Do the Democrats Have a ’God Problem’?
by Gregory A. Smith, Pew Forum on Religion & Public Life and Peyton M. Craighill, Pew Research Center for the People & the Press
In the immediate aftermath of George W. Bush’s 2004 victory over John Kerry, many journalists and other political observers declared the election to have been decided, in large part, on the basis of moral values and cultural issues. According to the national exit poll, more than one-in-five voters (22%) cited moral values as the most important issue in the campaign.1 Gay marriage was identified as being a particularly important issue, since same-sex marriage initiatives were on ballots in 13 states (including the crucial swing state of Ohio) and served, ostensibly, to mobilize turnout among conservative opponents of same-sex unions. These initial reactions turned out to have overstated the importance of moral values and cultural issues–and gay marriage in particular–in the campaign. (See, for instance, Pew Research Center 2004 and others.2 ) Still, careful analyses of the national exit poll and other post-election surveys3 revealed the persistence of what Prothero4 referred to as the “God gap”; those who attend church frequently and conservative Christians (especially white evangelical Protestants) supported Bush over Kerry by large margins. In other words, insofar as religion has become a political force, it seems generally to help Republicans and hurt Democrats.
But what, precisely, is the nature of the Democrats’ struggle with religion? At least two alternative conceptualizations of their problem present themselves. The first, suggested by the work of those who see a “culture war” being waged in American politics, is that Democrats, because of their liberal approach to social issues, have surrendered the vote of conservative, traditional Christians to the Republican Party, allowing the GOP to forge electoral victories and capture control, at least for now, of both the executive and legislative branches of the federal government. Abramowitz and Saunders, for example, show that among white voters in 2004 two religious variables, church attendance and born-again status, were “more strongly correlated with (Republican) party identification and presidential candidate choice [for Bush] than any other social characteristic including income, education, gender, marital status, and union membership.”5 Guth and colleagues conclude that “Evangelicals have become the religious mainstay of the GOP,” while Democrats have “established a strong link to the growing coterie of unaffiliated voters, perhaps building a large constituency of ‘secular warriors.'”6 For Democrats, the solution to the problem conceptualized in this way is simple in principle but likely to be quite difficult in practice: find a way to (re)capture at least some of the religious voters who have come to support the GOP.
But there is a second, less frequently discussed alternative conceptualization that may help to explain the Democrats’ recent struggles with religion. The root of the Democratic Party’s religious woes may not lie solely in the fact that they have been spurned by religious conservatives. Contributing to the Party’s electoral misfortunes may be that even moderate citizens and voters view the Democrats as being unfriendly toward religion. The Democrats may have had a problem in recent elections, in other words, not just because their policy prescriptions were rejected by conservative Christians but also because a large portion of the electorate who may be sympathetic to Democratic positions was nevertheless turned off by a perceived hostility on the part of the Party toward religion, which most Americans see as a positive force with an important public role to fill.
This possibility was suggested to us by the results of a poll conducted by the Pew Forum on Religion & Public Life and the Pew Research Center for the People & the Press during July 2005. The poll included the following question: “Do you feel that the Democratic Party is generally friendly toward religion, neutral toward religion, or unfriendly toward religion?” In 2005, a mere 29% of the public indicated that the Democratic Party was generally friendly toward religion, a noticeable decline of 11 percentage points since August 2004, when 40% viewed the Democrats as friendly toward religion.
Of course, it is difficult to determine exactly what individuals who describe the Democrats as unfriendly toward religion have in mind. But the fact that Republicans are perceived as friendly toward religion by a much larger proportion of the public (55%) than are Democrats, combined with the sharp decline of the Democrats’ score on this measure after the 2004 election, suggests that responses to this question are indicative of concrete attitudes on the part of the public.
In short, the trend shown in Table 1 suggests that perceived hostility toward religion on the part of the Democrats may be an increasing problem for the Party. This is, in part, a testable proposition that we explore in the remainder of this piece.
Analyzing the Democrats’ Problem
To start with we need to ask why viewing the Democratic Party as unfriendly toward religion might be negatively correlated with, for instance, views of the Democratic Party generally or casting a vote for John Kerry.
