A seminar held at the Ethics and Public Policy Center in late January featured two well-known commentators on American religion and public life. They made informal remarks and then took part in a general discussion. Moderator Michael Cromartie is vice president of the Center.
JOHN C. GREEN
Michael Cromartie: John Green is the director of the Ray C. Bliss Institute for Applied Politics at the University of Akron and a professor of political science there. He is the author or editor of about ten books. He and his colleagues did a study for us—funded by the Pew Charitable Trusts—to see what influence religion had upon voting behavior in the 2000 presidential election. John will report on his findings, and John DiIulio will comment. Following that we will have a general discussion of the survey results.
John C. Green: It is a pleasure to be here and to meet some of you whom I have spoken to on the phone. I enjoy speaking to journalists, mostly because I learn a lot from you—probably more than you learn from me.
The results I will report are from the third of three surveys we have done funded by the Pew Charitable Trusts in 1992, 1996, and now 2000. Their purpose is to measure the influence of religion on voting behavior more precisely than surveys typically do. Since we’ve done these surveys the same way three times, we can compare the behavior of religious groups over time with some confidence. In the spring of the election year, we take a large sample—this time it was a random sample of 4,004 adult Americans—and ask them a large number of religious questions. Then we go back after the election to ask these people how they voted. The post-election surveys tend to be fairly brief. Although we try to contact all the people we interviewed in the spring, we are able to re-survey a little more than one-half of them. In this case, we contacted 2,363. Our report covers the two-party vote only.
On the left side of Table One, we divide adult Americans into fourteen religious groups. The first two are white evangelical Protestants and white mainline Protestants, with the classifications “evangelical” and “mainline” based on denominational affiliation. We have found that denominational affiliation is very important to voting behavior. While it may not be the best way to define these groups in religious terms, in social and political terms, denominational affiliation is very important. Each of these large categories is broken into two parts, more observant and less observant based on church attendance. The more observant are those who claim to attend worship once a week or more often. This definition puts some people who attend fairly frequently in the “less observant” category, along with some who by their own admission rarely or never attend. We separate out black Protestants; we believe that the black Protestant churches represent a separate religious tradition. And politically, as we’ll see, black Protestants are very different from white Protestants. We also separate out Hispanic Protestants and Hispanic Catholics; although these groups are not really separate religious traditions, they differ politically from their non-Hispanic co-religionists. Roman Catholics are also divided into more observant and less observant. Then there are some smaller groups: we broke out Mormons, “other Christians,” Jews, and “other non-Christians.” Finally, there is a large category we call “secular” voters. These persons lack a religious affiliation but may have religious or spiritual values.
We measure religious affiliation as precisely as we can, and try hard to reduce what are called “social desirability” effects. Many people think it’s good to be a member of a religious group and may claim an affiliation even if they don’t really have one. Church attendance tends to be similarly over-reported. The result of our method of measuring religious affiliation is that we typically come up with fewer mainline Protestants and fewer Roman Catholics than other surveys. For a fair number of people in those traditions, the affiliation is purely nominal. Under our measures, these people end up in the “secular” category. This approach produces better relationships between religion and politics. Of course, surveys are statistical artifacts. These are best thought of as estimates of how these various groups behave. They are not written in stone.
Our numbers have an additional problem. We broke out some very small groups, such as Mormons and Jews, just because we thought they were interesting. But these small groups may include only a few dozen people, so one needs to view such figures with caution. Some of these smaller groups show such distinctive voting patterns that we suspect we would get very similar results if we surveyed hundreds of their members.
Table One shows that 84 per cent of observant white evangelical Protestants voted for Bush. This is a striking increase from our 1996 survey, when this group voted 70 per cent for Bob Dole. Less observant white evangelical Protestants did not vote as strongly Republican, just 55 per cent. But in our 1996 survey, Bill Clinton and Bob Dole broke even in this group, and some other people’s surveys showed Clinton winning this group. So less observant white evangelicals returned somewhat to the Republican fold. There are similar patterns among white mainline Protestants: 66 per cent for Bush, up from 58 per cent for Dole in 1996.
With Roman Catholics, a pattern continues that had been developing all through the 1990s: more observant Roman Catholics voted more Republican, less observant more Democratic. The Catholic community is indeed divided in political terms.
In contrast to their white counterparts, black Protestants voted overwhelmingly for Gore: 96 per cent. Of course, African Americans generally vote Democratic, but the black Protestant church is the strongest Democratic component of the African American community. Hispanics voted for Gore as well, particularly Hispanic Catholics, at 76 per cent. Hispanic Protestants were a little more divided but still gave Gore 67 per cent of their votes. With blacks and Hispanics, there was not much of a percentage change from 1996, but turnout increased.
We broke out the Mormons because they’re so solidly Republican. In most surveys, they turn out to be the most Republican of all religious groups, more so than observant white evangelicals. Some other smaller groupings are broken out in the survey also, and they all voted Democratic. Finally, there is the “secular” bloc, which voted—as in 1996—about two-thirds Democratic.
Table Two is more interesting from a political point of view. Here we show each of these religious groupings as a percentage of the Bush vote and the Gore vote. This gives us a sense of the relative weight of these groups in the presidential coalitions of the two major parties. More observant white evangelical Protestants were clearly the dominant religious category in the Republican coalition. In our numbers, almost a third of the Republican votes came from this group. That’s up from 1996 and has been growing all through the 1990s and indeed since the 1980s. If one adds in the less observant, who are less Republican and less numerous, then a full two-fifths of Bush’s votes came from white evangelical Protestants. Mainline Protestants also contributed to the Bush column, but they’re distinctly junior partners in the Republican Party. Not very long ago they would have been the distinctly senior partners. Roman Catholics also contributed; all together, about one-fifth of the Republican votes came from Catholics. The other groups didn’t contribute a whole lot to the Republican column.
