Sunday, August 06, 2006

The Rightwing Group Slander Of Liberals Refuted—Part 4

This is part of a series devoted to refuting the rightwing narrative slandering liberals as an alien force in American politics, not just out of touch with the American mainstream, but downright hostile to it. I began [part 1] by laying out the big-picture argument [links to previous parts at end of diary]:
  1. It’s not liberals, but ultra-conservative movement conservatives who are far outside America’s mainstream, based on data from the General Social Survey.
  2. This data shows that liberals and conservatives have significantly more overlap than disagreement across a range of controversial social issues, as well as national spending priorities.
  3. On the other hand, those wanting to downsize the welfare state represent only a small minority—even among self-described conservatives.

In part 2, I examined the data on spending. Part 3 highlighted the impact of race in driving the growth of conservative Republicans, particularly in the South, as a prelude to examining social wedge issues. This part is a previously unplanned look at the role religion played in that shift. It’s got some surprises. And more numbers than a campaign.

The Story So Far

This part represents an unplanned departure from my script. My first part laid out the big-picture argument. My second looked at spending, and the third looked at the role of race as a social wedge issue in altering the pattern of political identification—both in terms of party and ideology. The historical data from the General Social Survey (GSS) is quite clear, and mirrors the shift of the “Solid South” from solid Democratic to solid Republicans. The percentage of conservative Republicans has doubled in the white South from the 1972-1984 time-frame to the 1994-2004 time-frame: from 11.8% to 23.6%, while moderate Republicans gained 40%, from 7.5% to 10.5%, and liberal Republicans slipped slightly, from 3.7% to 3.2%. At the same time, outside the white South, the gains were far more modest, from 11.0% to 14.7%--and over half that 3.7% gain was offset by losses to moderate and liberal Republicans, totaling 1.9%.

Here, I'm taking a closer look at how religion figured in this process as sociological force. This does less to change the story than it does to add nuance and prevent the story from being hijacked. The complexity of cross-cutting trends is yet another warning against simplistic narratives and talking point explanations.

In the comments to Part 3, johnsonwax said:
I think the key to the numbers above

is realizing that it's not a zero-sum game. Many of those southern conservatives did not exist in the statistics in previous years. The religious right was ignored by both parties until Reagan and the religious right ignored the parties. It's a huge voter base that one party recognized and swallowed up whole. The Dems have largely focused on young voters, but that's not nearly as exclusive a set as the religious right.
This is not really a rightwing narrative, it’s part of the general conventional wisdom. And I deflected it somewhat, but didn’t challenge it in my response, Good Point, But... :
You make a very good point, but it's not the major thing we see going on here. This is not the electorate we are looking at here, but the American people as a whole. So the politicization of fundamentalists and evangelicals--those who were truly apolitical before--would not show up very clearly...

Does this mean that the effect you're talking about was negligiable? No, not at all. I think it was substantial. But I think it was overwhelmed by the shift out of the Democratic Party due to race. I also think that a lot of the impact of the religious right was not just to mobilize the formerly apolitical or marginally political, but to cement the effects of race among the Southern conservatives and moderates who left the Dems over race--and to raise their level of participation. (Remember, in the Democratic Solid South, it didn't even matter if they never voted!)

This analysis is somewhat speculative on my part, since I hadn't examined the data that could confirm it. I'm just reasoning from the data above.
Well, now I’ve looked at it, it looks like I maybe overstated it. Let’s say the impact in increasing voting by fundamentalists was “significant,” but “substantial” may be pushing it. Yes, there was an increase in fundamentalists voting. But the impact of that was entirely secondary compared to shifts in voting patterns—and the vast majority of those were already voting. The most important factors are (1) whether or not we’re talking about white Southerners, (2) the changing voting patterns of theological fundamentalists, moderates and liberals, and (3) how often they go to church.

Prelude: The White South vs. The Rest of Us

I’m not going to present a lot of analysis here. You’ll get more than enough below. I just want to present two sets of charts, showing the relative political power of conservative Republicans in the white South and outside it. First the white South:

Among regular churchgoers, conservative Republicans are the green line, streaking up into the stratosphere, while all the other lines converge into a relatively narrow range, far below. Among all white Southerners, the pattern is similar, though less extreme. Now let’s look at the rest of America:

Among regular churchgoers, conservative Republicans again break out of the pack to a position of clear numerical superiority. But the growth trajectory slows considerably from the second time-frame to the third. Its degree of dominance is considerably smaller. Among the whole non-white-Southern population, conservative Republicans aren’t even the largest group, much less the dominant one. What’s more, their growth hasn’t just slowed since the second time frame—its reversed itself into a decline.

