Chapter 5 • Partisanship and Party Change

“PARTISANSHIP IS A HELLUVA DRUG,” political scientist Brendan Nyhan told CNN’s Don Lemon on May 13, 2021, when describing why so many Republican politicians were saying things that were not true about the January 6 riot and insurrection at the U.S. Capitol. For good and for ill, people tend to see the world through partisan-colored glasses. Over the past few decades, as the two major parties have become more polarized, individuals’ partisan identity has come to shape how people think about themselves and their political opponents to the extent that even when people agree across party lines, they view the other side with distaste and distrust.1 One analysis showed that, after a contentious, partisan policy curtailing organizing rights in Wisconsin was passed in 2011, more than a third of Badger State residents stopped talking to people they disagreed with about the policy—a sign that partisan identity is deeply ensconcing itself into the daily lives of people.2

Partisanship is the centerpiece of American politics. Indeed, if you want to predict how someone is going to vote for president on Election Day and you can only learn one fact about that person before making your forecast, asking for that person’s political partisanship will give you enough information to make a correct prediction nine times out of ten. As E. E. Schattschneider put it in his 1942 book Party Government, “Modern democracy is unthinkable save in terms of parties.”3 Political parties are organized coalitions working to win elections and govern. They structure public debate, dominate news columns and airtime, and continue to play a central role in American politics.

This chapter addresses partisanship—the sense of attachment or belonging that an individual feels for a political party—and offers a brief history of American parties, an examination of the implications partisanship has for political behavior, and an assessment of the factors that influence partisan change. In Chapter 6, we explore how partisanship helps to shape and constrain public opinion, while Chapter 7 considers, in part, how partisan identification influences media choice and individual susceptibility to media effects.

Learning objectives for Chapter 5 include:

· Developing a historical understanding of party identity and party loyalty in the United States

· Understanding affective partisanship

· Exploring how partisanship influences voting behavior

· Analyzing the conditions for realignment

· Learning the factors that are correlated with partisan change

· Building a framework to think about the future of parties and partisanship


For a century and a half, the U.S. electorate has supported a two-party system of Republicans and Democrats in national politics. Such remarkable stability is largely unknown in other democracies. Within this stable party system, however, voter support for Republicans and Democrats has fluctuated widely, and significant numbers of voters occasionally abandon the traditional parties to support third-party or independent candidates, as occurred in 1992. The aggregate division of partisans in the electorate, shown in Figure 5-1, reveals a wide range of the structure of political conflict from 1952 to 2020, even in elections close together in time. While Democrats have outnumbered Republicans in the electorate for more than half a century, students of American politics know that holding the majority of partisans nationwide is not enough to promise victory at the ballot box. Moreover, the number of partisans in the electorate regularly vacillates. Years in which the total percentage of Democrats and Republicans drops a bit include elections with the third-party candidacies of George C. Wallace in 1968 and Ross Perot in 1992 and, to a lesser degree, 1996. Perot’s 1992 showing of 19 percent was the largest percentage won by a third-party candidate since 1912. Although Ralph Nader’s votes denied the presidency to Al Gore in 2000, his 2.7 percent of the popular vote was an unimpressive figure for a third-party candidate in recent years. Third-party candidates supporting the Green, Libertarian, and other parties garnered about 2 percent of the popular vote in 2012 and over 5 percent in 2016. A difference in vote choice of 0.006 voters across Wisconsin, Pennsylvania, and Michigan would have added an Electoral College vote victory to Hillary Clinton’s popular vote win in 2016. In 2020, just as many people answered the American National Election Studies (ANES) partisanship question by saying they were independents as did so by identifying as Republicans, 31 percent each.

Despite variations in election outcomes and American voters’ differentiated support for the candidates offered to them by the political parties, most voters have a basic and stable loyalty to one party or the other. This tendency of most individuals to be loyal to one political party makes the idea of partisanship, or party identification as it is often called, one of the most useful concepts for understanding the political behavior of individuals. After good survey data became available in the late 1940s, party identification assumed a central role in all voting behavior analysis.4

A line graph of the percent of Democrats and Republicans between 1952 and 2020.Description

Figure 5-1 Partisan Division of Americans, 1952–2020

Source: American National Election Studies, available at

Party Identification

Party identification is a relatively uncomplicated measure determined by responses to the following questions:

· Generally speaking, do you usually think of yourself as a Republican, a Democrat, an independent, or what?

· [If Republican or Democrat] Would you call yourself a strong [Republican/Democrat] or a not very strong [Republican/Democrat]?

· [If independent] Do you think of yourself as closer to the Republican Party or to the Democratic Party?

Leaving aside for the moment the people who do not or cannot respond to such questions, this yields seven categories of participants in the electorate according to intensity of partisanship ranging from strong Democrat at one end to weak Democrat and independent-leaning Democrat to pure independent in the middle, to independent-leaning Republican and weak Republican to strong Republican on the other end.

Because this self-identification measure of party loyalty is the best indicator of partisanship, political analysts commonly refer to partisanship and party identification interchangeably. While many other influences are at work on voters in U.S. society, and partisanship varies in its importance in different types of elections and in different time periods, partisanship is the single most important influence on political opinions and voting behavior.

Partisanship represents the feeling of identification with and loyalty to a political party that an individual acquires—sometimes during childhood—and generally holds through life, often with increasing intensity. This self-image as a Democrat or a Republican is useful to the individual in a special way. For example, individuals who think of themselves as Republicans or Democrats respond to political information partially by using party identification to orient themselves, reacting to new information in a way that fits in with the ideals and feelings they already have. A Republican who hears a Republican Party leader advocate a policy has a basis in party loyalty for supporting that policy, apart from other considerations. A Democrat may feel favorably inclined toward a candidate for office because that candidate bears the Democratic label. Partisanship orients individuals in their political environment, although it may also distort their picture of reality. Attaching a partisan label to an issue preference helps citizens determine what they find to be important and how they tend to think about their own preferences.5

Table 5-1 shows the stability of partisan conflict in the United States during the modern polling era. Democrats have held the party identification advantage, even in years such as 1956, 1972, 1980, and 1984, when Republicans ran away with the White House. One element of the data that jumps out is the fact that the percentage of self-identified independents has nearly doubled since the time series began. Even so, independent partisan leaners are among the most loyal party voters. Indeed, after asking independents which party they lean toward, under 10 percent of the public reveals itself to be truly independent from the two major parties.

Table 5-1

Source: American National Election Studies, available at

The dramatic role that the Southern United States has played in American political history—from seceding from the Union to kick off the Civil War to the days of “solid South” Democratic presidential voting to a region where Republicans have the advantage today—merits a careful examination of partisanship in the South and non-South. Southerners’ partisan loyalties shifted from a South that was almost twice as likely to identify as strongly Democratic compared to the non-South in 1952 to a region that displayed a significant loss of Democratic loyalty (but lack of gain for Republicans) by 1972 to one that was indistinguishable from the rest of the country across the party spectrum by 2012. We consider the changes in the South later, in the sections of the chapter dealing with party systems and realignment.

Partisanship is also interesting to political analysts because it provides a base against which to measure deviations in particular elections. In other words, individual voters’ long-standing loyalty to one party means that, “other things being equal,” or in the absence of disrupting forces, they can be expected to vote for candidates of that party. Of course, campaigns are not conducted in a vacuum; Republican presidential candidates do not simply say to themselves, “Well, there are more Democrats than Republicans, and I’ll never be able to convince any Democrats to vote for me, so I guess I’ll bow out of the race now!” Nor do Democratic presidential candidates say to themselves, “Well, looking at those party ID numbers, I see that there’s no need to campaign—we have this in the bag!” Indeed, voters are responsive to a great variety of other influences that can either strengthen or weaken their tendency to support their usual party. Variations occur from election to election in such factors as the attractiveness of the candidates, the impact of foreign and domestic policy issues, and purely local circumstances. These current factors, often called short-term forces, may move voters away from their usual party choices. If the political predispositions of all the individuals in the electorate were added up, the result would be an “expected vote” or “normal vote.”6 This is the electoral outcome that would be expected if all voters voted according to their party identification. Departures from this expected vote in elections represent the impact of short-term forces, such as issues or candidates.

In assessing the partisanship of the American electorate historically, we have no data to add up individual party identifications to find an expected vote. Survey data of this type have been available only for the past seventy-five years or so. For the period from 1824 to 1968, we base our estimates on the only available data—election returns for aggregate units.7 These data cannot tell us about the voting patterns of individuals, but they do allow us to make assessments of party loyalty and temporary deviations from party by collections of voters. Even though the same set of individuals does not turn out to vote in each election, we use the election returns over the years to indicate the collective partisanship of the electorate. From these data, an estimate is made of the expected vote for the Democratic and Republican Parties. It is then possible to say, for example, that the electorate deviated from its normal voting pattern in favor of the Republican Party in 1904 or that the voters departed from their normal Democratic loyalty in 1952.

Political scientist Philip E. Converse developed a method that can be used as an expectation about vote choice in the absence of short-term forces. This technique uses party identification, expected defection rates, and turnout to generate an estimate of the normal vote. Our analysis finds that the deviation of the actual Democratic vote meanders under the predicted Democratic normal vote, meaning that Democratic presidential candidates since 1968 have rarely done as well as would be expected, given the distribution of party identification. The elections of 2004, 2008, and 2012 were exceptions, as the Democratic candidates’ performances matched Democratic partisan strength in the electorate. In elections in which a third-party candidate won a significant number of votes—1968, 1992, and 1996—both the Democratic and Republican candidates performed below what the normal votes would predict. Strong party support and turnout for Joe Biden in 2020 in Wisconsin, Michigan, and Pennsylvania overcame a problem of lower enthusiasm for Hillary Clinton in those same states in 2016.