The Gore coalition was more diverse, as Democratic coalitions tend to be. The largest group was black Protestants, who contributed almost one-fifth of Gore’s votes. But notice that the secular category contributed about the same, almost one-fifth. Roman Catholics, interestingly enough, provided about the same proportion of the Bush vote as the Gore vote, a little over one-fifth. Other minority groups tended to support the Democratic coalition, and if you add them together they are a fairly significant bloc. The Democrats had some problems with white Protestants: in our survey, the Gore vote among white evangelicals was the lowest proportion of that group that the Democrats received in the 1990s, lower than Clinton in 1996 and in 1992. This year Democrats had some similar problems with white mainline Protestants.
Thus the survey shows some old patterns and some new patterns. Religious traditions do indeed matter in politics, very much as we’ve come to expect. What we see that’s new in 2000 is a polarization of some key groups. Observant white evangelical Protestants are voting more Republican than they have in the past, balancing out the strong Democratic vote among black Protestants and Hispanics.
JOHN DiIULIO
Michael Cromartie: Thank you, John. We’ll have some comments now by John DiIulio, the Frederick Fox Leadership Professor of Politics, Religion, and Civil Society at the University of Pennsylvania. He previously taught at Princeton, in the department of government and politics. [A few days after this seminar, on January 29, John DiIulio was appointed director of the new White House Office of Faith-Based and Community Initiatives.]
John DiIulio: Political scientists in general did not fare very well in the 2000 campaign. We were the big losers. Our very econometrically sophisticated “it’s the economy, stupid!” models had generally predicted a walk-away Gore victory, not the fifty-fifty split that we got, and so I am especially glad to have a fellow member of the tribe doing a very interesting and cogent analysis that may help to redeem our credibility a bit.
This is a very good piece of survey research, using 2,363 randomly selected adults. I agree that the data analysis by Professor Green and his colleagues suggests what they say it suggests. Nonetheless, it’s always important to remember what statistical analysis of this kind is: it’s about drawing inferences about large populations from small samples. The inferences are only as good as the samples, and the samples, even when you have a nice big group, are not always perfect. A second thing to remember about statistical analysis is that it has three rules: disaggregate, disaggregate, disaggregate. John has done a very good job of disaggregating these data denominationally. The most interesting finding, I think, is that the more religiously observant people voted for Bush and the less observant voted for Gore.
There are three slightly contrarian points that I would like to make. First, I think that peeking out from behind certain of the data here, from behind the religious variable in the analysis, is a race variable. Research by my University of Pennsylvania colleagues Ram Cnaan and Byron Johnson and by others who have studied the contemporary life of black congregations suggests that there is probably as much variance in patterns of religious observancy among black Protestants as among whites. It would appear, therefore, that almost all self-identified Protestant blacks, whether they are weekly church goers, twice-weekly goers, or not-very-often goers, voted for Gore. From this I think we can conclude that race, not just religion or degree of religious observancy, was at least in part predictive of the vote. I don’t mean to imply that Professor Green suggested otherwise. Both he and I will explain to anyone who cares to hear that after a hundred years of this kind of research by social scientists, we can agree on the fact that it’s a multivariate world.
Second, I would agree wholeheartedly that religion was unusually prominent in the 2000 election. Three of the four top candidates seemed unable to remain mum on the Almighty. Only Mr. Cheney was able to do that. The first candidate to give a powerfully faith-friendly speech was Vice President Gore in May 1999, when he called for an expansion of the “charitable choice” provision of the 1996 welfare-reform law. Around that time, he talked repeatedly about religion and about his own faith as a born-again Christian. I thought he seemed fairly sincere, and he said it quite often, at least as often as Bush.
Why then were more religiously observant people in general, and Gore’s fellow born-again Protestants in particular, so much more inclined to vote for Bush? For the same reason that blacks, despite Bush’s appeal to compassionate conservatism, were inclined to vote for Gore: namely, that party identification, independent of both religion and race, continues to be a major determinant of voting behavior even in this age of independent voters, fewer yellow-dog Democrats, fewer rock-ribbed Republicans, and lots of split-ticket voters. Blacks are disproportionately registered Democrats, while highly religious whites tend to be disproportionately Republican. So while religion was certainly an important factor, there are other mediating variables here, especially of race and party ID, but also such things as ideology, positions on issues, and indeed the state of the economy. I would be very interested to learn whether the data on congressional voting would show a similar pattern of religion—mediated or not by these other variables. I suspect that it would, though not quite as strongly. That’s the nice thing about an empirical question: you can get data and find out.
Finally, Professor Green’s paper reminded me of a question I had long had about the electoral significance of groups like the Christian Coalition. While the media have often written about such groups as if they were a huge factor in mobilizing white evangelical turnout and deciding elections, I’ve never actually seen any good evidence that these groups did more than essentially increase the winning margins of essentially conservative Republican candidates who, given the usual turnout in voting, would have won anyway. So I don’t know the answer to the question, Would the white evangelical vote be as one-sided and cohesive as it now appears to be were it not for the Christian Coalition and other such groups? I suspect that it probably would be, but I do not know.
The first interesting question about elections is of course who won and who lost, but what most interests us academics is the broad voting trends and what they imply about the attitudes of voters, the operation of the electoral system, the fate of the political parties, and the direction of public policy. Just as we’re beginning to recognize that religion has a much stronger influence on volunteering behavior than we had realized before, so we’re also beginning to understand better that religion has a powerful pull on voting behavior. I think that Professor Green’s analysis gets us much further down that path of understanding.
DISCUSSION
Michael Cromartie: Thank you, John. Now everyone else is invited to join the conversation. [These other participants will be identified at the end.]