This is what the two Americas look like in terms of religion—the white South, dominated by avidly church-going conservative Republicans, and the rest of us, dominated by no one.

And now, let the slice-and-dice begin.

Church-Goers Vote More Often

First, let’s look at a fairly dramatic chart. It shows how the percentage of voters rises steadily and dramatically with church attendance among Southern whites, in every time period. Indeed, those who never attend church vote less than 50% of time:

The figures are less dramatic outside the white South. One reason we would expect this is that voting is related to how well integrated people are into social and institutional networks that make up what’s commonly called “civil society,” and the South has relatively little of civil society outside of the church, and church-related institutions. The North and the West have many more non-church forms of public participation for people to be civically engaged in. As the following chart shows, those who never go to church outside the white South voted 5.8% above white Southerners for the whole time period. For regular churchgoers—almost weekly or more often—the gap falls to 1.4%, less than a quarter of the gap for non-church-goers.
This also demonstrates yet another way in which Pat Robertson & company are mistaken. Their fear-mongering about secular humanism—which he and others like him often falsely label as a religion—taking over the country is the complete reverse of what the data shows. Not only are their no secular humanist churches, but those who don’t attend them vote significantly less often than regular church-goers do. In fact, the most noticeable increase in voting percentage occurred among white Southern regular churchgoers—Robertson’s core audience. Their voting percentage increased from 71.7% in the first time-frame to 77% in the second time-frame, and 78.1% in the third time-frame. (These gains were offset, however, by declines in churchgoing, as we’ll see below.)

This series was originally inspired in response to the argument that liberal dominance of the Democratic Party with anti-mainstream values (including secularism) was responsible for propelling conservatives out of the party and into the GOP. This is a popular narrative on the right, promoted by a few high-profile figures, such as Ronald Reagan, in the 50s and early 60s. In fact, the opposite was the case.

There was a modest white exodus from the party at this time, primarily due to growing affluence that resulted directly from unionization, the GI Bill, federally-subsidized suburbanization and the like—all products of the Democratic New Deal. People left the Democratic Party because it made them so affluent that they started to feel like Republicans.

The Democratic Party did not “leave” anyone until it firmly embraced civil rights with the Civil Rights Act of 1964 and the Voting Rights Act of 1965. And the loss of the Democratic “Solid South” is proof positive that the leaving was all about race. The notion that other, later factors were important—that secular humanists had taken over the Democrats, for example—is all part of the narrative to hide the central role of racism in building the modern conservative movement. The low rates of voting among non-church-goers is yet another piece of hard data that directly contradicts the movement conservative narrative.

Changing Patterns Among Theological Orientations

A comment about the increased participation of fundamentalists was the spark that touched off this diary. But far more important has been the shifting patterns of voting among fundamentalists as well as those with other theological orientations. One look at the flip-flop in fundamentalist party ID should be enough to make that point:

Fundamentalists went from 55.3% Dem/31.8% Rep to 32.8% Dem/51.1% Rep. This 41.8%+ swing was clearly far more important than any increase in the number of fundamentalist voters. After all, as long as Democrats got more of those voters, additional voters helped them. It was not the additional voters, but who they were voting for that mattered.

Moderates and liberals also roughly flip-flopped, but with less extreme results: 31.1% among theological moderates, and 17.2% among theological liberals, who started off as the most pro-Republican, and ended up as the most pro-Democratic.

Outside the white South—due, in part to black Protestants—the picture is significantly different. There have been shifts, to be sure, but no flips. In fact, among theological liberals, the shift was towards the Democrats. The Democrats remain a majority among all three theological orientations:

The biggest shift was amongst fundamentalists—20.5% toward the Republicans, just under half of their gain in the white South. But the GOP would need another comparable shift, just to draw even in this group. The shift amongst theological moderates was 12.5%, and again, the GOP would need another shift of this magnitude to draw even. Among theological liberals, Democrats gained 7.6%, and the balance there is almost identical to what it is amongst fundamentalists.

But most tellingly, perhaps, the GOP lost ground among all three orientations between the second and third time frames. Indeed, Democrats had even lost ground amongst theological liberals between the first and second time frames—7.5% to be precise. But they came roaring back to gain just more than twice that in the next time frame, to end up with a net 7.6% gain. This same picture will show up however we slice the data: Democrats lost ground everywhere from 1972-1984 to 1985-1993, but then came back outside the white South, with varying strength across a wide range of different groups.

