict and no polarization. Essentially, the politically unconnected produce nonpolarized noise.

This point has several implications. First, we should not expect average citizens or voters ever to appear as polarized as political leaders (even when they may actually be as polarized). The views of leaders are more refined and focused and even practiced in their expression. The views of everyday citizens and voters are not nearly as well-honed or crystallized. As V. O. Key observed long ago, much of public opinion is latent or hibernating opinion that should not be confused with non-attitudes (an absence of opinion) or, for that matter, with moderate opinions. Latent views are not expressed with the same clarity, certainty, and confidence as the views of leaders, but they may find a better expression in the more sharply articulated views of leaders. This is essentially a broader application of U.S. Supreme Court Justice Potter Stewart's observation about recognizing obscenities. Justice Stewart admitted that he had trouble defining what an obscenity was, but wrote that "I know it when I see it." Voters also may not be able to identify their policy preferences at a moment's notice in a survey, but they know what they like and what they don't like when they hear positions articulated by political leaders.


Excerpted from "Polarized"
by .
Copyright © 2018 Princeton University Press.
Excerpted by permission of PRINCETON UNIVERSITY PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

Table of Contents

List of Figures and Tables, ix,
Acknowledgments, xi,
Introduction, 1,
Part One. Preparing the Foundation, 13,
Chapter 1. Knowns and Unknowns, 15,
Chapter 2. History and Theories, 39,
Part Two. The Polarized Electorate, 59,
Chapter 3. Ideology and Polarization, 61,
Chapter 4. Issues and Polarization, 91,
Chapter 5. Circumstantial Evidence, 117,
Part Three. The Polarized Parties, 143,
Chapter 6. Why Are the Parties More Polarized?, 145,
Chapter 7. One-Sided Party Polarization?, 173,
Chapter 8. Why Are the Parties Polarized at All?, 197,
Chapter 9. Polarization and Democracy, 221,
Afterword, 247,
Appendix A. Five Ideological Series, 257,
Appendix B. Regression Analyses of Ideological Orientations, 259,
Notes, 263,
References, 293,
Index, 323,

Customer Reviews

Most Helpful Customer Reviews

See All Customer Reviews