Robert Shogan: John Green, could you speculate about the effect of the third or maybe the fifth man in the campaign, Bill Clinton?
John Green: In looking at the other items in this survey, we saw clearly that the country had ambivalent views toward Clinton. A lot of other survey research has demonstrated this as well. Even among groups that voted Republican there was a general appreciation of Clinton’s job performance, and even among groups that voted Democratic there was a generally negative view of his personal behavior. Among white Protestants, particularly the less observant ones, the personal problems of President Clinton seemed to be more important.
Robert Shogan: Among the less observant?
John Green: The less observant. It wasn’t that they were necessarily more troubled about Clinton’s character, but their troubles with it seemed to influence their vote. Although the more observant evangelical Protestants and even Catholics didn’t care much for the President either, their voting behavior can be explained by a lot of other things. But with these less observant people, the character issues seemed to make a difference. I think that some of the character questions about Vice President Gore seemed to make a difference also.
Richard Ostling: The polarization you described as being greater in 2000 than in 1996, was it also greater in 1996 than in 1992?
John Green: Yes, the religious polarization that we highlight does seem to have been increasing all during the 1990s, election by election, usually by just a few percentage points.
Fred Barnes: Why didn’t Gore’s appeal to religious voters work?
John Green: Many of the white religious groups that Gore was trying to mobilize had over the 1990s become increasingly Republican, and so there was a perceptual screen there. I think the Clinton problems and Gore’s association with the Clinton problems made it difficult for a lot of these religious persons to respond to Gore’s religious appeals. Although Al Gore is actually a very religious man, most people did not know that about him. Then his faith suddenly appeared on the radar screen, and it surprised a lot of people, including journalists. I remember getting a number of calls asking, “What is this all about?” People seemed to see it as something odd. In that sense, George W. Bush had an advantage; the partisan cues worked in his direction—well, at least with whites. His expression of faith didn’t seem like something new, and anyway people think that Republicans tend to be this way. Bush had similar problems with black Protestants and with Hispanic Christians. He simply was not as credible, even though he made tremendous efforts to appeal to them.
John Dilulio: That May speech by Gore was a bold speech for a Democrat, at least as bold as Governor Bush’s speech in July. In it Gore said what role he thought religion should have in public affairs. But as the Democratic convention drew nearer, there was a fight, apparently, about whether even to keep the language favoring charitable-choice expansion in the party’s official talking points. Finally they agreed to keep it, but the point was somewhat watered down. Gore had a problem in that his appeals were episodic and rather sketchy; he was never able to flesh out the implications, even when Senator Lieberman was nominated and it seemed like a perfect moment to put flesh on the bones of these sentiments. It never happened.
Larry Witham: One interesting thing that came out of the election postmortems in GOP circles was that the Republicans didn’t get as much of the white evangelical vote as they had hoped. By one reckoning, the turnout had been 19 per cent of all voters in 1996 but was only 15 per cent in 2000, and so the Republicans lost potentially six million conservative evangelical voters. There were two explanations. One was that their electoral strategy required setting priorities, so they didn’t go out to the states with a lot of white evangelicals and pump them up. But by not trying to secure those votes, they lost six million of them. The other view was that the Jim Dobsons and the Christian Coalition certainly didn’t deliver this time, but maybe they never really could.
John Green: In terms of the folks who did show up at the polls, the Republicans did well among both evangelical and mainline Protestants; it would be difficult—though possible—to do better. But I think there is some truth to both of the points you suggested. Many states where white Protestants are common were not considered competitive in 2000, so there wasn’t as much pressure to turn out the vote. Alabama didn’t matter; Texas didn’t matter. In our data, we have some evidence that turnout among white Protestants was down a little. Black Protestants were way up, but white Protestants were down a bit. There’s a strong regional component in this pattern. Florida aside, there was not a lot of competition in the South.
I think that the organizations of the Christian Right, the Christian Coalition and other groups, were quite active and reasonably effective, though not as effective as in 1996 or 1994. Even if they didn’t bring any new people to the polls, they certainly certified to their constituency that George W. Bush could be trusted. Early in the campaign, particularly in the primaries, there were some serious questions as to whether white evangelicals would stay with Bush. He seemed to be too moderate, and in fact was being advertised in many circles as very moderate. But he did manage to do very well in this voting bloc. So I think the Jim Dobsons and the Pat Robertsons, if nothing else, succeeded in certifying George W. Bush to this community. By the way, I think a similar thing happened in the Catholic community; a lot of priests and lay leaders went out of their way to indicate to their followers that Bush could be trusted.
Michael Cromartie: John, isn’t there an untold story here as to how Bush got Christian Right leaders to keep a low profile during the campaign?
John Green: I was absolutely amazed that those leaders actually worked quite hard for the Republican ticket and didn’t make the kind of headlines they have typically made that would disillusion other voters. Yes, there is a good untold story there.
Terry Mattingly: Does your survey form include specific questions about social issues, such as sexual freedom and abortion rights?
John Green: Yes, we had batteries of questions on abortion, gay rights, a whole range of social issues.
Terry Mattingly: Do you have tabulations by those issues?
John Green: Not in these results, but we’ll soon make them available.
Terry Mattingly: At the height of the Lewinsky affair, the New York Times columnist Maureen Dowd said that the country was divided into two groups, those who opposed Woodstock and those who liked it, and that this was a sexual-morality cultural divide. Did you see evidence of this in your survey?
John Green: I think there is a cultural and moral divide in the United States, though not quite as simple as Woodstock versus traditional values. The most controversial issues like abortion are in some ways the least significant part of that cultural divide. Throughout the South and the West and around middle-sized cities, there is a much more traditional view of sexual behavior, family relations, and the public role of religion than in major metropolitan areas, where there is a much more cosmopolitan view. That tension underlies to a substantial degree what we saw in the election. A very interesting survey on this issue was done by Richard Morin and Claudia Deane of the Washington Post; it documents this divide more fully than we do.