Party identifiers are members of various demographic and social groups. There are a wide range of relationships in the U.S. electorate between social characteristics and political behavior. American journalists and party strategists often attribute political trends to such categories as “white populists,” “soccer moms,” or “born-again Christians”; frequently, these explanations rely on so-called bloc voting, such as “the Black vote,” “the senior citizen vote,” or “the Latin@ vote,” implying that some social factors cause large numbers of people to vote in certain ways.

Social groups have a major impact on individual attitudes and behavior, including partisanship. As we discussed in Chapter 4, some of this impact occurs directly through face-to-face interactions with primary groups and social networks, but the secondary groups with which one identifies can also have an impact. The two major political parties in the United States have courted certain social groups and passed policies that benefit some groups over others. These connections have led to the political parties having an image or “brand” that people have in mind when they think of the parties. For example, the Democratic Party is widely associated with the poor and with people of color, whereas the Republican Party is widely associated with the wealthy and white people.

Table 5-2 shows the percentage of party identification in the United States by social characteristics. Men are almost evenly split between the two major parties and are nearly mirror images of each other in terms of partisan strength. However, more women identify as Democrats, particularly as strong Democrats, continuing the reversal of how things were in the 1950s and early 1960s when men were more likely to be Democrats. White people are more likely to be Republicans, especially strong Republicans, while both Black people and Latin@s are more likely to be Democrats. That said, it is worth noting that 19 percent of Latin@s identified as pure independents, the largest percentage of independents in Table 5-2.

Age is another factor that is related to party identification. Adults under forty-four are more likely to be Democratic identifiers. Those in the prime of their professional working years, forty-five to fifty-four, are equally distributed across the two parties while those between fifty-five and sixty-four and those seventy-five and older are more likely to identify as Republicans. Those sixty-five to seventy-four are stronger Democrats.

Education, union membership, and religious beliefs are also related to partisanship. Those who have a high school education or less as well as those with some time in college are equally distributed across the two parties while college graduates are more likely to be Democrats. Union members are more likely to be Democrats, especially strong Democrats. With respect to religious identification, there are more Republican mainline Protestants but more strong Democrats than strong Republicans who are mainline Protestants. Evangelical Christians are overwhelmingly Republican, and atheists or agnostics, the fastest-growing religious group in the country, are overwhelmingly Democrats. Catholics are evenly distributed across the parties.

Identification with these groups is related to vote choice as spouses tend to vote alike, as do neighbors, fellow churchgoers, and even coworkers. Even in workplaces, where crosscutting political conversations are more likely to occur, people are more likely to find like-minded people to talk to, if they talk about politics at all.

Table 5-2

Source: American National Election Studies, available at

Affective Partisanship

As we have noted throughout the book, partisan polarization is on the rise in the United States. One way in which the divide between the parties is particularly acute, and that is different from issue divides among partisans, is with respect to the intense feelings partisans have for their own party as compared to the feelings they express toward the other major party. Alan Abramowitz attributes the rise of this emotional, or “affective,” polarization to the growing intensity of ideological differences between liberals and conservatives.8 The average intensity of one’s partisan preferences, measured by a feeling thermometer asking how warmly or coolly (100 is warmest, 0 is coldest) one feels about one’s party, has increased over 15 percent in the past few decades. Meanwhile, the average difference in temperature one feels toward one’s party and the other party has gone from about 20 degrees to more than 40!

A horizontal stacked bar graph of the partisan feeling thermometers toward Joe Biden and Donald Trump in 2020.Description

Figure 5-2 Partisan Feeling Thermometers Toward Joe Biden and Donald Trump, 2020

Source: American National Election Studies, available at

Figure 5-2 shows the differential levels of animus that partisans felt toward Donald Trump and Joe Biden before the 2020 election. In general, Republicans loved Trump and were cool to Biden while Democrats liked Biden but were downright frozen in their feelings about Trump. Specifically, strong Republicans felt incredibly warmly toward Trump (87 degrees) and just as coolly toward Biden (13 degrees). Strong Democrats were not quite as warm to their party’s candidate (81 degrees) as Republicans were to Trump, but they were almost wholly negative toward President Trump (5 degrees). Independents were warmer to Biden (44 degrees) than Trump (34 degrees), and the gap between feelings of independent-leaning Republicans for Trump as compared to Biden was smaller (41 degrees) than the gap between what independent-leaning Democrats felt for Biden and what they felt for Trump (58 degrees).

Some scholars argue that the roots of affective polarization are in the media messages partisans consume—many of which are quick to denounce the other side not just as wrong but as evil, dangerous, and a threat to the American way of life. Thinking of partisanship as a social identity enables analyses of partisans as members of an in-group and people in the other party as members of an out-group (see Chapter 4). Shanto Iyengar and colleagues found evidence, using this social identity perspective, that some partisans simply loathe the other side, and consuming messages that treat the other party as an out-group exacerbates the negative views partisans ascribe to their opponents across the aisle.9

One reason this happens is that many partisans get their news from different sources than folks in the other major party. Figure 5-3 shares data from the Pew Research Center that show how different Republican and Democratic Party identifiers are when it comes to the news sources they trust the most. Democrats trust more traditional, mainstream sources the most. The only source that garnered a majority of trust for Republicans was Fox News. Notably, three of the top five most trusted sources are the same for each group of partisans (ABC, CBS, and NBC), though Democrats are twice as likely to trust them as Republicans are. The same Pew study also showed that of the thirty prominent news sources Pew asked people about, Republicans distrusted twenty-two of them, whereas Democrats only distrusted eight. We take up other relationships between the media and polarization in Chapter 7.

Two horizontal bar graphs of the percentage of partisans who trust each source for political and election news.Description

Figure 5-3 Partisan Trust in Major National News Sources, 2020

Source: Mark Jurkowitz, Amy Mitchell, Elisa Shearer, and Mason Walker, “U.S. Media Polarization and the 2020 Election: A Nation Divided,” Pew Research Center, January 24, 2020,


As we noted earlier, the standard party identification question, used in almost all political surveys, asks respondents whether they are Republicans, Democrats, or independents and whether they are “strong” or “not very strong” Republicans or Democrats. The likelihood of voting loyally in support of one party varies with the strength of individuals’ partisanship. The defection rates of strong and weak partisans in each presidential election since 1952 are shown in Figure 5-4. Declining party loyalty is apparent as the intensity of partisanship decreases. Strong partisans consistently support the candidate of their party at higher rates than do weak partisans. In most years, Republicans have been a bit more loyal to their party than Democrats, although this is partly accounted for by Southern Democrats who regularly deserted their party in presidential elections in the late twentieth century. By the end of the twentieth century, Southern Democrats were no longer distinctive in this regard; previously defecting Democrats had become independents or Republicans.

Differences in candidate appeal affect the propensity to defect. Few Republicans deserted Dwight D. Eisenhower in the 1950s, Richard M. Nixon in 1972, or Ronald Reagan in 1984; many more left Barry Goldwater in 1964. Similarly, most Democrats were loyal to Lyndon B. Johnson in 1964 but abandoned George McGovern in large numbers in 1972.

Another potential cause of defection is attractive third-party candidates. In 1992, Ross Perot drew defectors from both parties, although more from the Republican side. Ten percent of strong Republicans and 25 percent of weak Republicans defected to Perot. Although few strong Democrats defected to Perot, 17 percent of weak Democrats did. John B. Anderson in 1980 and George C. Wallace in 1968 similarly account for part of the upsurges in defections in those years. In 2020, Libertarian Party candidate Jo Jorgensen, Green Party candidate Howie Hawkins, rapper and entrepreneur Kanye West, and a host of other third-party candidates earned fewer than 3 million of the 158.3 million votes cast.10

Historically, third-party candidates often have been viewed as “halfway houses” for partisans moving from one party to another. Not as dramatic for a partisan as defection to the opposition party, a vote for such a candidate may be a first step away from party loyalty or a temporary blip related to a party’s nominee for president, prevailing economic conditions, or an unusually effective third-party candidate. In any event, support for third parties and an increase in defection rates have generally been symptomatic of the loosening of party ties in eras of dealignment. (See discussion in the Partisans, Realignment, and Party Systems section.)

A different pattern—one of high party loyalty on both sides—was exhibited in the presidential elections of 1976, 1988, 1996, 2000, 2004, 2008, 2012, 2016, and 2020. In these elections, partisans of both parties remained loyal to candidates who were relatively balanced in their appeal.

In the twenty-first century, the high degree of loyalty is also a reflection of partisan polarization.11 Although strong partisans vary in their loyalty from year to year, depending on the candidates offered by their party, this tendency is much more pronounced among weak partisans. For example, the defection rate of strong Republicans falls in a narrow range from around 2 percent in a good Republican year to 10 percent in a bad year. In contrast, weak Republicans are almost as loyal as strong Republicans when an attractive Republican candidate is on the ticket, but nearly 50 percent defected in the disastrous 1964 election. The behavior of Democrats is similar, although both strong and weak Democrats are more likely to desert their party than are Republicans. Clearly, marked departures from the expected vote of a party are accomplished by wooing away the weaker partisans of the opposite party. Figure 5-4 shows that 2016 brought higher defection rates from both strong Republicans and strong Democrats as compared to 2012. However, fewer strong Democrats defected in 2016, a year Democrat Hillary Clinton lost, as compared to 2008, the year Democrat Barack Obama won. While weak partisans on both sides tended to hold the party line, their defection rates were higher than those of weak Republicans and Democrats in 2012.