Churchgoing in Decline

As the following table shows, regular church attendance is in decline, while the number of non-churchgoers grows:
The decline is actually sharper in the white South: In absolute percentage (the percentage of the population in each category, not the percent change) regular attendance is down 5.5%, compared to 4.7% in the rest of country. Frequent attendance is also down 1.0%, compared to a 0.9% increase in the rest of the country. The pattern was reversed with rare churchgoing: up 1.5% in the white South, down 1.0% in the rest of America. The percentage of non-churchgoers rose 4.9% in the white South and 4.7% outside it.

The story in terms of percent change (not shown) sounds even more dramatic: In the white South, regular attendance is down 14.0%, compared to 13.4% in the rest of country. Frequent attendance is also down 6.0%, compared to a 5.8% increase in the rest of the country. Rare churchgoing: up 4.5% in the white South, down 2.8% in the rest of America. Non-churchgoing rose 45.0% in the white South and 34.0% outside it.

These declines are only part of a larger picture, however. A July 20, 2004 press release from NORC (which conducts the GSS), “America’s Protestant Majority is Fading NORC Research Shows” stated:
The increasing secularization of American society has taken a particular toll on Protestant identity, presenting the prospect that after more than 200 years of history, the United States may soon no longer be a majority Protestant country, according to a new study by the National Opinion Research Center (NORC) at the University of Chicago.

The percentage of the population that is Protestant has been falling and will likely fall below 50 percent by mid-decade and may be there already, the research reported.

From 1972 until 1993 the Protestant share of the population remained stable. But then a decline set in. In 1993, 63 percent of Americans were Protestant, but by 2002, the number was 52 percent, NORC research found. During the same time, the number of people who said they had no religion went up from 9 percent to nearly 14 percent....

The survey shows a continual erosion on many measures for Protestants in the past generation, while the proportion of people who say they are Catholic has remained fairly steady at about 25 percent of the population. People who said they belonged to other religions, including Eastern faiths and Islam, Orthodox Christians, interdenominational Christians, and native-American faiths increased from 3 percent to 7.

Because of this sharp decline from 1993 onward, the figures from the third timeframe surely understate the declines in churchgoing as of today.

These declines took their toll in the voting booth as well:
Regular churchgoers declined as a percentage of the electorate, both in the white South (45.6% to 40.3%) and in the rest of the country (40.1%to 34.9%). As noted above, there was an increase in voting among regular churchgoers in the white South, but this was more than offset by the decline in their numbers, relative to the population as a whole. What we have seen within that overall decline is a net shift from regular mainline church attendance to regular fundamentalist and evangelical churches. The racist shift I wrote about last time is facilitated by that shift. It’s not that these churches were openly racist—a handful were, but the vast majority were not. Rather, they were simply unconcerned with the social justice dimension of the Gospels—a dimension that figured prominently among mainline denominations. Typically, when they do profess concern, they express it in terms of charity, not justice.

It was, after all, the mainline churches that formed the strongest institutional support for the civil rights movement in the white community, second only to the more progressive unions, most notably the UAW under Walter Ruether. Both the churches and the unions, however, had significant resistance amongst their memberships. And both lost significant membership during the periods covered by these surveys, although for different reasons. Nonetheless, the end result is a decline in social and political infrastructure supporting racial equality, and egalitarianism in general.

Compensating for these declines is a vital need for progressive politics, one that is commonly ignored. The reason is simple: it has low immediate payoff. But the long-term benefits are enormous, as correlation between church-going and voting suggests.

Church Attendance Vs. Orientation

There are two different trends in voting. First, as noted above, church attendance correlates with increased voting. Second, the more liberal one’s theological orientation (fundamentalist, moderate, liberal), the more likely one is to vote. There’s one more trend that matters here: The more conservatives one’s theological orientation, the more likely one is to attend church more often. This last, church-attendance trend works against the second voting trend. The way the two interact can be seen in the following table:
(This table comes with a major caveat: in 1984, the GSS refined the way it coded denominations—which in turn refined the assignment of respondents’ theological orientation. The survey has never attempted to ask about orientation directly, instead relying on that of the denomination named. So even the post 1984 data has some unknown amount of assignment error. That’s one reason I’m not going to devote my most intensive analysis to this data. The conclusions I will draw are based on broad patterns that we can be confident of because they show up so consistently. I’m staying away from fine distinctions were the assignment errors might play a significant role. Just to be on the safe side, I discussed this briefly on Friday with Tom W. Smith, Director of NORC's General Social Survey. There is reason to believe that voting participation data is relatively comparable pre-and post-1984.)

At every attendance level during every time frame, theological liberals voted at higher rates than fundamentalists, from an astonishing high 22.1% to a low of 8.4%. This shows the higher rates of voting amongst theological liberals at similar levels of church attendance. The total difference is shown in the unweighted average for each time frame. It takes no account of the attendance differences. But the higher rates of church attendance amongst fundamentalists offsets this considerably. The result is shown on the last (“total”) line in each time frame, where the liberal-fundamentalist difference is consistently smaller than the unweighted average, by roughly around seven or eight percent.