The answer lies in American attitudes toward religion and religion’s role in public life. Though there is much that divides Americans across religious lines, it is also true that there is much about religion on which Americans largely agree. In July 2005, for instance, a Pew poll found that upwards of 95% of Americans continue to express belief in a supreme being, with 85% expressing belief in God and another 11% saying they do not believe in God but do believe in a universal spirit or higher power (see Table 2).
Just as they are largely unified in their belief in a supreme being, so Americans are also largely united in their belief that religion should inform public life and political leaders. In August 2004, for instance, when a Pew poll asked Americans whether it is important for a president to have strong religious beliefs, seven-in-ten agreed that this is an important trait for presidents to have. Table 3 indicates, furthermore, that while more than nine-in-ten white evangelicals agreed that presidents should have strong religious beliefs, 72% of white mainline Protestants and 68% of white, non-Hispanic Catholics also desire religiously grounded leaders. Only among seculars did fewer than half agree that presidents should have strong religious beliefs.
Additionally, when asked whether there has been too much, too little, or the right amount of expression of religious faith by political leaders, the majority of the public indicates that there has been either the right amount (27%) or too little (39%) religious expression by political leaders. Once again, as shown in Table 4, this holds true for majorities of all religious groups, and is even true of a substantial minority (43%) of seculars.
Finally, Table 5 indicates that Americans are also comfortable with a role for religion in public schools. Pew polling finds, for instance, that nearly two-in-three Americans favor teaching creationism along with evolution in public schools. This includes 67% of white evangelicals as well as a majority of seculars.
Testing Two Hypotheses
Americans, then, are clearly comfortable with a public role for religion and desire a politics and political leaders that are informed by religious values. This suggests that to the extent that the Democratic Party is viewed as unfriendly toward religion, its electoral prospects may be damaged. The Pew survey from July 2005 that asked whether or not Democrats are friendly toward religion also asked respondents to provide a favorability rating for the Democratic Party as a whole and about their voting behavior in the 2004 election. It is thus possible to test the following two hypotheses:
Individuals who view the Democrats as friendly toward religion will have more favorable views of the Democratic Party than will individuals who view the Democrats as neutral or unfriendly toward religion.
Hypothesis 2: Individuals who view the Democrats as friendly toward religion will be more likely to have voted for Kerry over Bush than will individuals who view the Democrats as neutral or unfriendly toward religion
Table 6 summarizes the relationship between perceptions of Democrats’ friendliness toward religion (or lack thereof) and overall views of the Party, and offers preliminary support for Hypothesis 1.
Among the 20% of the population that views the Democratic Party as unfriendly toward religion, a mere 22% express favorable views of the Party as a whole. Among the 51% of the public who say the Democrats are neutral or express no opinion as to the Democrats’ friendliness toward religion, more than twice as many (50%) hold favorable views of the Democratic Party. And, as predicted, among the 29% of the public who view Democrats as friendly toward religion, a still greater proportion (66%) views the Party favorably.
Of course, it might be objected that looking at the relationship between Democrats’ perceived friendliness toward religion and views of the Party among the public as a whole could obscure the true dynamics of the relationship. For example, it may be that evangelical Protestants (who make up the core of the culture warriors on the religious right) tend both to view the Democrats as unfriendly toward religion and unfavorably as a whole. In other words, antipathy toward the Party generally could lead its detractors to judge it also hostile to religion rather than the other way around. At the other end of the religiosity spectrum, seculars too pose a complication in analyzing the strength of the positive correlation between Democrats’ perceived friendliness toward religion and favorable views of the Party, since seculars might plausibly be expected to feel more favorably toward a party that is less — rather than more — friendly toward religion. And, of course, it must be acknowledged that “friendliness” itself is a somewhat amorphous concept.
Table 7 attempts to address these potential objections by excluding both seculars and white evangelical Protestants from the analysis. It shows that even among the centrists who remain, the relationship between perceptions of the Democrats’ approach toward religion and views of the Party holds strong; those who view the Democrats as friendly toward religion are more than two-and-a-half times as likely to hold favorable views of the Party as those who view the Democrats as unfriendly toward religion.