One way that the 2020 election was different from 2016 can be found in the behavior of weaker Democrats and weaker Republicans. Figure 5-4 shows that weak Democrats were far less likely to defect from their party’s nominee in 2020 than they were in 2016 and weak Republicans exhibited the highest level of defection in two decades. Even so, 2020 was still an election where partisans, especially strong partisans, were comparatively very loyal to their party’s candidates.

A line graph of the defection rates by party identifiers in presidential voting between 1952 and 2020.Description

Figure 5-4 Defection Rates by Party Identifiers in Presidential Voting, 1952–2020

Source: American National Election Studies, available at

The tendency of both strong and weak partisans to vote according to their party identification becomes even more pronounced as one moves down the ticket to less-visible and less-publicized offices. Figure 5-5’s focus on congressional races shows that this is a product of the dominant two-party system nationwide. Even highly successful third-party or independent candidates down the ticket are merely local disruptions that have virtually no impact on national patterns. (Data limitations for some midterm election years—the ANES stopped fielding midterm surveys early in the twenty-first century—prevent breaking down partisans by strong and weak for the entirety of the time series.) The voting behavior of partisans in congressional races since 1952 differs from the presidential data in two significant ways. First, differences between the party loyalty of strong and weak partisans are usually smaller. Second, the defection rate does not fluctuate from year to year nearly as much as in the presidential elections. Both differences are attributable to the lower visibility of congressional races. In a presidential election, the flood of available information means that a particularly attractive candidate or a stirring issue may touch the consciousness of the weak partisans, causing them to defect from traditional party ties; the firmly attached, strong partisans are more likely to resist. In the less-publicized congressional races, the information that might cause weak partisans to defect is less likely to even reach them. In the absence of information about the candidates and issues, weak partisans vote according to their party identification. Recent midterm races, in particular, have had historically low levels of partisan defection on both sides of the aisle. In 2020, Democrats defected in their congressional voting less than they did four years earlier while Republicans defected slightly more in 2020 than in 2016.

Two bar graphs of the defection rates by party identifiers in congressional voting between 1952 and 1920.Description

Figure 5-5 Defection Rates by Party Identifiers in Congressional Voting, 1952–2020

Sources: American National Election Studies, available at Data for 2006 and 2010 from Pew Research Center, November 2006 Postelection Survey and November 2010 Postelection Survey, available at

The intensity of partisanship affects political behavior beyond its influence on the likelihood of voting for or defecting from a party’s candidate. Strong partisans are also more likely to vote in all kinds of elections than are either weak partisans or independents. Though it seems quaint today, one explanation sometimes offered for the low turnout in the late twentieth century was the declining partisanship of the American public.12 The turnout rates of the various categories of partisans and independents for three types of elections—presidential, off-year congressional, and primary—are illustrated in Figure 5-6. Presidential primaries, despite all their accompanying publicity and frenetic campaigning, typically have a lower average turnout than off-year congressional elections. This is especially true in uncontested primaries, which often occur when an incumbent president is seeking reelection. No serious opponents squared off against Donald Trump in 2020 (so much so that the Grand Old Party [GOP] in eight states canceled their primaries). Even so, Figure 5-6 shows the power of strong partisanship, as strong Republicans voted at comparable rates to weak Democrats in the primaries even though Democrats had a highly populated field of would-be nominees for their party.

In congressional voting, unlike presidential voting, Democrats were regularly more party loyal than were Republicans until 1994. This was both cause and effect of the recent disjuncture of national politics, whereby Republicans were stronger in presidential politics and Democrats dominated in congressional politics. Throughout the 1970s and 1980s, the Republicans were able to field more attractive presidential candidates than the Democrats, leading more Democratic partisans to defect in presidential races. In contrast, congressional races saw Republican partisans often defecting to vote for a long-term Democratic incumbent running against token Republican opposition. The situation changed dramatically in 1994, when the Republicans gained control of the House of Representatives in part by fanning the flames of anti-incumbent, anti-Democratic sentiment. Thereafter, with more Republican incumbents for whom to vote, Republican partisans were noticeably more party loyal than they had been in previous congressional elections. Even when the Democrats took back the House of Representatives in 2006, the percentage of GOP defectors in the electorate was low.

A line graph of the voter turnout for president and in a primary in 2020 by partisan.Description

Figure 5-6 Reported Voter Turnout by Partisans and Independents in Presidential and Presidential Primary Elections, 2020

Source: 2020 American National Election Studies, available at

What can candidates do to try and orient persuadable voters to the things the candidates’ parties do well? It turns out that the public has some strong opinions about which issues each major party “owns”—that is, handles better than the other party. The issue ownership hypothesis is that when parties can focus voter attention on the issues they are perceived to own, they are more likely to be successful at the ballot box. Figure 5-7 shows the issues the public thought Democrats and Republicans owned, respectively, in 2020. Perhaps most importantly, the public gave Democrats higher marks on handling the COVID-19 pandemic as compared to Republicans. Breaking down the results by the party of the survey respondent (not shown in the figure), 86 percent of Democrats and 44 percent of independents thought the Democrats would handle the pandemic better. Forty percent of independents and 30 percent of Republicans felt it would not make a difference. Less than 1 percent of Democrats thought Republicans would do a better job, and only (comparatively) 63 percent of Republicans thought their party was the better choice on COVID-19.

As is commonly the case, Democrats were perceived to own environmental and health care issues while Republicans were perceived to own the economy and taxes. With the economy and the pandemic serving as the issues that topped voters’ minds in 2020, we might have expected for Trump to focus on advertising about the economy and taxes and Biden to focus on advertising about COVID-19 and health care. As noted in Chapter 2, Biden focused on COVID-19 in his ads, but actually aired more ads about taxes than anything else—a strategy political science research suggests might have worked on some Republicans. This is because partisans prefer it when candidates in the other party discuss the issues owned by the partisans’ own party as compared to out-party candidates who only focus on what their party does well.13

Six horizontal stacked bar graphs of the issue ownership in the 2020 presidential election.Description

Figure 5-7 Issue Ownership in the 2020 Presidential Election

Source: 2020 American National Election Studies, available at

Are Independents Apolitical?

Independents, who now account for more than one-third of the national electorate if leaners are included, are the most obvious source of additional votes for either party. Although partisans, especially weak partisans, sometimes abandon their party, year after year independents are the largest bloc of uncommitted voters available to both parties. Theoretically, in the current closely divided electorate, the vote of the independents can easily determine the outcome of an election.

The independents’ capacity for shifting back and forth between the major parties is shown in Table 5-3. Each party has, on occasion, successfully appealed to the independents, winning over a large majority to its side. In 1984, the independents voted almost two to one for Reagan over Walter F. Mondale, and Johnson held a similar advantage over Goldwater in 1964. During the years in which the Democrats had a clear advantage in partisan identifiers, Republicans had to win a healthy majority of the independent vote even to stay in close contention. This was the case in the elections of 2012 (Romney vs. Obama), 1976 (Carter vs. Ford), and 1960 (Kennedy vs. Nixon). The election of George W. Bush in 2000 depended on, among other things, the substantial advantage he enjoyed over Al Gore among independents. Bush’s reelection in 2004 was a different story as Democratic candidate John Kerry handily carried the independent vote.14 Obama matched Kerry’s appeal to independent voters in 2008, though Mitt Romney scored more independent support than Obama did in 2012. In 2020, independent voters supported Joe Biden 58 to 36 percent, the largest gap since 1984.

Table 5-3

Source: American National Election Studies, available at

Third-party or independent candidates find unaffiliated voters a major source of votes. In 1992, 27 percent of the independents voted for Perot. His failure to hold those votes in 1996 turned his earlier, impressive showing into a minor story. In 1968, more than 20 percent of the independents gave their votes to Wallace, and in 1980, 14 percent voted for Anderson. Looking at the composition of third-party candidates’ votes, one sees more than half typically come from independents. Furthermore, independents may shift dramatically in voting for president and remain stable in voting for Congress.

On what basis do independents make their vote choices? Two views of independents have competed for popularity. The pundit’s view is of an intelligent, informed, dispassionate evaluator of candidates and issues who, after careful consideration, votes for “the person, not the party.” An alternate view—generally attributed to scholars—is of an uninformed and uninterested voter on whom issue-oriented appeals are less effective. Further analysis will help determine who is right.

The first thing we need to do is make two distinctions among independents. We note these distinctions and then drop them because they complicate the analysis and are usually ignored. First, important differences exist between nonpartisans who identify themselves as independents and those who lack any political identification. A sizable segment of the electorate answers the party identification question by saying that they identify as nothing or that they do not know what they are. According to the coding conventions used by the ANES, most nonidentifiers are included with the independents, but important conceptual distinctions may exist between them and self-identified independents.15 The two types of nonpartisans are included in Table 5-4. Those in one set identify themselves as independents; the others do not think of themselves in terms of political labels. Since 1972, between about one-sixth and one-third of the nonpartisans failed to identify themselves as independents. Even though the electorate generally has become more nonpartisan, it is not necessarily more independent. These situations present different implications for the political parties. Self-identified independents think of themselves as having a political identity and are somewhat antiparty in orientation. The nonidentifying nonpartisans have a less clear self-image of themselves as political actors, but they are not particularly hostile to the political parties. They are less self-consciously political in many ways.