Although the comparisons to the first time frame are inexact, because of the recoding GSS did in 1984 to capture more specific information on denominations, we can see that (1) the trend is clearly toward reducing the gap in voting rates, but (2) a gap still remains. It’s the attendance gap and the growth of fundamentalism more generally that are more significant, however. If theological liberals attended church as often as their fundamentalist brethren, and voted accordingly, the voting gap from 1994 to 2004 would have been more three times what it was—10.7% rather than 3.1%. Among other things, Al Gore would have unambiguously won the Presidential election in 2000.

Church Attendance And Political Identity: Parties First

With all the above in mind, let us turn to less ambiguous data for our refined analysis. This involves church attendance and political identity. Following the earlier data, showing that the white South has lead the shift toward conservative Republicanism, we examine the white South first. We begin by looking at party ID and church attendance. Below we’ll break that down further into liberals, moderates and conservatives within each party ID.
This table above reveals a very interesting pattern. In the first time frame, Democratic dominance was still in place, at least as far as party ID was concerned. Democratic identification hovered around 40% for all four categories of church attendance. Republicans and independents split the remainder, which was almost 50/50 among regular church-goers, with Republicans losing strength fairly evenly to fall 10.2% among non-churchgoers, where they trail independents by a 5-2 margin.

In the second time frame, Democratic support had dropped over 12 points, but was even more consistent, hovering around 28% of the vote. Meanwhile, the Republican skew toward regular churchgoers had more than doubled: Republicans took 46.4% of regular church-goers compared to just 24.2% of those who never attend—a 22.2% gap.

We find a surprisingly similar pattern to begin with in the rest of the country, with a much more modest pattern of change:
Again, we start out with the Democrats taking roughly 40% across the board, with slightly more variance than in the South. Republicans and independents split the remainder with the same pattern seen in the South. They start off even closer to 50/50 among regular church-goers, and the Republicans again drop almost 10 percent (9.8% to be precise) and trail independents by a little over 5-2.

Outside the white South, the changes follow a similar pattern—a broad Democratic decline matched with Republican gains centered on regular church-goers—but the magnitudes are so much smaller that end result looks entirely different.

The Democrats drop only about three points overall, dropping most among regular church-goers, least among frequent church-goers, and roughly the same three point drop among the rest. As a result, they outnumber Republicans at every level of church attendance—something they only did for non-church-goers in the South. They outnumber Republicans among regular church-goers by 6.3%, compared to being outnumbered in the South by almost three times as much—18.6%. In addition, they hold a massive 20.2% margin among frequent church-goers.

The reason for this is clear: The Republicans picked up 4.1% among regular church-goers, but could only manage one other increase larger than 1%--a pickup of 2.5% among rare church-goers. In fact, they actually lost ground in every category since the 1985-1993 time frame—from a 2.3% decline in frequent church-goers to a 5.1% decline in rare church-goers. If these had been increases on top the pervious ones, the story would have been quite different. As it is, this is not the picture of a growing majority party. It is the picture of a declining national party, buoyed by a still-growing regional base.

With that in mind, we turn to the combined national picture:
The only group that the Republicans lead among are the most regular voters—but only by 1.2%. In contrast, they trail in two other categories by double digits, and in a third by over 8%. They are in much better shape than they were in 1972-1984, but they are going backwards in three categories, and are virtually stalled in their strong suit—regular church-goers. Again, this is not the picture of a growing national party. It’s the picture of one that has peaked. Of course, this is no guarantee of Democratic success. It merely indicates that Democrats have a real opening. Their fate is in their own hands.

Church Attendance And Political Identity: Party-Ideology Groups

The fate of the parties is the big picture. But the theme of this series is ideological as well as partisan. And so we turn to the finer-mesh picture we get from examining church attendance and political identity—defined as the combination of ideology and party ID:
In raw percentages, the biggest gains were by conservative Republicans among regular attendees (18.6%) and frequent attendees (12.4%). However, the next figure is a bit of a surprise: Conservative Republicans gain more (8.7%) among non-attendees than among rare attendees (7.8%). There are now more white conservative Republicans who don’t go to church in the South (14.6% of all non-church-goers) than there are liberal Democrats (13.1%). Still, this remains the only group in which conservative Republicans aren’t the most numerous group. They’re only #2, behind moderate independents (19.2%). The gain among non-church-goers reflects the fact that racial politics cuts across all the white South.