A similar pattern emerges upon examining the relationship between perceptions of Democratic friendliness toward religion and casting a vote for John Kerry (over George Bush). Table 8, for instance, indicates that among those who view the Democrats as unfriendly toward religion, Kerry received a mere 11% of the two-party vote. Among those who viewed the Democrats as friendly toward religion, however, Kerry’s proportion of the two-party vote was more than five times as large (61%).
Table 9, which again excludes white evangelicals and seculars from the analysis, demonstrates that this relationship holds even among centrists. Kerry’s share of the two-party vote was more than four times as large among those who viewed the Democrats as friendly toward religion as among those who perceive the Party as unfriendly toward religion.
The preceding analyses, then, provide relatively strong support for both hypotheses proposed here. Those who view the Democrats as friendly toward religion are both much more favorably disposed toward the Party and much more likely to have voted for Kerry over Bush than are those who view the Party as unfriendly toward religion. These relationships hold true even when evangelical Protestants and seculars are excluded from the analyses.
But providing assurance that perceptions of Democrats’ friendliness toward religion are independently related to views of the Party and voting for Kerry requires simultaneously considering the role of other potential explanations for views of the Party and vote choice. It could be, for instance, that the apparent relationship between views of the Democrats’ friendliness toward religion and views of the Party as a whole is simply a byproduct of the fact that liberals and Democrats, for instance, may be both more likely than others to see the Party as friendly toward religion and more likely to view the Party favorably as a whole.
Accordingly, as a final test of Hypothesis 1, we conducted a statistical analysis that simultaneously considered both the impact of Democrats’ perceived friendliness toward religion and the impact of a number of other factors on views of the Party as a whole. Specifically, in addition to perceptions of Democratic friendliness toward religion, we simultaneously analyzed the impact on views of the Party of several religious factors (religious affiliation, frequency of attendance at religious services and belief that the Bible should be interpreted literally), political factors (party identification and political ideology) and demographic factors (sex, education, race, income, region, marital status and union membership).7
Table 10 presents the results of this analysis, which provide strong support for our hypothesis. Controlling for a variety of other factors, one’s view of the Democrats’ friendliness toward religion is positively correlated with favorable views of the Party, and the impact of those views is quite large. Indeed, Table 10 shows that except for party identification, views of the Democrats’ friendliness toward religion are more strongly related to favorable attitudes toward the Party than any other factor we considered. Indeed, the strength of the correlation is approximately three times that of frequency of religious attendance, and more than three times that of political ideology.
Perhaps of even greater interest is the question of the effect of views of Democrats’ friendliness toward religion on vote choice in the 2004 presidential election. Using a slightly different statistical technique, we analyzed the impact of views of Democrats’ friendliness toward religion on vote choice while simultaneously considering the impact of the other factors described above.8 Again, the findings provided strong support for Hypothesis 2. Even after controlling for a host of factors that are also correlated with vote choice, views of the Democrats’ friendliness toward religion are strong predictors of voting behavior.
A sense of the findings from this analysis is provided by Chart 1, which illustrates the relationship between a number of political and demographic factors and vote choice. As indicated earlier, among those who view the Democrats as unfriendly toward religion, only 11% reported voting for Kerry; 61% of those who view the Democrats as religion friendly, by contrast, voted for Kerry, yielding a 50 percentage point gap.
The electoral impact of views of Democrats’ friendliness toward religion is far weaker than the impact of party identification, where there is an 81 percentage point gap between Democrats (86% of whom voted for Kerry) and Republicans (only 5% of whom crossed party lines to support Kerry). Political ideology is also more powerful, with a 61 percentage point gap between the number of liberals who voted for Kerry (83%) and the number of conservatives who voted Democratic (22%).
But the correlation between views of the Democrats’ friendliness toward religion and vote choice rivals the magnitude of the impact of a number of other important factors known to be correlated with vote choice, including gender and frequency of church attendance. In sum, though it must be remembered that “friendliness toward religion” is not a well-defined concept, it does appear that one’s view of whether or not the Democrats are friendly toward religion is a fairly powerful predictor of presidential vote choice.