Second, within the large group of people who do not identify with either the Democratic or the Republican Party are many who say they “lean toward” one or the other. These leaners make up two-thirds of all nonpartisans, and they complicate analysis in a significant way. On crucial attitudes and in important forms of political behavior, the leaning independents look like partisans. Independents who lean toward the Democratic Party behave somewhat like weak Democratic partisans, and independents who lean toward the Republican Party behave like weak Republicans.16 As can be seen in Figure 5-8, independent leaners are more similar to weak partisans than strong and weak partisans are to each other.

Table 5-4

Source: American National Election Studies, available at

Samara Klar and Yanna Krupnikov have found, in their research on independents, that people who have preferences that should land them in either the Republican or the Democratic Party are increasingly identifying as independents because they are so disgusted with partisan politics and the parties’ favored candidates for office.17 Increasingly, then, independents have preferences that make them appear like partisans, but they identify as independents because of their distaste for the parties and their candidates. These independents hide their partisanship from pollsters—and even their friends! The same forces that drive undercover partisans to identify as independents—media coverage of partisan sniping and polarization—also drive independents to candidates like Bernie Sanders and Donald Trump, who run under a party label but criticize the party system and party elites. Indeed, Figure 5-8 shows that pure independents favored Joe Biden. Higher percentages of independent-leaning Republicans and independent-leaning Democrats favored Biden in 2020 than favored Hillary Clinton in 2016.

A line graph of the turnout and democratic vote for president by partisanship.Description

Figure 5-8 Percentage of Turnout and Democratic Vote for President by Partisanship, 2020

Source: 2020 American National Election Studies, available at

How appropriate, then, is it to include all independents in one category? On some characteristics, such as ideological self-identification and interest in public affairs, much more variation is evident within the three independent categories than between the several partisan categories. The differences between leaners and pure independents are often greater than those among Republicans or Democrats. Because the concept of independent embraces these three dissimilar groups, there is little wonder that some disagreement exists over what the true independent is like. As we briefly consider in Chapter 6, some of these independents have ideological views on social and economic issues that directly contradict the views of other independents on the same issues. In any case, the point here is that independents are a very diverse group, and efforts to paint them with a broad brush and use that painting to make precise electoral predictions is a fool’s errand.

In other words, as a consequence of including various types of people under the label “independent,” making generalizations about the degree of political interest and information of independents is difficult. Some independents have considerable interest in politics, and others are apathetic. There are more informed, concerned voters among the leaning independents than among other nonpartisans, and the leaning independents are more likely to register and to vote. So are independents attentive or apathetic toward politics? The answer is they are both.

To the student of contemporary American politics, these characteristics of the independent remain important because they determine the independent’s susceptibility to political appeals. We and others have argued that the American electoral system is presently at a time when a fairly large group of potential voters has weak ties to the political parties. The argument is that, when a large portion of the electorate either is independent or exhibits more independent behavior, these people form a pool of potential recruits for one of the parties or a new party. Given the diverse nature of beliefs among the ranks of independents, the likelihood of the emergence of a third party that appeals to a majority of independents is unlikely. Indeed, between 2006 and 2008, anti-Bush and pro-Obama sentiment combined to move some previous independents to begin calling themselves Democrats, at the same time moving some Republicans into the independent category. Several efforts were made to generate interest in a Libertarian Party and a Green Party in the 2016 elections, but each party’s candidate for president struggled to get media attention, struggled to raise money, and, consequently, struggled at the ballot box. In 2020, many independent campaigns also struggled at getting on the ballot. Kanye West’s campaign sued to get on the ballot in Wisconsin, but lost the case since the campaign turned in the required signatures supporting his nomination between 5:00 p.m. (the deadline) and 5:01 p.m. (what West’s team argued the deadline should be). The court decided that time has meaning—and the meaning is that the deadline was at 5:00 p.m.18 Since West turned in his signatures late, he was left off the ballot.


One type of departure from the expected level of partisan voting in an election is usually referred to as deviating change: the temporary deviations from normal party loyalty attributable to the short-term forces of candidate images or issues.19 The amount of deviating change in an election tells how well a candidate or party did relative to the party’s normal performance. In these terms, the Eisenhower victories in 1952 and 1956 and the Nixon landslide in 1972 appear even more dramatic because they represent big Republican margins during a time when the Democratic Party held an advantage in party loyalists. These deviating elections involved substantial departures from the underlying strength of the two parties in the electorate.

Temporary deviating changes may be dramatic and reflect important electoral forces, but another type of change is of even greater interest. On rare occasions in American national politics, a permanent, or realigning, change in voting patterns occurs. In such instances, the electorate departs from its expected voting pattern but does not return to the old pattern afterward. The changes sometimes are large and durable enough to alter the competitive balance between the parties, with significant consequences for the policy directions of the government. Such a period of change is usually referred to as a partisan realignment.20 Electoral analysts usually discuss three major realignments in American history. One occurred during the time of the Civil War and the emergence of the Republican Party; another followed the depression of 1893 and benefited the Republicans; and the most recent followed the depression of 1929 and led to Democratic Party dominance. These abrupt changes in the expected votes of the parties can be seen in Figures 5-9 and 5-10. Each realignment of partisan loyalties coincided with a major national crisis, leading to the supposition that a social or economic crisis is necessary to shake loose customary loyalties. But major crises and national traumas have not always led to disruptions of partisanship, suggesting that other political conditions must also be present for a crisis to produce a realignment. The nature of the realignment crisis has political significance, however, because it generally determines the lines along which the rearrangement in partisan loyalties will take place, as different segments of the electorate respond differently to the crisis and to attempts to solve it.

In general, realignments appear to happen in the following way. At a time of national crisis, the electorate rejects the party in power, giving a decisive victory to the other party—a victory that includes not only the presidency but also large majorities in both houses of Congress. The new party in office acts to meet the crisis, often with innovative policies that are sharp departures from the past. If the administration’s policy initiatives are successful in solving the nation’s problems (or at least if they are widely perceived as successful), then significant numbers of voters will become partisans of the new administration’s party and continue voting for this party in subsequent elections, thus causing a lasting change in the division of partisan strength in the electorate. If the administration in power is not perceived as successful in handling the crisis, then in all likelihood the voters will reject that party in the next election, and its landslide victory in the previous election will be regarded, in retrospect, as a deviating election.

In a realignment, the people who become partisans of the new majority party likely are independents and previously uninvolved members of the electorate, not partisans of the other party. In other words, in a realignment, few Democrats or Republicans switch parties. It appears more likely that independents drop their independent stance and become partisan. Thus, for a realignment to occur, a precondition may be a pool of people without partisan attachments who are available for realignment. This, in turn, suggests a longer sequence of events that forms a realignment cycle.

First there is the crisis that, if successfully handled, leads to a realignment. This initiates a period of electoral stability during which the parties take distinct stands on the issues that were at the heart of the crisis. Party loyalty is high during this period, both within the electorate and among the elected political leaders in government. However, as time passes, new problems arise, and new issues gradually disrupt the old alignment and lead to greater electoral instability. During this period, often referred to as a dealignment, voters are much more susceptible to the personal appeals of candidates, to local issues, and to other elements that might lead to departures from underlying party loyalty. As the time since the last realignment lengthens, more and more new voters come into the electorate without attachments to the symbols and issues of the past that made their elders party loyalists. This group of voters, who have no strong attachments to either party, may provide the basis for a new realignment should a crisis arise and one or the other of the parties be perceived as solving it. One conceptual problem for the realignment perspective is its “either-or” nature. Declaring whether an election is or isn’t a realigning one can mask slow but important shifts in the electorate. As campaigns become more sophisticated and ideologues sort themselves into the party that is best for them, dramatic realigning elections have become less common.

Political historians often divide American electoral history into five party systems—eras that are distinguished from each other by the different political parties that existed or by the different competitive relationships among the parties.21 The transition from one party system to another has usually been marked by a realignment.

The first party system, which extended from the 1790s until about 1824, saw the relatively rapid formation of two parties, the Federalists and the Jeffersonian Republicans. The issue that divided the parties most clearly was their attitude toward the power of the central government. The commercial and financial interests supported the Federalist position of increasing the authority of the central government, whereas Jeffersonian Republicans distrusted the centralizing and, in their view, aristocratic tendencies of their rivals. The parties began as factions within Congress, but before long they had gained organizations at the state and local level and had substantially broadened the base of political participation among the voting population. After 1815, competition between the two parties all but ceased as the Jeffersonian Republicans gained supremacy, moving the country into the so-called Era of Good Feelings.

The second party system is usually dated from 1828, the year of the first presidential election with substantial popular participation, which marked the resurgence of party competition for the presidency. Emerging ultimately from this renewed competition were the Democrats and the Whigs, parties that competed almost evenly for national power until the 1850s. Mass political participation increased, and party organizations were strengthened as both parties sought electoral support from the common people. Although the Democratic Party had come to prominence led by frontiersman Andrew Jackson, by the 1850s both Democrats and Whigs had adherents in all sections of the nation. Thus, when the issue of slavery broke full force on the nation, the existing parties could not easily cope with the sectional differences they found within their ranks. As the Whigs and Democrats compromised or failed to act because of internal disagreements, a flurry of minor parties appeared to push the cause of abolition. One of these, the Republican Party, eventually replaced the floundering Whigs as one of the two major parties that would dominate party systems thereafter.