The only other group to gain ground among all four groups is the moderate Republicans, though in two cases their modest gains come after giving back a lot of gains they earned in the second time frame. Among regular church-goers, they gained 3.6% from the first to the second time frame, only to drop 2.2% back to 9.8%. At the other extreme, among non-church-goers, they jumped from 6.5% up to 7.3%, only to tumble back down to 6.6%. It’s only among the rare attendees (up 5.8% to 12.4%) and frequent attendees (up 3.2% to 12.1%) that moderate Republicans gained ground twice in a row, and significantly improved their standing.

The only other group to gain ground more times than it lost was liberal Democrats, on the other extreme. They gained 1% each among non-churchgoers (up to 13.1%) and frequent churchgoers (up to 9.2%). They broke even on rare churchgoers and lost ground among regular churchgoers (down 3.2% to 6.4%).

In between these two extremes were two sets of three groups with similar patterns. One set—moderate and conservative Democrats and liberal independents—lost ground in all four groups. They were the biggest overall losers. Collectively, they lost 17.2% among non-churchgoers, 14.6% among rare churchgoers and frequent churchgoers, and 14.1% among regular churchgoers. The other set—moderate and conservative independents and liberal Republicans—lost ground twice. One—the moderate independents—gained ground twice, while the other two gained ground once and broke even once. Collectively, they pretty much broke even. They gained 7% among non-churchgoers and 1% among rare churchgoers. They lost 3.1% among frequent churchgoers and 2.8% among regular churchgoers.

In the rest of the country, things looked very different, as the following table shows:
Among regular churchgoers, conservative Republicans made large gains—up from 14.1% to 22%, in first place by more than 6%. But that gain of 7.9% came at the expense of a 3.8% loss among liberal and conservative Republicans. In every other category liberal democrats took a larger share—as did moderate Democrats. They came in fourth among frequent churchgoers, fifth among rare churchgoers, and sixth among non-churchgoers.

In short, what we’re seeing is pretty much the same thing we saw from a slightly different angle before: the growth of conservative Republicans among regular churchgoers is both more modest than in the South, and lacks broader resonance. Indeed, almost half the growth comes at the expense of less conservative Republicans.

When we look at America as a whole, the white South pattern is submerged, and resulting pattern looks more like the pattern seen among minorities and non-Southern whites:
Conservative Republicans post strong gains among regular churchgoers, but partly at the expense of other Republicans. Although they make gains among all other attendance groups, they do not dominate any of them, coming in second among frequent churchgoers, third among rare churchgoers and sixth among non-churchgoers. Other Republicans fare worse: Moderate Republicans lose ground among frequent churchgoers as well as regulars, while liberal Republicans lose ground in all four attendance groups.

In all groups except regular churchgoers, liberal Democrats made modest gains, and liberal plus moderate Democrats combined outnumbered conservative and moderate Republicans combined.

A Broader View

All the above data are consistent with my original thesis—that the driving force behind political polarization is racial politics. It has taken a long time to set in at the level of political self-identification, thus cementing Republican power in the white South. But it was evident long before in presidential voting, as noted in the previous installment. Liberal activism around social issues—such as women’s rights and gay rights—were natural outgrowths of the civil rights movement. Likewise, conservative opposition to them was an outgrowth of opposition to the civil rights movement, and the GOP’s use of them as wedge issues to split the Democratic base has become a defining characteristic of American politics over the past 30-40 years.

The conservative narrative blames liberals for creating the current climate of polarization. It denies its racists roots, as well as its own pro-active role in fomenting resentment and social divisions. Instead of race, it points to issue like abortion and gay marriage. Yet, as I’m about to demonstrate in my next installments, the issue-by-issue attitudes on wedge issues are nowhere near as polarized as our political narratives suggest. The 72% gap between liberals and conservatives in the 2004 presidential vote cannot be found in any issue area, even for the most contentious of specific questions. To understand this systemic polarization, we have to look beyond issues to framing narratives, and beyond the narratives to the larger political system that generates them, and the social system of which it is part.

That will come at the end of the series. For now, we can simply conclude that religion reflects a divided America. The white South is increasingly dominated by church-going conservative Republicans. The rest of the country remains far more balanced in its views. It remains a place for dialogue—if we stand up for it.

Reactionary movement conservatives have been trying to convince us that the South is a model for all America, and lectures from the pulpit are the prototype for all political speech. But the data we’ve looked at shows that’s just not true. The successes of Kathleen Sebelius in Kansas and Brian Schweitzer in Montana are not flukes. They reflect the fact that slaveholders were defeated in Bloody Kansas in the 1850s, and never got anywhere near Montana.