Michael Cromartie: The authors of that survey are here today. When did you do the survey, Rich and Claudia?
Richard Morin: Last year and also two years ago.
Claudia Deane: The Washington Post published it on September 30, 2000 [“Among These Voters, Values Matter—but Vary Widely”].
John DiIulio: Let me add something on the cultural-divide question.
Michael Cromartie: Were you at Woodstock, John?
John DiIulio: Nobody from south Philadelphia was at Woodstock—Democrat or Republican! Let me try to be a little bit of a contrarian. That’s a wonderful map that the survey results provide. But there’s also a state-legislatures map and a mayors map and a governors map. With eighty-plus thousand elected officials, we’ve got lots of elections in this country. My, my, how interesting things get when people who voted one way on that particular Tuesday in November seem at mid-term election two years later to vote for members of Congress whose stance on controversial moral questions looks very different. Now, I’m not saying there is not a divide—I think you would have to be more than a little out-of-it to think that there’s none—but I’m suggesting again that there are reasons to be a bit more circumspect about how deep it is and how much of a driving force it is.
Bob Jones: There was a lot of analysis of the surge for the Democrats among white suburban voters, particularly in the industrial states. Given your survey numbers, that seems a little counterintuitive, since those suburbs are strongholds of the evangelical megachurches and the white mainline Protestant churches. I’m wondering if you think the white Protestant vote would look different if you could break out suburban versus rural or urban.
John Green: Yes. There are regional and size-of-place distinctions in our data. Gore did better among all these groups in suburbs and Northeastern/Midwestern states than in the rest of the country. So it goes back to John’s point about a multivariate world. Let me add another variable: gender. There is a gender gap even among the most observant white Protestant evangelicals. I think gender was a particular factor in the suburbs of the Northeast.
Steve Wagner: The thing that struck me when I examined the exit polls was the degree to which what you call secular voters—those people who also think the country is morally on the right track—coalesced around Al Gore. Do you detect an emerging political identity among secular voters?
John Green: Yes, I do. Throughout most of the 1990s, secular voters largely voted Democratic in reaction to what was going on with religious voters in the Republican Party. Like all these other groups, secular voters are diverse in many respects. But one thing they tend to have in common is very liberal positions on sexual and family issues. In recent times, there does seem to be something of an identity developing among at least some secular voters; they really see themselves as something apart. Of course, secular voters don’t have institutions to nurture and maintain that identity.
Jody Hassett: In the last week or two, we’ve seen a split among evangelicals between the pragmatists—who are parsing Bush’s language when he talks about the fact that the country is not ready for overturning Roe v. Wade, and who are taking the long-haul, incremental view—and the purists who want him to review his predecessor’s executive orders, to get that office on faith-based organizations set up in the White House, to make things start to happen right away. Any thoughts on that?
John Green: It’s hard to tell how the voters themselves view this, but if you look at the leadership, there has always been this gradient between the pragmatists and the purists. Lots of evangelical and other religious leaders were very pragmatic in 2000, figuring that, for all his faults, Bush was better than Gore. I see this breaking down, in terms of expectations post-election, very much according to pre-existing lines. The hard-core pragmatists think that Bush ought to be given time and that only so much can be achieved. The hard-core purists never went along with the strategy—they were induced to be quiet during the campaign, but they never bought into it. Now some of the pragmatists are falling away—they’re moving toward expecting a lot more, having less realistic expectations. This presents the President and the Republicans with some interesting problems, because they are going to have to deliver on some things fairly quickly to keep the new pragmatists—not to mention the more purist groups—in line. But this is very common in American politics: people coalesce around a candidate, he wins, they divide up the spoils and become discontented.
Michael Barone: You’ve taken these surveys now in three successive elections. Which of these groups is growing larger as a percentage of the electorate and/or population, and which is growing smaller? Which is the wave of the future?
John Green: Evangelical Protestants appear to be growing. When we first did this survey in 1992, using the same kinds of measures, evangelicals made up slightly less than 25 per cent of the adult population. By 2000 they were just over 26 per cent. That’s not a huge increase, and it may just be the result of sampling problems, but since these are fairly big samples, it may represent growth. We also see continuing decline in mainline Protestants and some decline in white Roman Catholics over that period.
Michael Barone: What about seculars, the second-largest group?
John Green: They seem to be edging up as well. Again, we’re not talking about huge changes here. Of course, the Hispanic population continues to grow. There are more Hispanic Catholics, but there are also more Hispanic Protestants, something a lot of people don’t realize. We don’t see much growth in the “other non-Christian” category, among Muslims or people of Eastern faiths—a little increase, but not a dramatic one.
Michael Barone: So the two growing groups seem to be the ones that politically are among the most one-sided in their preferences: evangelical Protestants and seculars.
John Green: That’s what our figures show.
Gerard Perseghin: Why didn’t Hispanic Catholics vote like the rest of the Catholics?
John Green: I think a lot of it was ethnicity. Hispanics are largely new immigrants, and new immigrants do not identify as strongly with their co-religionists in the United States as they do with their own heritage. I suspect that the message of the Catholic Church on issues like abortion does not come through as clearly in the ethnic churches as in other congregations, while the messages on social welfare resonate more with those communities. It’s interesting that Hispanic Protestants, not just in this election but also in previous ones, are somewhat more Republican. It may be that the more conservative messages of the evangelical clergy somehow get through. But they don’t get through all that strongly, because Hispanic Protestants vote more Democratic than their non-Hispanic counterparts.