The intense conflicts that preceded the Civil War led to the basic regional alignment of Democratic dominance in the South and Republican strength in the North that emerged from the war and that characterized the third-party system. But the extreme intensity and durability of the partisan loyalties were also significantly dependent on emotional attachments associated with the war. The strength of partisan attachments after the Civil War was not lessened by the sharp competitiveness of the two parties throughout the system. Electoral forces were so evenly balanced that the Republican Party could effectively control the presidency and Congress only by excluding the Southern Democrats from participation in elections. Once Reconstruction relaxed enough to permit the full expression of Democratic strength, the nation was narrowly divided, with the slightest deviation determining the outcome of elections.

The most dominant characteristic of the Civil War realignment was the regional division of party strongholds, but considerable Republican vote strength was found throughout much of the South and Democratic strength in most of the North. Especially in the North, states that regularly cast their electoral votes for Republican presidential candidates did so by slim margins. Within each region, persistent loyalty to the minority party was usually related to earlier opposition to the war. The intensity of feelings surrounding the war overwhelmed other issues, and the severity of the division over the war greatly inhibited the emergence of new issues along other lines. Thus, a significant feature of the Civil War realignment is its “freezing” of the party system.22 Although later realignments have occurred and a fourth and fifth party system can be identified, after the Civil War the same two parties have remained dominant. New parties have found it impossible to compete effectively (although they may occasionally affect electoral outcomes). The subsequent realignments changed only the competitive position of these two parties relative to each other. Thus, although the choices were frozen following the Civil War, the relative strength of the parties was not. Political scientist Nathan Kalmoe has shown that the Civil War was a partisan war, where partisan voters were stable supporters of their party before and during the war, and created stable party cultures long after the war—as late as 1912!23

Toward the end of the nineteenth century, Civil War loyalties weakened enough to allow new parties, particularly the Populists in the Midwest and South, to make inroads into the votes of both major parties. Following the economic recession of 1893, for which the Democrats suffered politically, the Republican Party began to improve its basic voting strength. In 1896, the formation of a coalition of Democrats and Populists and the unsuccessful presidential candidacy of their nominee, William Jennings Bryan, resulted in increased Republican strength in the East and a further strengthening of the secure position of the Democratic Party in the South. Republican domination was solidified in the Midwest by the popularity of Theodore Roosevelt in the election of 1904. By the early twentieth century, competitive areas were confined to the border states and a few mountain states.

The realignment of 1896 and the fourth party system that followed are appropriately viewed as an adjustment of the Civil War alignment. Few areas shifted far from the previous levels of voting; most individuals probably did not change their partisanship. The issue basis of the alignment was economic. The Republicans advocated development and modernization while opposing regulation of economic activity. The Democrats supported policies intended to provide remedies for particular economic hardships. At a minimum, these issues led the more prosperous, more modern areas in the North to shift toward the Republicans and the more backward, more depressed areas in the South to shift toward the Democrats. These tendencies are based on normal vote patterns and should not obscure the considerable variation in the vote for president during these years, particularly in the elections of 1912 and 1916.

Following the onset of the Great Depression in 1929 under a Republican president, Democrat Franklin D. Roosevelt rode the reaction to economic hardship to a landslide victory in 1932. In his first administration, Roosevelt launched a program of economic recovery and public assistance called the New Deal. The Democrats emerged as the majority party, signaling the start of the fifth party system. The New Deal realignment resulted in far greater shifts than the earlier realignment of 1896, because it moved many of the northern states from Republican to Democratic status. Because the policies of the Democratic administration during the New Deal appealed more to the working class than to the middle class, and more to poor farmers than to prosperous farmers, these groups responded differently to Democratic candidates. The New Deal and the electorate’s response to Roosevelt’s administration considerably sharpened the social class basis of party support. Especially for younger voters during these years, class politics was of greater salience than it had been before or has been since.

This realignment resulted in adjustments in previous loyalties, but it did not override them completely. The New Deal coalition was based on regional strength in the South, which was independent of social class, and further reinforced an already overwhelming dominance in the region. The most incompatible elements in the New Deal coalition were Southern middle-class whites, mainly conservative, and northern liberals, both white and Black, and this incompatibility led to the later unraveling of the New Deal alignment. The erosion of the New Deal coalition occurred first in presidential voting with the departure of Southern white voters from the Democratic Party. In 1964, the states of the Deep South were the only states carried by Republican candidate Barry Goldwater, a stark reversal of one hundred years of history. This pattern continued for the next four decades. Only when the Democratic candidate was a Southerner (Jimmy Carter in 1976 and Bill Clinton in 1992 and 1996) did the Democrats have a chance to carry some Southern states. In 2000, Al Gore, also a Southerner, was given a chance of winning only two Southern states—Florida and his home state of Tennessee. Ultimately, he won neither. (Of course, he was running against another Southerner, George W. Bush.) However, Obama’s success in winning Florida, North Carolina, and Virginia in 2008 and Florida and Virginia again in 2012 needs to be viewed against this recent history.

The departure of the South from the Democratic fold is the major reason for the decline of the New Deal coalition. To a degree, working-class whites in the North also have been attracted to the Republican Party on occasion, and middle-class voters—particularly those in service professions—have shifted toward the Democrats.

Survey data on party identification over the past fifty years yield evidence of the New Deal alignment as well as its later deterioration (recall Table 5-1). In the early years of this period, the advantage that the Democrats enjoyed nationwide was largely a result of having an overwhelming Democratic majority in the South. The increased strength of the Republicans in the South after 1964 led to a number of years of fairly even balance nationwide between Democrats and Republicans. Since 2006, the Democrats have gained an advantage over the Republicans, as Democrats increased their strength in the North.

Another important element regarding shifts in the New Deal alignment, also reflected in these tables, was the increase beginning in 1966 in the proportion of independents. Supporters of George Wallace in the South represented part of this increase initially, but an even larger portion is composed of young voters who, since the early 1970s, have not chosen sides in politics as quickly as their elders did. The increase leveled off in the 1970s, and although the proportions have fluctuated, the number of independents remains near its highest point since the era of survey research began. The 42 percent of independents in 2000 is the largest proportion of independents in the history of the ANES.

The New Deal partisan realignment established in the 1930s remained intact longer in congressional voting. However, by the 1970s, additional shifts in the New Deal alignment became evident, as conservative Republicans began to show strength in races for other offices in many parts of the South. Long-standing Southern Democratic incumbents in Congress were safe from competition. As they stepped down, though, their seats were won more often than not by Republicans. Conversely, in some areas of the North, moderate Republicans were replaced by liberal Democrats. In the 2008 congressional elections, not one Republican was elected to the House of Representatives from New England. Democratic strength in the South was crucial for the Democrats’ control of the House of Representatives for much of this historical period.

We should be clear about what is changing and what is not. White Southerners have always been conservative, especially on matters concerning race. From the Civil War until the 1960s, the Democratic Party was at least as conservative as the Republican Party on the crucial issue of race. When the national Republican Party took the more conservative position on race in 1964, white Southerners began to vote for Republican presidential candidates; they continued to vote for Southern conservative Democratic candidates in state and local races. Meanwhile, for the same reasons in reverse, newly enfranchised Black Southern voters were moving into the Democratic Party. Over the years, the positions of the two parties have become more clearly distinguished—the Democratic Party as the more liberal party on racial as well as economic issues, the Republican Party as the more conservative party. Particularly in the South, voters have changed their partisanship and their votes accordingly. Figure 5-9,24 showing the party identification of white voters, North and South, from 1952 to 2016, highlights the dramatic shift in the partisanship of white Southerners over this period. The congressional elections of 1994 were perhaps the moment when the shift became complete. Even so, white Southern Democratic Party identifiers have continued to drop in number while non-Southern white Democrats have continued to fluctuate around the 40–50 percent mark. When the Republicans took control of the House of Representatives and Senate, their leadership was predominantly Southern.

A line graph of democratic identification among southern and non-southern whites between 1952 and 2016.Description

Figure 5-9 Democratic Identification Among Southern and Non-Southern Whites, 1952–2016

Source: American National Election Studies, available at

Are Conditions Right for a Realignment?

The decline of the New Deal coalition is best seen, we believe, as more of a process of conflict extension than of traditional dealignment.25 Voter movement and electoral volatility have been in evidence since the 1960s. Furthermore, much of this movement has been a sorting-out process, whereby some voters are finding their natural home in a political party that shares their views on issues that concern them most. Over this same time period, however, a sizable number of voters have found neither political party a congenial place and have chosen instead to become independent—that is, not to adopt a party identification in the first place or support independent candidates such as Perot or, to a lesser extent, Nader. Through the 1990s, neither party was able to gather the political support to take firm control of government or complete initiatives that would appear to solve societal problems and win converts to their ranks. This sounds a lot like dealignment, so what gives?

