Thomas Carsey is the Pearsall distinguished professor of political science and director of the Odum Institute for Research in Social Science at the University of North Carolina at Chapel Hill. Geoffrey Layman is professor and director of graduate studies in political science at the University of Notre Dame.

In our prior post, we argued that the last 40 years of American politics has been characterized by what we call conflict extension, which we define as the growth of party polarization across all of the major issue agendas in domestic politics. Here, we discuss a key contributor to conflict extension: identification with and commitment to political parties.

There are many other explanations of growing party polarization among party activists and citizens at large. Some focus on the emergence of new political issues, others point to the increasing association between ideology and party affiliation, some to deeper racial and religions divisions and others note strategic efforts by political candidates to draw new groups of issue advocates into party politics. Some assert that sorting rather than polarization better characterizes citizens’ changing political views. While all have some merit, they all concentrate on how changing political winds might reshape party identification and commitment. They do not view partisanship and party loyalty themselves as the winds of change. We do.

Party loyalty drives politics in part because parties help determine “what goes with what” in politics. For example, opposition to tax increases and opposition to abortion go together as “conservative” positions in our current politics because the Republican Party put them together. Of course, the Democratic Party has done so for the opposite sides of these issues as well.

More importantly, decades of research on citizens’ political attitudes behavior suggests that many, perhaps most, people feel an emotional connection to and social identity with their political party, causing them to view the political world through party-colored lenses. When parties stake out positions on new issues or move to more extreme positions on old issues, party identifiers tend to gravitate toward those positions themselves — and this, we have shown, is even true of party activists.

Recent trends in public opinion on NSA monitoring of citizen communications provide a good example. The Pew Research Center for the People and the Press reported on June 10, 2013 that the percentage of Democratic identifiers who found NSA surveillance programs acceptable increased from 37 percent in January 2006 to 64 percent in June 2013. In contrast, the percentage of Republican identifiers saying these programs were acceptable decreased from 75 percent to 52 percent over this same time period. We doubt these changes emerged from a large influx of anti-surveillance advocates into the GOP or of pro-surveillance supporters into Democratic ranks between 2006 and 2013. Rather, the shift likely occurred because we had a Republican president in 2006 and a Democratic president in 2013, and many people simply adjusted their views on NSA activities to fit with their prior partisan attachments.

Such changes have driven party conflict extension more generally among both activists and the public. The figure below uses panel data from the 2000 Convention Delegate Study, focusing on national convention delegates who responded to surveys in both 1992 and 2000 and were active in the presidential campaigns in both years. It shows the difference between the average position of continuing Republican and Democratic activists (on scales ranging from 0 for the most liberal position to 1 for the most conservative position) on six issues: abortion, levels of government services and spending, government providing health insurance, spending on welfare programs, spending on programs to help African Americans, and spending on national defense.

Party polarization among convention delegates grew from 1992 to 2000 on every issue. Activists committed enough to their parties to stay involved changed their views on issues to reflect the growing divide between the two parties. This is striking when you consider that party activists are among those with the most strongly held views on political issues. Given that a large majority of convention delegates in 2000 were also delegates in 1992, the majority of the increase in overall polarization observed between 1992 and 2000 can be traced to those repeat delegates adjusting their views rather than the mobilization of first-time delegates to the 2000 convention. This means that commitment to the party itself plays a major role in producing and maintaining party polarization among party activists.


The next figure uses data from the 2008-2009 American National Election Study Panel Study. We examine the positions held by citizens who were interviewed in both January and October of 2008 and who identified with the same party at both times. We show the difference between the average positions (on scales ranging from 1 for the most liberal position to 7 for the most conservative position) of Democratic and Republican identifiers on two of the key issues of the campaign: whether government should pay for medical care for all citizens and whether taxes should be raised on incomes over $200,000 per year. Among all partisans, but especially among those reporting a strong identification with their party, the degree of party polarization over this short time period increased.


The next figure (just below) uses data from the 2012 American National Election Study to show that party identification and awareness of party differences combine to produce a higher correlation between the views citizens hold on abortion and social welfare issues. (We measure social welfare issues by combining respondents’ views on whether government has the responsibility to ensure that citizens have jobs and good standards of living and whether government should provide health insurance.) We divide respondents between Independents, Weak Party Identifiers, and Strong Party Identifiers. We further distinguish between those who are aware that the the Republican Party is more conservative than the Democratic Party on both issues, those who recognize this for only one issue, and those who are not aware of this for either issue.

The third figure above shows that when people do not recognize party differences on both issues, there is essentially no connection between their attitudes on abortion and social welfare programs (the blue and red bars). However, people who realize that the GOP is to the right of the Democrats on both abortion and social welfare (the green bars) adopt much more consistent views on the two types of concerns, but only if they identify with a particular party, and especially if they do so strongly. This pattern conforms to our previous findings using panel surveys.

For decades, many political scientists have pointed out how party loyalty shapes the views held by Democrats and Republicans. However, we believe that the role played by party identification and commitment in producing party polarization has been underemphasized. A large part of why Democrats and Republicans hold increasingly divergent issue positions is simply because they are Democrats and Republicans committed to their parties. Parties signal to voters where they stand on issues, and those who identify with and are committed to a party view those signals through partisan colored lenses and adopt the positions of their partisan “teams.” For those hoping to reduce party polarization, ours is not a terribly optimistic view—it may be more difficult to reduce party polarization if attachment to a party is a core cause. However, we think our view is a realistic one.

Further Reading

This is the latest post in our ongoing series on political polarization. The previous posts are listed below.  -Dan Hopkins