Mary Leonard: Professor DiIulio, to the extent that evangelicals have a political agenda, how do they expect it to be advanced through Bush’s proposal on faith-based action?
John DiIulio: I really don’t know. I don’t know what their expectations are. One thing I would say is that there are different communities within the evangelical community. There are liberal evangelicals, and even liberal evangelical organizations, as well as what are conventionally thought of as conservative evangelicals. I think that when then Governor Bush talked about what he intended to do, he made it pretty clear. What he said won’t satisfy all evangelicals, but there was a lot in it that should please most of them.
John Green: For many evangelicals, the expectation is pretty straightforward, I think: that religious institutions will be honored and will have an important place in public life. When you get beyond that kind of general symbolism, though, I think there are some real disagreements. A lot of leaders we talk to in the evangelical community are deeply suspicious of charitable choice; for good, historic, evangelical reasons, they don’t want entanglement between religion and government.
Michael Cromartie: For theological reasons.
John Green: Yes, theological reasons. Others are just champing at the bit; they want federal money, and they really think they can do a tremendous amount of good with it. They think they can really take care of alcoholics, for instance, and troubled youth.
Carlyle Murphy: Do any of you have details about President Bush’s faith-based plans?
Thomas Pratt: I might be able to shed a little light on that since we, Prison Fellowship, have a major project in Texas as a result of Governor Bush’s office of faith-based activity. He wanted a unit in the governor’s office that would encourage faith-based organizations of repute and capability to take on some aspects of the state’s social responsibilities, such as operating prisons and offering drug treatment. You had to meet certain requirements, of course. We run a prison outside Houston. We chose to do it without government money, so we were not beholden. But we had to pass muster with regard to constitutional issues. We had to spend time with constitutional lawyers, with the ACLU, with the unions, and so on, to get the lines straight. We’ve been operating there for about three years. We operate in two other states under the same set of rules, and in those cases the states provide dollars for the programs. It is possible to do this. You do have to watch the lines and have them set before you get started.
Karlyn Bowman: Do you think we have a deeply or closely divided electorate? I’m thinking here of the work of Alan Wolfe in which he finds that people of different stripes nonetheless have a lot in common.
John Green: It’s clear that the country is deeply divided over the choice of president, and that probably reflects disagreement about the future direction of the country. I’m not sure, though, that this reflects deep-seated social divisions. Many of these groups that in our data would appear to be warring with one another actually get along pretty well and have a lot in common. One of the great lines from Alan Wolfe’s book One Nation After All was that the culture war is “within us.” For a lot of the groups in our survey, that statement seems accurate. They have contradictory views. A few of them, notably black Protestants and more observant evangelicals, have less of a conflict. But for a lot of the other groups, the choice for president was tough because their values pull them different ways. While they may disagree with their neighbors, they don’t hate them. Most people wanted the election to go away so they could get back to regular life, which is really pretty good these days.
John DiIulio: If you look at the electorate as a whole and try to divine from the data what you might say about the degree of polarization according to race, region, religion, and so forth, the picture gets murky pretty quickly. Remember the “new progressive coalition” that Clinton brought us in 1992? In 1994 that was gone and there was a new voting bloc of “angry white males,” who in 1996 had been replaced by “soccer moms.” Now we are on the verge of “religious identity politics.” These things have a remarkable volatility in American politics. Perhaps in the year 2004, given a different set of economic and other issues and different candidates with different positions, the Catholics might turn back and become once again more reliable Reagan Democrats. What’s important is the fact that you must look at religion continually to explain political outcomes.
Joe Loconte: We hear quite often that African Americans, particularly church-going African Americans, are one of the most socially conservative population sectors. But this doesn’t seem to be reflected in these election results. Is there something going on in that community? Are they becoming less concerned with some of these social issues? Stephen Carter in his latest book accuses the African American leadership of being utterly co-opted by the Democratic Party and sacrificing its prophetic voice.
John Green: In our data, church attendance and other measures of religious commitment have essentially no effect on the voting behavior of African Americans but tremendous effect on their attitudes. Regularly attending black Protestants have more conservative views on abortion, on gay rights, on traditional morality in general, but this social conservatism is not connected to their votes. I suspect there are two reasons for this. First, there are issues of racial justice and civil rights that trump that connection. Second, African American leaders offer a very consistent message to their followers that the Democratic Party is superior to the Republican Party. In this last election, we saw that mobilization reach new heights of effectiveness. It’s interesting that back in the 1950s, most evangelical Protestants in the South voted Democratic because of class reasons, although they had the same conservative views on morality that evangelicals have today. There were other issues that connected them to the Democratic Party. I think there is a potential for conservative or Republican votes in the African American community if Republicans can get around the issue of race. I know that George W. Bush tried hard to do that, but the election results show that he was very unsuccessful.
John DiIulio: I think there is no question at all—and I’ll invoke the name of George Gallup, who has more good kinds of data on these subjects than journalists and academics have generally noticed—that African Americans are, by every measure, the most consistently religious people in America, and they hold strongly what would be considered conservative views on a whole range of social issues. Then why don’t they vote with the more conservative party? One reason, I think, is that the Republican Party is still identified as an anti-government party. Government is not a bad thing in the African American community. It’s not all good, either, but there are very few libertarians in the African American community. And here comes the interesting dynamic that we’ll watch over the next several years: candidate Bush put out a message that was consistently targeted to say what I think was in his heart about the poor children, youth, and families in urban neighborhoods. He made every effort to get that across at the Republican convention in my hometown, Philadelphia; his speech was extraordinary in that regard. Then came the 92-8 split. It’s a bad thing, and those within the Republican Party who say, “Well, let’s not think about this anymore,” are making a terrible social, moral, and political mistake. I think the African American electorate is saying, “We heard this and we didn’t believe it, but maybe now we’ll know you by your works.”