To the extent that the Democrats’ recent electoral shortcomings can be explained, in part, by their struggles with religion, this analysis demonstrates that this struggle is at least two-dimensional in nature. On the one hand, scholars who observe and document an ongoing culture war and the polarization of the American electorate are correct to point out that highly religious Americans, and especially white evangelical Protestants, have largely abandoned the Democratic Party, which has contributed to their electoral misfortunes. Conceptualized in this way, overcoming their struggles with religion will require the Democrats to recapture the support of at least a fraction of those who now form the core of the GOP.
But there is a more subtle, less frequently discussed dimension to the Democrats’ religion problem. As scholars who doubt the existence of a culture war point out, there remains in the United States a (very) large corps of moderate citizens and voters, and these voters truly hold the balance of power in American elections. These voters, and even a large portion of seculars, have overwhelmingly positive views of religion and desire an important public and political role for religious symbols and values. The analyses reported here suggest that even among many of these centrist citizens and voters, the Democratic Party is not seen as friendly toward religion, and these analyses show that this is strongly related to the Party’s general reputation and electoral outcomes.
This second dimension of the Democrats’ problem also suggests an alternative route to overcoming their recent struggles with religion. That is, instead of having to peel away at the conservative Christian base of the GOP, the Democrats may benefit simply from convincing centrists of their general friendliness toward religion. Attempting to convince the public of their friendliness to religion, however, may carry risks of its own for the Democrats. Our analysis indicates that among seculars, who have become one of the core constituencies of the Democratic Party, those who view the Democrats as friendly toward religion were actually less likely to have voted for Kerry than were those who view the Party as unfriendly toward religion.
Of course, none of this is to suggest that perceptions of Democrats’ friendliness to religion are the new linchpin of American politics or the single key to understanding electoral outcomes. But in a nation where the electorate is as closely divided as the American electorate has been in recent years, any one of a number of factors could, conceivably, serve to tip the balance in one direction or another. Perceptions of the Democrats’ friendliness toward religion may be one such factor.
A longer version of this article was presented at the 61st Annual Conference of the American Association for Public Opinion Research, May 18-21, 2006, Montreal, Canada.
1Pomper, Gerald M. 2005. “The Presidential Election: The Ills of American Politics After 9/11.” In Michael Nelson, ed., The Elections of 2004. Washington, D.C.: CQPress, p. 56
2Pew Research Center for the People & the Press. 2004. “Voters Liked Campaign 2004, But Too Much Mudslinging – Moral Values: How Important?” ; Freedman, Paul. 2004. “The Gay Marriage Myth: Terrorism, Not Values, Drove Bush’s Re-election.” Slate, Nov. 5; But also see Campbell, David E. and J. Quin Monson. 2005. The Religion Card: Evangelicals, Catholics, and Gay Marriage in the 2004 Presidential Election. Paper presented at the Annual Meeting of the American Political Science Association, Sept. 1-4, Washington, DC.
3Guth, James L., Lyjman A. Kellstedt, Corwin E. Smidt and John C. Green. 2005. Religious Mobilization in the 2004 Presidential Election. Paper presented at the Annual Meeting of the American Political Science Association, Sept. 1-4, 2005. Galston, William A. and Elaine C. Kamarck. 2005. “The Politics of Polarization.” A Third Way Report, October.
4Prothero, Stephen. 2004. “Democrats: Get Religion!” Boston Globe, November 10.
5Abramowitz, Alan and Kyle Saunders. 2005. “Why Can’t We All Just Get Along? The Reality of a Polarized America.” The Forum: A Journal of Applied Research in Contemporary Politics. Vol. 3: No. 2. p. 15.
6Guth, James L., Lyman A. Kellstedt, Corwin E. Smidt, and John C. Green. 2005. Religious Mobilization in the 2004 Presidential Election. Paper presented at the Annual Meeting of the American Political Science Association, September 1-4, Washington, DC. pp. 15-16.
7We estimated a linear regression model to predict favorable views of the Democratic Party. For more details and complete results of the model, contact the authors at firstname.lastname@example.org.
8We estimated a logistic regression model to predict the likelihood of having voted for Kerry. For more details and complete results of the model, contact the authors at email@example.com.