Thomas Carsey and Geoffrey Layman have argued that what has occurred in American politics can best be described not as dealignment but as “conflict extension.”26 Along with other scholars, they show that the New Deal divide between Republicans and Democrats has remained on economic issues and that political conflict has been extended to include issues of race and social issues like abortion and gay rights. As Edward Carmines, Michael Ensley, and Michael Wagner have shown, some Americans have liberal attitudes across these issues and are strong Democrats, whereas others have conservative attitudes across these same issues and are strong Republicans.27 However, plenty of people have views that are liberal on one set of issues and conservative on the other. (We discuss this in greater detail in Chapter 6.) Libertarians (conservative on economic issues, liberal on social issues) and populists (the opposite of libertarians) regularly identify as partisans but are more likely to switch their party identification due to short-term forces, are more likely to cast split-ticket ballots, and are less likely to participate. Thus, while elite conflict has extended to a left–right divide across the panoply of political issues currently contested in Washington, the public continues to organize its attitudes separately across economic and social issues. In other words, the New Deal divide has weakened, but it still remains at the same time the parties have adopted competing positions on a whole new set of issues that have pulled some people closer to the parties and pushed others further away. Those who have attitudes on social issues that match the positions offered by their political party have “extended” their conflict with the party on the other side of the aisle. Those who have attitudes on social issues that match one party and attitudes on economic issues that match the other have found themselves stuck in the middle, even though most (weakly) identify with a party.

Into this scenario rode a charismatic young candidate for president who promised change and won the election in 2008 with the largest margin of victory in twenty-four years. His party increased the size of the majorities it had recaptured in both houses of Congress in 2006, giving the new administration control of the legislative and executive branches of the federal government. Six months into the new Congress, the Democrats reached the magic number of sixty votes needed to shut off a filibuster in the Senate, after one switch of allegiance (Arlen Specter of Pennsylvania, who changed from Republican to Democrat) and one resolution to a long-running recount (Al Franken of Minnesota). Public opinion polls found fewer people who identified with the Republicans and more identifying with the Democrats. The new administration took office amid great euphoria, sky-high approval ratings, and an ambitious agenda of health care reform, energy independence, stopping climate change, and restoring international respect and prestige. Several of the ingredients for a realignment were there—a crisis (or crises); an electorate willing to throw the rascals out; a pool of voters without affiliation to either party, available for conversion; and unified control of government, giving the new administration the ability, in principle, to enact its policy agenda. Pundits argued that the opportunity was there for the new administration to capture the imagination of those available independents, turn them into Democrats, and change the partisan division in the country for the foreseeable future. The United States was poised for a true realignment and a new party system.

But . . .

As we contemplate this scenario, we need to keep in mind that many of the voters up for grabs in any given election have a series of issue preferences that do not match up perfectly with either party. In Chapter 6, we describe how Populist voters were more likely to be strong Trump supporters during the 2016 primaries—a departure from their relatively consistent association with Democrats over the previous few decades. If the 2008 election portended a true realignment, Hillary Clinton would have been elected president in 2016. On the one hand, she did win the popular vote, but on the other hand, Republicans made gains in Congress as well, something we would not expect in a realignment advantaging the Democratic Party.

Thus, long-term conversion of these voters is a difficult prospect. Carsey and Layman have also shown that people are willing to change their attitudes to align with their political party on issues that are not very important to them, but this has not occurred enough, nor has it been sufficiently durable, to result in a realignment.28 Though the administration was moderately successful at enacting its policy agenda (with big victories, including health care reform, the auto bailout, and the economic stimulus plan), it also had some notable failures (failing to close the detention center at Guantánamo Bay, increase the minimum wage to $9.50 per hour, and usher in a new era of bipartisan cooperation in Washington, DC). The Democrats’ dramatic loss of seats in Congress in 2010 reflected, in part, the frustration of citizens over the unified government of Democrats not getting more accomplished, which also helped to prevent a cementing of a new and durable Democratic majority.


Partisanship can be thought of as a basic attitude that establishes a normal or expected vote, an estimate about how individuals or populations will vote, other things being equal. However, partisanship itself is not unchangeable. Individuals may change not only their vote but also their long-term party identification from one party to another. More important, over extended periods of time, the partisan composition of the electorate may be altered as new voters of one political persuasion replace older voters of another. When the basic partisan division of the electorate changes, a partisan realignment occurs.

In the past, the absence of survey data limited analysis of realignments, but during the current period, the individual processes of partisan change that underlie aggregate shifts in the partisan division of the electorate can be studied. These processes have been a matter of some controversy. One perspective holds that individual partisans are converted from one party to the other during a realignment. Other analysts, noting the psychological difficulty in changing long-held and deeply felt attachments, argue that such change probably comes about through mobilization, not conversion. In other words, the independents or nonpolitical individuals, perhaps predominantly young voters just entering the electorate without strong partisan attachments, fuel a realignment by joining the electorate overwhelmingly on the side of one party.

Some evidence on these points comes from the New Deal era. Although survey research was then in its infancy, some scholars have creatively used data from early surveys to try to answer these questions. Research by Kristi Andersen, reported in The Changing American Voter, reveals high levels of nonvoting and nonpartisanship among young people and new citizens before the Great Depression.29 Those uninvolved, uncommitted potential participants entered the electorate in the 1930s disproportionately as Democrats. Andersen’s findings on the electorate of the 1920s and 1930s support the view that realignments are based on the mobilization of new, independent voters instead of on the conversion of partisans. In contrast, Robert S. Erikson and Kent L. Tedin argue on the basis of early Gallup poll data that much of the increase in the Democratic vote in the 1930s came from voters who had previously voted Republican.30 In the next section, we examine the processes of partisan change in the contemporary period. Although we are in a better position to do so than we were for earlier eras, efforts are still hampered by a scarcity of panel data—that is, repeated interviews with the same individuals at different times. In most cases, it will be necessary to infer individual changes from the behavior of different individuals over time.

Changes in Individuals Over a Lifetime

Two types of change in partisan identification can be distinguished, both of which have significant implications for political behavior. First, an individual may change from one party to another or to independent or from independence to partisanship. Such change is important if a large proportion of the electorate shifts in the same direction at about the same time. Second, an individual’s partisanship may strengthen or weaken in intensity. A long-standing hypothesis states that the longer individuals identify with a party, the stronger their partisanship will become.31 In the electorate as a whole, the two types of change are not necessarily related to one another, so the occurrence of one form of change does not dictate or prevent the other. For example, recent decades saw an increase in the number of independents in the electorate, which can be accounted for by young people not choosing a party or by partisans moving to independence or both. At the same time, the remaining partisans have become more firmly committed and more party loyal, and polarization between the parties has increased.

Analysts have attempted to explain partisan change by referring to three types of causal effects: (1) period effects, or the impact of a particular historical period that briefly affects partisanship across all age groups; (2) a generation effect, which affects the partisanship of a particular age group for the remainder of their political lives; and (3) a life-cycle effect, which produces changes associated with an individual’s age. In current political behavior, all three can be illustrated: a period effect that resulted in increased independence in all age groups, a generation effect that keeps Democratic partisan loyalty high in the generation that entered the electorate during the New Deal, and a life-cycle effect that yields greater independence among the young than among their elders.

The change in particular age cohorts is worth considering. For example, the youngest age cohort in 1972 was more than forty years old in 1992 and had a lower level of independence than it did when entering the electorate. By 2012, that same cohort was more than sixty years old and had stayed about the same in its level of independence. How do age cohorts identify as independents today? The difference between those thirty and under, those thirty-one to fifty-nine, and those over sixty in the proportion of independents in various age groups is shown in Figure 5-10. In 2020, the younger the cohort, the higher the percentage of people who identified themselves as independents. Of those under thirty, 37.2 percent did not prefer to affiliate with a party as compared to only 26.4 percent of those over sixty.

A pie chart titled Percent Independents shows the following three parts in the decreasing order of size: 30 and under, 31 to 50, and over 60.

Figure 5-10 Distribution of Independents, by Age Cohorts, 2020

Source: American National Election Studies, available at

Contrary to political folklore, little evidence exists that people become Republicans as they grow older—that is, that a life-cycle effect favors Republicans. Older members of the electorate were, for some years, more likely to be Republicans than younger members. The generation of young people who came of age before the Great Depression contained large proportions of Republicans, an understandable situation given the advantage the Republicans enjoyed nationally at that time. Relatively few members of this generation changed partisanship over the years, and these individuals constituted the older, more heavily Republican segment of the electorate. By the same token, the generation that entered the electorate during the New Deal was disproportionately Democratic. Because they also remained stable in partisanship, older voters looked increasingly Democratic as this generation aged.

However, when party voting is frequently disrupted, this reinforcement of partisanship may not occur. Even if the strength of partisanship does not increase with age, older partisans are less likely to abandon their party and become independents. This explains in part why older partisans are less likely to vote for independent or third-party candidates than are younger partisans. In 1992, 19 percent of the Republican and Democratic partisans aged twenty-five and younger voted for Perot, but only 11 percent of partisans aged forty-five and older voted for him. In 2000, Nader’s vote, although small, was greatest among the young.

Gradual changes in individual partisanship have not been assessed satisfactorily for the entire public because the few election studies based on repeated interviews of the same individuals have covered at most four years. Nevertheless, the possibility that individuals change their partisanship over longer time periods is of considerable interest. Speculation has focused on the possibility that the large number of young independents will become identified with one party or the other, thus creating a substantial shift in the overall partisan balance of the electorate. Obama’s appeal to young people raised this as a possibility, but 5 percent fewer voters between eighteen and twenty-nine voted for Hillary Clinton in 2016 than Barack Obama in 2012.32

The best evidence of this type of change in the past comes from a major study of political socialization led by M. Kent Jennings. He surveyed a national sample of high school students and their parents in 1965, with follow-up interviews in 1973, 1982, and 1997.33 This study provides a before-and-after picture of young people during the political traumas of the late 1960s and early 1970s, as well as later snapshots after a more quiescent period.