Ramesh Ponnuru: Could you give us a breakdown of age among these religious groups? How relevant is age?
John Green: In the South, younger evangelicals are becoming more Republican and older evangelicals are less Republican, reflecting generational change. Outside the South, there doesn’t seem to be much of a pattern by age.
George Weigel: Let’s discuss the candidacy of Senator Lieberman. Publicly, at least, it seems that the ancient prejudice of anti-Semitism simply doesn’t exist at a national level. Was there anything to suggest that Lieberman’s Jewishness was an issue in the campaign?
John Green: I’ve seen no poll evidence—ours or others’—that it was.
George Weigel: This would not have been the case forty years ago.
John Green: Yes, I think it is a big change. Of course, it’s possible that there is some deeper anti-Semitism that we are unaware of, and maybe in another context it might rear its ugly head. But certainly, on the face of it, there does seem to be a change. Interestingly, our evidence suggests that Joe Lieberman probably helped Al Gore with a lot of people, including conservative Christians.
Hillel Fradkin: Professor Green, you mentioned that evangelicals hope to get respect for their churches from this administration. How much of their political outlook at this point stems from this sense that they are held in disrespect and want to be vindicated in the public mind?
John Green: Within the mass public and particularly among political activists and elites, evangelicals feel severely put upon. They particularly dislike journalists, but I think college professors are only slightly behind. They see all of the cultural elites as being very hostile to them. I speak to a lot of different religious groups on these sorts of issues, and I generate the most suspicion among evangelicals. You get the sense that they’re thinking, “You all don’t respect us, you don’t understand us, and because of you, the powers of the earth don’t treat us as they should.” Now, I suspect that a lot of other religious groups feel the same way. But part of what makes it especially difficult for evangelicals is that when they talk to people outside the evangelical community, the reaction usually is disbelief. “You’re the biggest religious tradition in America! Your values dominate huge sections of our society! How can you feel discriminated against?” But nonetheless they really do. It’s a serious thing.
Hillel Fradkin: So this is “religious identity politics”? What you just said suggests that it wouldn’t necessarily be tied to the actual positions people have on sexual matters or family matters but rather to a sense of themselves as a group that wants respect like any other group.
John DiIulio: I’m trying to think of a term for this new variable, this sort of religious Rodney Dangerfield-ism—“I don’t get no respect.” Where surveys ask, “Do your views on this, that, or the other issue set you apart from most people you know, most people in your neighborhood, in your state, in your country?” the response is likely to be “no, it’s fine,” until you get to the “cultural/academic/media elite.” There aren’t many evangelicals visible in those institutions. How many Assemblies of God folks have been cabinet secretaries lately? It’s interesting to see in the coverage of Senator Ashcroft how the fact that he doesn’t dance is received quite differently from the fact that Joe Lieberman doesn’t drive on the Sabbath or somebody else doesn’t do something else because of religious scruples. It’s an odd thing. Evangelicalism is a religious tradition that causes people to say, “My goodness, isn’t that odd,” and yet 47 per cent of the American people, according to Gallup data, define themselves as born-again Christians.
John Omicinski: Does the split in the Catholic vote make those votes irrelevant, so that politicians don’t have to deal with them?
John Green: No, it makes them especially valuable, because these groups are more likely to be moved.
John Omicinski: Have they been moved in the last few elections?
John Green: The more observant Catholics have been moving—let’s say creeping—in the Republican direction, and the less observant have been creeping in the opposite direction. We hear a lot about “swing voters.” The middle-of-the-road Catholics are the swing voters, and politicians spend a lot of time trying to appeal to them because they can be moved. In this election, the Catholic swing voters split pretty much down the middle: the more observant moved to the Republican side and the less observant to the Democratic side. Forty or fifty years ago, when Catholics were a very reliable voting bloc for the Democrats in the Midwest and Northeast, Republicans didn’t spend much effort on them at all, because there weren’t very many votes to be gained. But now Catholics are very much up for grabs, and a lot of effort is spent on trying to persuade them.
John DiIulio: I think that’s exactly right. It’s interesting that non-Hispanic white Catholic voters have a perfect record of voting with the winner in the last four or five elections. They are the bellwether voters. Republicans were hoping they were reliably Reagan Democrats, but they’re not. The fact that they are split, or nearly split, probably means that both parties will try to do more vote-hunting in that segment of Catholic voters in elections to come.
John Omicinski: Besides the Baptists, what other denominations are in the evangelical category?
John Green: The Assemblies of God, other Pentecostal denominations, the Presbyterian Church in America (not to be confused with the Presbyterian Church USA, which is a mainline church). The one that does confuse people is the Evangelical Lutheran Church in America, which is not an evangelical church. But other Lutherans, the Missouri Synod and the Wisconsin Synod, are evangelical. Another big piece is non-denominational churches, the fastest growing part of the evangelical community.
Michael Cromartie: Tell us again what a “less observant evangelical” is. I thought the very essence of evangelicalism was fervency.
John Green: It’s a person who is affiliated with one of the denominations we classify as evangelical but who claimed to attend church less than once a week. So if you were a Southern Baptist who didn’t show up at church every week, we would categorize you as a less observant evangelical.
Michael Barone: With Jewish voters, there seems to be a pattern emerging. My theory has been that it has really been fear of Christian conservatives that has moved Jewish voters toward a very heavily Democratic, very anti-Republican kind of feeling. They are voting as they did sixty years ago, which is not true of most of these other groups. It’s as if Roosevelt was on the ballot again against some goy.