Table 5-5 shows the amount of change in partisanship between each wave of the study. Partisanship was least stable when the respondents were youngest, between 1965 and 1973. About two-thirds of the sample reported the same partisanship when interviewed in 1982 as in 1973 and, again, between 1982 and 1997. Most of the changes that did occur were between partisanship and independence; relatively few reported switching from Democrat to Republican or vice versa.

Table 5-5

Source: Youth-Parent Socialization Panel Study, 1965–1997, Youth Wave. Data provided by Inter-university Consortium for Political and Social Research, available at

Note: Dem. = Democrat; Ind. = Independent; Rep. = Republican. The highlighted cells (along the diagonal) represent those individuals who remained stable in their partisanship from one time period to the next. The off-diagonal cells represent individuals who changed their partisan identification.

Changes Across Generations

A shift in the partisan composition of the electorate owing to generational change is ordinarily a gradual one, because political attitudes, including partisanship, tend to be transmitted from parents to their children. Normally, more than two-thirds of the electorate identify with their parents’ party if both parents had the same party identification. Adoption of parents’ partisanship by their children is consistent with the notion of family socialization, but it is also consistent with the notion that political views are, in part, biological in nature. Children pick up the partisanship of their parents while young, but the parents’ influence diminishes as the child comes into contact with other political and social influences during the teenage years. For most individuals, the political influence of their surroundings will be consistent with their family’s political leanings, so the similarity between parents’ and offspring’s partisanship remains strong. In contrast, people who remember their parents as having conflicting loyalties are more likely to be independents than either Democrats or Republicans. This is even more true of the children of parents without any partisan attachments. Thus, in each political generation a sizable number of voters lack an inherited party loyalty.

The Jennings study permits the empirical examination of the process of generational change because it allows a comparison of party identification for parents and their children. Figure 5-11 shows that 58 percent of the seventeen-year-olds in 1965 had adopted the party identification of their parents. Of the high school seniors, 30 percent were Democratic and came from Democratic families. Another 10 percent of the seniors were Democratic but came from independent or Republican families. Although not explicitly shown in Figure 5-11, Democrats had a somewhat higher transmission rate than either Republicans or independents. Despite the higher transmission rate, there were so many more Democratic parents that their children also contributed substantial numbers to the independent ranks.

The latest wave of the Jennings study allows an examination of generational change in a more recent time, by comparing the partisanship of the 1965 high school seniors, now parents, with the partisanship of their high school–aged children (see Figure 5-12). (Not all the original 1965 students had children of that age in 1997, so the focus is on only a subset of those in the original sample reinterviewed in 1997. Therefore, the distribution of partisanship of these parents will not be the same as for the whole 1997 sample covered in Table 5-5. The group of parents is somewhat less Democratic and more Republican than the full group.) Figure 5-12 suggests that parents transmitted their partisanship to their children at a lower rate in the 1990s than they had a generation earlier. Hidden in these numbers, however, are traces of a modest recovery in partisanship. Unlike 1965, the younger generation is only slightly more independent than the parents, and the number of children leaving their parents’ parties for independence is about equally offset by the children of independents adopting a partisanship. In 1965, twice as many children opted for independence as moved toward partisanship.

An illustration of the party identification of high school seniors and their parents in 1965.Description

Figure 5-11 Party Identification of High School Seniors and Their Parents, 1965

Source: Adapted from Paul A. Beck, Jere W. Bruner, and L. Douglas Dobson, Political Socialization (Washington, DC: American Political Science Association, 1974), 22.

Note: On the left of the figure is the distribution of the parents’ party identification, and on the right is their children’s. The numbers in the three boxes highlight the percentages of the children who had the same party identification as their parents. The numbers on the remaining arrows show various amounts of change from their parents’ partisanship by the children. For example, 7 percent of the total number of children had independent parents but became Democrats.

An illustration of the party identification of high school seniors and their parents in 1997.Description

Figure 5-12 Party Identification of High School Seniors and Their Parents, 1997

Source: Youth-Parent Socialization Panel Study, 1965–1997, Youth Wave. Data provided by Inter-university Consortium for Political and Social Research, available at


Since the 1970s, some political observers have commented on the weakness of political parties, citing especially the overall increase in independents and the appeal of independent candidates, such as Anderson in 1980 or Perot in 1992 and 1996. These factors, combined with declines in trust and confidence in government, turnout, and attention to political news, have suggested to some that the American public has lost its capacity to identify with political parties in a meaningful way. A corollary would suggest there would likely never be another realignment because political parties would not be able to attract new partisans to their camps.

A contrary point of view argues that many of these trends slowed or stopped in the late 1970s, and that partisan stability and party-loyal behavior since then have been nearly as high as in the 1950s. Analysts cite increases in party-line voting in Congress, sharper ideological division between the parties, and an increase in uncivil political discourse in the mass media and in Congress as evidence of the increased commitment to, as well as the polarization of, political parties.

So is partisanship becoming stronger or weaker? It seems to us that both these phenomena are occurring—in different parts of the electorate. On the one hand, among political elites and party activists, the polarization and hostility are becoming greater. On the other hand, a large pool of individuals remain who do not strongly identify with either of the major parties and for whom the increased intensity of the partisan debate is off-putting. More generally, in the words of political scientist Julia Azari, the United States has weak parties and strong partisans. Party organizations have little power over office holders and candidates, Azari argues, but partisanship is strong in the electorate. This combination could be dangerous when it comes to the health of political institutions.34

The close competitiveness of recent presidential elections has raised the intensity of feelings about politics. As we showed in Chapter 3, almost half the public reported trying to influence other people’s votes in 2004 and again in 2008. This is a substantial increase over percentages reported in any election in the past fifty years. Higher percentages than in previous elections reported having a strong preference for their presidential choice and caring who won the election. However, this does not seem to translate to stronger partisanship. The percentage of strong party identifiers has not increased in twenty years and is not as high today as it was in the 1950s and early 1960s. The percentage of people who call themselves “extremely liberal” or “extremely conservative” has not increased either and is generally a small fraction of the population.

A sizable segment of the electorate distrusts political parties, preferring divided government to keep either party from doing too much. The 2008 ANES found that a majority of the public (51 percent) preferred to see divided control of government, though a Gallup survey just four years later saw that number drop to 23 percent. In 2016, Gallup estimated preferences for divided government at 20 percent.35

For at least some of these nonparty people, the problem with the parties is the same partisan and ideological intensity that has been increasing. They see the party elites and activists as extreme in their views, whereas they see themselves as moderate. They view party conflict in Washington as divisive and contributing to, instead of solving, the country’s problems. These are the people attracted to Trump’s and Sanders’s attacks on the party system. For such people, heightened partisan debate is unlikely to move them to embrace a political party. Becoming more engaged in political discussion, turning out to vote, and trying to influence the views of others are not unimportant aspects of the public’s behavior, and they may signal changes in the partisan feelings of American citizens. However, the largest changes in partisan behavior are among leaders and political activists.

Study Questions

1. What is party identification, and how has Americans’ attachment to their party changed over time?

2. What are some of the key characteristics of affective polarization?

3. How do different ways of categorizing independents influence how we understand the role they play in American elections?

4. What are some key characteristics of realignment?

Suggested Readings

Blum, Rachel M. How the Tea Party Captured the GOP: Insurgent Factions in American Politics. Chicago: University of Chicago Press, 2020. A book that chronicles how the Tea Party insurgents in the Republican Party increased their political influence in the two-party system.

Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. The American Voter. New York: Wiley, 1960. The classic study of public opinion and voting behavior in the United States.

Green, Donald, Bradley Palmquist, and Eric Schickler. Partisan Hearts and Minds: Political Parties and the Social Identities of Voters. New Haven, CT: Yale University Press, 2004. A strong argument for party identification as a social-psychological orientation and a powerful determinant of vote choice and political attitudes.

Klar, Samara, and Yanna Krupnikov. Independent Politics: How American Disdain for Parties Leads to Political Inaction. New York: Cambridge University Press, 2016. Unique account of how some citizens have become embarrassed by their partisan attachments and started to identify as independents; Klar and Krupnikov then show the consequences of independent identification on political action in the United States.

Lavine, Howard G., Christopher D. Johnston, and Marco R. Steenbergen. The Ambivalent Partisan: How Critical Loyalty Promotes Democracy. New York: Oxford University Press, 2012. A fascinating analysis of partisans’ willingness to view the world accurately and pay attention to issues and how this contributes to good democratic citizenship.

Lewis-Beck, Michael S., William G. Jacoby, Helmut Norpoth, and Herbert F. Weisberg. The American Voter Revisited. Ann Arbor: University of Michigan Press, 2008. A rich reanalysis of the themes from the classic work using mainly 2000 and 2004 data.

Mason, Lilliana. Uncivil Agreement: How Politics Became Our Identity. Chicago: University of Chicago Press, 2018. A careful and far-reaching study that highlights the rise of social polarization in American party politics.

Noel, Hans. Political Ideologies and Political Parties in America. New York: Cambridge University Press, 2013. A theoretically and empirically innovative study that shows how the growth of modern liberalism and conservatism developed independent of the Republican and Democratic Parties—and then restructured the parties in their image.