John Green: I think there’s a lot of truth to that. In some of our surveys, we find Jewish political activists who in terms of their issue positions would seem to fit better in the Republican camp, but who report to us that the involvement of the Christian Right in the Republican party is deeply troubling to them. Part of it is the church-state issue—they are really worried about that. But it’s also broader. There is a cultural element. The Christian Right would certainly favor more conservative cultural policies, and many Jews would not, because they are culturally liberal.
Hillel Fradkin: Right after the election, some polls showed that younger Jewish voters were more Republican this time around than their elders by some significant degree.
John Green: I’ve seen reports of those surveys, but our numbers are too small to show that. An interesting study in Philadelphia showed a Republican trend among younger Jewish voters.
George Weigel: Within the white vote, can one say, on the basis on your data and other data, that going to church is the single most effective predictor of voting behavior?
John Green: I’m not so sure I’d go so far as to say it’s the single best, but it is certainly one of the best predictors.
Tim Shah: Is religiosity a good predictor of voter turnout?
John Green: Church attendance and other measures of religious behavior are very strongly correlated with voter turnout, and that’s true in all religious traditions. It seems to have to do mostly with social connectedness. People who go to places of worship frequently are more connected socially and are more likely to vote.
John DiIulio: One of my colleagues at Harvard, Sidney Verba, published a book called Voice and Equality: Civic Voluntarism in American Politics. He wasn’t looking for religion at all, but he discovered that, lo and behold, just what people had found with respect to religious institutional affiliation and volunteering. Especially among Latino populations, there was this interesting effect that people who were associated with a church were more likely to register and more likely to turn out to vote. Why was that? Well, church is a place where people learn certain skills, such as how to join together in a common cause, or how to write a letter to a congressman. They learn language skills also. That’s important to watch because I think, especially with this growing population of Latino voters, that the mediating role of the local church is going to be a non-trivial factor, not only in whether those populations get to the polls, but also in what they do when they get there.
DISCUSSION PARTICIPANTS
John Green, University of Akron; John DiIulio, University of Pennsylvania; Michael Cromartie, Ethics and Public Policy Center; Fred Barnes, The Weekly Standard; Michael Barone, U.S. News & World Report; Karlyn Bowman, American Enterprise Institute; Claudia Deane, Washington Post; Hillel Fradkin, American Enterprise Institute; Jody Hassett, ABC World News Tonight; Bob Jones, World; Mary Leonard, Boston Globe; Joe Loconte, Heritage Foundation; Terry Mattingly, Scripps Howard News Service; Richard Morin, Washington Post; Carlyle Murphy, Washington Post; John Omicinski, Gannett News Service; Richard Ostling, Associated Press; Gerard Perseghin, Catholic Standard; Ramesh Ponnuru, National Review; Thomas Pratt, Prison Fellowship; Tim Shah, Ethics and Public Policy Center; Robert Shogan, Los Angeles Times; Steve Wagner, QEV Analytics; George Weigel, Ethics and Public Policy Center; Larry Witham, Washington Times.
[Top]
Study Report
The 2000 presidential race will be remembered not only as one of the closest in American history but also for the unusual prominence of religion. From George W. Bush’s proclamation of Jesus Christ as his favorite political philosopher to Joseph Lieberman’s quotes from Hebrew Scriptures, religious rhetoric played an important role in appealing to America’s diverse faiths.
But how did the faithful vote? A just completed national survey offers some answers to this question, revealing both old and new patterns. Longstanding political differences among the religious groups undergirded the Bush and Gore vote, but in addition there was an increased polarization among the faithful. Both factors contributed to the closeness of the contest.
The Bush vote was substantially an alliance of more observant white Christians (Protestant and Catholic), led by evangelical Protestants; they were joined by less observant white Protestants. Together these groups made up about three-quarters of the Texas governor’s total. In contrast, the Gore vote essentially came from members of minority faith groups, especially black Protestants, plus secular voters and less observant white Christians. In total, these groups accounted for about three out of four of the vice president’s votes.
The Survey
The survey was conducted at the University of Akron as part of a larger project of the Ethics and Public Policy Center and was supported by a grant from the Pew Charitable Trusts. A random sample of adult Americans were interviewed in the spring of 2000 and contacted again right after the election. The data were weighted to match the demographic characteristics of the U.S. adult population, and a statistical model was used to correct for over-reported turnout. These modifications produced more accurate estimates of the vote, but the results were comparable to those found in the raw data. The weighted sample size was 2,363; the margin of error was plus or minus 4 per cent. This survey is the third in a series of election studies, and comparisons were made with the 1996 results.
Faiths and Politics
The great diversity of American religious faiths can be captured in two politically relevant ways: by religious tradition and by religious commitment. Four large religious groupings are commonly recognized: (1) white evangelical Protestants, (2) white mainline Protestants, (3) black Protestants, and (4) Roman Catholics. Another large group is (5) the secular population, those not affiliated with organized religion. These groups are listed in the tables at the end of this report, along with (6) Mormons, (7) other Christians (such as Eastern Orthodox, Christian Scientists, and Unitarian/Universalists), (8) Jews, and (9) other non-Christians (including Muslims, Buddhists, and Hindus). Although they do not constitute separate religious traditions, (10) Hispanic Protestants and (11) Hispanic Catholics are grouped separately because their voting behavior differs from that of their white counterparts.
Religious traditions influence the vote both by shaping their members’ values and by providing them with information on issues. Religious commitment can have an independent impact on the vote as well. For example, regular worship attenders often have more conservative values and issue positions than their less observant co-religionists. They are also more likely to vote, partly because of their greater social involvement and partly because they are easier targets for electioneering.
The impact of religious commitment is represented in the accompanying tables by the division of white Protestants (evangelical and mainline) and Catholics into self-reported regular (once a week or more) worship attenders, called more observant, and less regular (less than once a week) worship attenders, called less observant. Similar patterns obtained for some of the smaller groups, but for others, such as black Protestants and Mormons, regular worship attendance had no significant observable impact on the vote.
The 2000 Presidential Vote
Table 1 reports the two-party presidential vote in 2000 for the fourteen religious groups identified above. Each horizontal row adds up to 100 per cent; the total vote for the weighted sample is at the bottom of the table. The very few votes for minor-party candidates were excluded for clarity of presentation. (These voters were concentrated among the seculars and the less observant groups.)
The Republicans received especially strong support from white Protestants. For example, 84 per cent of more observant evangelicals voted for Bush. This figure was considerably higher than the comparable figure in 1996, when Bob Dole received 70 per cent of their votes. Observant mainline Protestants were also strongly Republican, backing Bush with 66 per cent, a margin that also grew from Dole’s 58 per cent. However, the most Republican religious group was the Mormons, 88 per cent of whom voted for Bush, about the same level as in 1996.
Less observant white evangelical and mainline Protestants resembled each other, backing Bush by 55 and 57 per cent, respectively. In 1996, Dole and Clinton nearly tied among these voters (some surveys showed the Democrats prevailing by a small margin).
Catholics were more evenly divided than white Protestants. The more observant supported Bush with 57 per cent of their vote, while the less observant backed Gore with 59 per cent. These margins were greater than in 1996, when the Republican-Democrat split was similar. In 2000 the Democrats won 76 per cent of the Hispanic Catholic vote, roughly the same proportion as in the previous election.
Black Protestants were the strongest Democratic group, giving Gore 96 per cent of their votes. This margin differed little from 1996, but increased turnout benefitted the Democrats. Gore also won Hispanic Protestants with 67 per cent, a gain for the Democrats over 1996, despite George W. Bush’s purported appeal to this group.
Jews were Gore’s second strongest group at 77 per cent. However, this figure may actually represent a decline from the comparable Democratic figure in 1996, despite the presence on the ticket of Joseph Lieberman, the first Jewish vice presidential nominee. In addition, Gore received strong support from other Christians (72 per cent) and non-Christians (80 per cent). Here, too, there appears to have been a decline compared to 1996. Secular voters strongly backed Gore with 65 per cent, a figure similar to their vote for Bill Clinton four years earlier.
Presidential Coalitions
Table 2 reports these same data from a different perspective: as a percentage of each candidate’s total vote. Vertical columns add up to 100 per cent. These figures show the relative weight of each religious group among the Bush and Gore voters (and in the third column, the electorate at large).
The first column looks at the religious composition of the Bush vote. More observant white evangelical Protestants were by far the most important group, accounting for almost one-third of the total. Less observant evangelicals contributed 8 per cent. Together evangelicals provided fully two-fifths of Bush’s popular vote.
Each mainline white Protestant group (more observant, less observant) accounted for about one-tenth of the Bush vote; together, then, they made up one-fifth of the total. White Catholics also made up one-fifth of the Bush column, with the more observant providing one-eighth and the less observant about one-twelfth.
Seculars contributed only one-tenth of the Bush vote, and all the remaining groups combined for one-twelfth. Note, however, the relative importance of Mormons: their three per cent was equal to the combined Bush vote from black Protestants and both Hispanic groups.
Looked at another way, evangelicals were the dominant group in the Republican presidential vote. Only a combining of the Bush votes from mainline Protestants and from Catholics equals the support provided by evangelicals. Regular church attenders were also dominant.
The religious composition of the Gore vote in the second column of Table 2 presents a sharp contrast. Black Protestants and seculars were the largest Democratic constituencies, each accounting for about one-fifth of the Gore total. Jews made up one-twentieth, and the other smaller groups (Hispanics, other Christians, and other non-Christians) combined for about one-twelfth.
Gore’s support from white Catholics made up one-fifth of his total, the same as Bush’s support from that group. And the impact of worship attendance was the mirror image, with the more observant providing slightly less than one-tenth of Gore’s vote and the less observant slightly more than one-tenth. Mainline Protestants and evangelicals each contributed less than one-seventh.
To look at the results another way, there was no dominant religious group in the Democratic presidential vote. All the less observant white Christians added up to one-quarter of the total, a little more than either black Protestants or seculars. And more observant white Christians accounted for one-fifth of the Gore total, equaling the contribution of black Protestants and of seculars.
The Faithful Mattered
These findings persist even when other demographic factors such as gender, income, and education are taken into account, which confirms the basic religious underpinnings of the presidential vote. Bush and Gore successfully mobilized the core religious constituencies of their parties, and in the process further polarized the faithful.
The Republicans’ deft handling of traditional moral issues, from abortion to presidential scandals, helped attract more observant white Christians, especially evangelical Protestants, while at the same time gaining support among less observant white Protestants. But the narrowness of this religious alliance nearly cost Bush the election: he lost New Mexico, Oregon, Iowa, and Wisconsin by tiny margins—and faced a contested result in Florida. Indeed, if Bush had done as well as Bob Dole among black Protestants and Hispanics, he could have won the popular as well as the electoral vote.
In contrast, the Democrats used other moral questions, including appeals to racial, environmental, and social justice, to rouse key groups, including black Protestants, seculars, and less observant white Christians. But here, too, the limitations of Gore’s religious coalition created problems. Partly because of his weakness among white Protestants, Gore lost his home state of Tennessee, Bill Clinton’s Arkansas, and the Democratic stronghold of West Virginia. Victory in any of these states would have given Gore a majority in the Electoral College, no matter how the contested ballots in Florida were resolved.
Thus, old patterns of religious group voting and a new polarization of the faithful help explain the closeness of the 2000 election. America’s many faiths were part of the divisions revealed at the ballot box, reflecting the unusual prominence of religion in the campaign.