Wolbrecht, Christina. The Politics of Women’s Rights: Parties, Positions and Change. Princeton, NJ: Princeton University Press, 2000. A detailed and persuasive analysis showing that the parties were transformed by the women’s rights debate in American public life.


1. Lilliana Mason, Uncivil Agreement: How Politics Became Our Identity (Chicago: University of Chicago Press, 2018).

2. Chris Wells, Katherine J. Cramer, Michael W. Wagner, German Alvarez, Lewis A. Friedland, Dhavan V. Shah, Leticia Bode, Stephanie Edgerly, Itay Gabay, and Charles Franklin, “When We Stop Talking Politics: The Maintenance and Closing of Conversation in Contentious Times,” Journal of Communication 67 (2017): 131–157.

3. E. E. Schattschneider, Party Government (New York: Farrar & Rinehart, 1942), 1.

4. The most important research cited on party identification is in Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes, The American Voter (New York: Wiley, 1960), 120–167. For an updated treatment, see Michael S. Lewis-Beck, William G. Jacoby, Helmut Norpoth, and Herbert F. Weisberg, The American Voter Revisited (Ann Arbor: University of Michigan Press, 2008), Chapters 6 and 7.

5. Michael W. Wagner, “The Utility of Staying on Message: Competing Partisan Frames and Public Awareness of Elite Differences on Issues,” The Forum 5 (2007): 1–18; Michael W. Wagner and Michael W. Gruszczynski, “When Framing Matters: How Partisan and Journalistic Frames Affect Public Opinion and Party Identification,” Journalism and Communication Monographs 18, no. 1 (2016): 5–48. DOI: 10.1177/1522637915623965.

6. For the most important statement of these ideas, see Philip E. Converse, “The Concept of a Normal Vote,” in Elections and the Political Order, eds. Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes (New York: Wiley, 1966), 9–39.

7. This discussion and data presentation are based on earlier work by the authors of the first twelve editions of this book, William H. Flanigan and Nancy H. Zingale, “The Measurement of Electoral Change,” Political Methodology 1 (Summer 1974): 49–82.

8. Alan I. Abramowitz, “Partisan Nation: The Rise of Affective Partisanship in the American Electorate,” in The State of the Parties: The Changing Role of Contemporary American Parties, eds. John C. Green, Daniel J. Coffey, and David B. Cohen (Lanham, MD: Rowman and Littlefield, 2014).

9. Shanto Iyengar, Gaurav Sood, and Yphtach Lelkes, “Affect, Not Ideology: A Social Identity Perspective on Polarization,” Public Opinion Quarterly 76, no. 3 (2012): 405–431.

10. “Presidential Election, 2020,” Ballotpedia, accessed July 4, 2021,,_2020.

11. Larry Bartels argues that partisan loyalty has been steadily increasing since its nadir in 1972, but this trend has been overlooked by analysts focusing on the weakness of the political parties. See Larry Bartels, “Partisanship and Voting Behavior, 1952–1996,” American Journal of Political Science 44 (January 2000): 35–50.

12. Paul R. Abramson and John H. Aldrich, “The Decline of Electoral Participation in America,” American Political Science Review 76 (September 1982): 502–521.

13. Kevin K. Banda, “Issue Ownership Cues and Candidate Support,” Party Politics 27, no. 3 (2019): 552–564.

14. Had Bush adviser Karl Rove not implemented his strategy to turn out the vote among Christian conservatives in safe Republican states in the South, Bush would have again lost the popular vote nationwide, while winning in the Electoral College.

15. Arthur H. Miller and Martin P. Wattenberg, “Measuring Party Identification: Independent or No Partisan Preference?,” American Journal of Political Science 27 (February 1983): 106–121.

16. John Petrocik, “An Analysis of Intransitivities in the Index of Party Identification,” Political Methodology 1 (Summer 1974): 31–47.

17. Samara Klar and Yanna Krupnikov, Independent Politics: How American Disdain for Parties Leads to Political Inaction (New York: Cambridge University Press, 2016).

18. Nick Vivani, “Kanye West Won’t Be on the Wisconsin Ballot, Elections Commission Rules,” WSAW, August 20, 2020,

19. This and most discussions of the classification of elections are based on the work of V. O. Key and Angus Campbell. See V. O. Key, “A Theory of Critical Elections,” Journal of Politics 17 (1955): 3–18; and Angus Campbell, “A Classification of Presidential Elections,” in Elections and the Political Order, eds. Angus Campbell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes (New York: Wiley, 1966), 63–77.

20. This and the following discussion draw heavily on Jerome M. Clubb, William H. Flanigan, and Nancy H. Zingale, Partisan Realignment: Voters, Parties, and Government in American History (Boulder, CO: Westview Press, 1990). See especially Chapters 5 and 8.

21. See, for example, William N. Chambers and Walter Dean Burnham, eds., The American Party Systems: Stages of Political Development (New York: Oxford University Press, 1975).

22. This concept was developed by Seymour Martin Lipset and Stein Rokkan in their discussion of the development of the European party systems in Party Systems and Voter Alignments (New York: Free Press, 1967), 1–64.

23. Nathan P. Kalmoe, With Ballots and Bullets: Partisanship and Violence in the American Civil War (New York: Cambridge University Press, 2020).

24. At the time we submitted our book for publication, the ANES had not released data including which state the respondents lived in, so we could not update Figures 5-9 and 5-10 with 2020 data.

25. Some analysts have argued that the movement of white Southerners into the Republican Party and that of Blacks and some northern whites into the Democratic Party constitutes a realignment and should be regarded as the start of a new party system. Disagreement arises about when this realignment occurred. Some date it from the 1960s, with the start of Republican dominance in presidential voting; others view it as a Reagan realignment of the 1980s.

26. Thomas M. Carsey and Geoffrey C. Layman, “Changing Sides or Changing Minds? Party Identification and Policy Preferences in the American Electorate,” American Journal of Political Science 50, no. 2 (2006): 464–477.

27. Edward G. Carmines, Michael J. Ensley, and Michael W. Wagner, “Political Ideology in American Politics: One, Two, or None?,” The Forum 10, no. 4 (2012): 1–18.

28. Carsey and Layman, “Changing Sides or Changing Minds?”

29. Norman H. Nie, Sidney Verba, and John R. Petrocik, The Changing American Voter (Cambridge, MA: Harvard University Press, 1976), Chapter 5.

30. Robert S. Erikson and Kent L. Tedin, “The 1928–1936 Partisan Realignment: The Case for the Conversion Hypothesis,” American Political Science Review 75 (December 1981): 951–962.

31. Philip E. Converse, The Dynamics of Party Support: Cohort-Analyzing Party Identification (Beverly Hills, CA: SAGE, 1976).

32. Jon Huang, Samuel Jacoby, Michael Strickland, and K. K. Rebecca Lai, producers, “Election 2016: Exit Polls,” New York Times, November 8, 2016,

33. The parents were also reinterviewed in 1973. By 1997, many of the original high school seniors were parents of high school–aged students. The study also interviewed the children of the original sample, creating a second set of parent–child interviews to compare with the original data from 1965. See Figure 5-11.

34. Julia Azari, “Weak Parties and Strong Partisanship Are a Bad Combination,” Vox, November 3, 2016,

35. Art Swift, “In U.S., Preference for Divided Government Lowest in Fifteen Years,” Gallup, September 28, 2016,

Descriptions of Images and Figures

Back to Figure

The horizontal axis ranges from 1952 to 2020 in increments of 4. The vertical axis ranges from 0 to 70 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 0 to 120 in increments of 20. The vertical axis shows different levels of political affiliations. The approximate data from the graph are tabulated below.

Back to Figure

The data from the graph for Democrats are as follows.

The data from the graph for Republicans are as follows.

Back to Figure

The horizontal axis ranges from 1940 to 2030 in increments of 10. The vertical axis ranges from 0 to 60 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

In both the graphs, the horizontal axis ranges from 1952 to 2016 in increments of 4 and the vertical axis ranges from 0 to 40 in increments of 10. The approximate data from the graphs are tabulated below.

Back to Figure

The horizontal axis shows partisans. The vertical axis ranges from 10 to 80 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

In all six graphs, the horizontal axis ranges from 0 percent to 100 percent in increments of 10, and the vertical axis shows different issue ownerships. All data are approximate.

The data from the first graph are tabulated below.

The data from the second graph are tabulated below.

The data from the third graph are tabulated below.

The data from the fourth graph are tabulated below.

The data from the fifth graph are tabulated below.

The data from the sixth graph are tabulated below.

Back to Figure

The horizontal axis shows partisanship. The vertical axis ranges from 0 to 100 in increments of 10. The data from the graph are tabulated below.

Back to Figure

The horizontal axis ranges from 1952 to 2016 in increments of 4. The vertical axis ranges from 0 to 80 in increments of 10. The approximate data from the graph are tabulated below.

Back to Figure

N equals 1,036. The data from the illustration are tabulated below:

The amounts of change from parents’ partisanship by the children are as follows:

· Democratic to Independent: 13

· Democratic to Republican: 3

· Independent to Democratic: 7

· Independent to Republican: 5

· Republican to Democratic: 3

· Republican to Independent: 11

Back to Figure

N equals 731. The data from the illustration are tabulated below:

The amounts of change from parents’ partisanship by the children are as follows:

· Democratic to Independent: 11

· Democratic to Republican: 4

· Independent to Democratic: 9

· Independent to Republican: 11

· Republican to Democratic: 6

· Republican to Independent: 12

If you find an error or have any questions, please email us at Thank you!