Friday, August 25, 2006

The Rightwing Group Slander Of Liberals Refuted—Part 5F

Abortion: A Summing Up

This concludes the abortion section of my argument (based primarily on data from the General Social Survey [GSS]) that it’s not liberals, but ultra-conservative movement conservatives who are far outside America’s mainstream. In contrast, ordinary conservatives and liberals agree much more often than not. This is the last of 6 sub-parts. [Links at end of post.]

At the end of the last post, I noted:
If abortion is murder, there are no extenuating circumstances.

This is, however, an extreme minority position, much like opposition to welfare state spending.
One key to reactionaries’ success is their ability to tap into extremist emotions, but escape responsibility for doing so—whether in stirring up violence or in launching arguments whose conclusions they shy away from. The position that abortion is murder does both these things. Here, we fix our gaze on the reactionaries’ extremist game, and what it says about who’s really out of touch with the American mainstream.

A Fringe Minority

As I repeated just above, the position that abortion is murder (not homicide—which can justifiable, in the case of rape, incest, or threat to the mother’s life, for example—but murder) is a fringe minority position, much like opposition to welfare state spending. In fact, it’s even more of a fringe position. A mere 8.1% of the population opposes abortion in all three cases in the AbThreat scale during the most recent timeframe.

It’s even more revealing if we put these two conditions together: opposition to welfare state spending and to all abortions on the AbThreat scale. The NatSpend6Sp scale measures attitudes toward six welfare state spending questions. Running cross-tabs between NatSpend6Sp and AbThreat yields the following:


The bold numbers are percentages. In the lower right-hand corner, the intersection of two fringe positions—one 16.6%, the other 6.4% (a bit lower than the 7.3% in the total survey population)—comes out to just 1.2%. This is the percentage of people who hold two key movement conservative positions: that abortion is murder, and that the welfare state should be cut. In contrast, the two boxes in the upper left-hand corner total to 55.9%. That’s the intersection of those wanting to increase welfare state spending and supporting abortion in all three cases in the AbThreat scale. That’s the liberal position that’s “out of touch with the American people,” according to folks in the 1.2% lower right-hand corner.

It’s amazing how much power these people have, given how marginal their attitudes are. They have been the driving force behind 30+ years of culture war, that has produced remarkably little over-all change in levels of support and opposition to abortion rights. And yet they remain a tiny fringe.

Questioning The Fringe

Part of the key to their power is simply that it’s never questioned. Reactionaries claim to be conservatives, and nobody questions them, really. By constantly demonizing liberals as “other,” they automatically draw attention away from the gaps that divide them, not just from the mainstream of American opinion, but from the mainstream of conservative opinion—the very body of opinion that they claim to represent, and have the most prominent influence over. Yet, for all their influence, they cannot convince a majority of self-identified conservatives to do away with the welfare state, or to oppose all abortions. Indeed, they can only maintain their positions of power and influence by hiding the full implications of these positions.

Consider what it means to call abortion murder. If that is so, then there are tens of millions of murderesses running around the country. If we were to take this seriously, America would have virtually no other industry except for the prison-industrial complex. The task of apprehending, convicting and imprisoning so many people would take up almost all of our national resources. Does anyone seriously advocate this? No. Of course not. To do so is to reveal the utter absurdity of the position.

One does not have to like abortion even a little to recognize that it is not murder. We do not treat it as murder because we cannot treat it as murder. However fervently a small minority of people may feel, if society were to act as if it were murder, society itself would fall apart.

Even the anti-choice leadership knows this, and so they resort to deception and bluff. They continue using the language of murder, but their aim is not to punish abortion as murder. Their aim, instead, is mass intimidation, for it is the only by intimidating women who want and need an abortion that they can be prevented en masse from getting one. And what kind of society can we be, living in such a state of mass intimidation?

Law & Morality: Separation Strengthens Both

The root problem here really is a confusion of law and morality, which in turn is a reflection of (not the same as) the confusion of church and state. The reasons for keeping them separate are diverse, but one reason should be crystal clear: too much mixing of the two undermines both of them. This was the view set out by John Locke in his Letter Concerning Toleration. As the Dictionary of the History of Ideas entry on liberalism explains:
Locke's Letter Concerning Toleration (1689), shorter
than Bayle's Commentary and more popular and less abstract than Spinoza's argument in the Tractatus, is the classical apology for liberty of conscience....

The proper business of civil government, according to Locke, is to protect and promote men's interests. Though everyone has the right to try to persuade others to hold beliefs which he thinks are true and important, nobody has the right to use force to that end. The civil magistrate has no authority from either God or man to require anyone to profess or refrain from professing a belief on the ground that it is true or false, necessary to salvation or incompatible with it. It is not for him to dispute with his subjects or to persuade them to a particular religion. Even if he could force them to adhere to it, he would not thereby save their souls, for salvation depends on a free adherence to what is true. A church is no more than an association of men who come together to worship God in the manner they think acceptable to him, and no church can claim authority from God to be the only teacher of the true faith. Like any other voluntary association it may make rules for its members, may admonish and exhort them, and may expel them for disobeying the rules. But it may not deprive them of their civil rights, or of any rights other than those they acquire by joining it, nor may it call upon the civil power to do so. No belief is to be suppressed merely because it is heretical, nor any practice merely because it is offensive to God. No doubt, what is offensive to God is sinful, but what is sinful is not punishable by man. No man deserves punishment at the hands of other men, unless he has offended some man, unless he has invaded his rights.

Locke, in this Letter, seems at times to come close to saying what J. S. Mill was to say long afterwards: that men are answerable to civil authority only for their harmful and not their immoral actions. Yet he does not say it outright, nor even clearly imply it.

What he does say is that all beliefs are to be tolerated “unless they are contrary to human society” or to moral rules “necessary to the preservation of civil society.” This is not a clear saying. What is to be reckoned contrary to human society or necessary to the preservation of civil society?
Indeed, these are not easily-answered questions. But the simple fact that abortion has been widely legalized for 30 years or more counts heavily against any argument that tolerating a belief in abortion rights falls into either category. To the contrary, as pointed out above, it is the outlawing of abortion that would create a state of mass intimidation incompatible with our notions of a free and democratic society.

Laws must stand by the consent of the governed, freely given, or respect for law in general will suffer for it. But morality can and must strike out for what is seen as right, however few agree. The two represent distinctly different approaches toward changing human behavior. A widely disrespected law, based on moral arguments, undermines respect for both.

What liberalism stands for, more than any particular result, is the unfettered exploration of multiple different perspectives and chains of thought. This can deeply frustrate those who are guided by a simple moral vision. But America was founded in the aftermath of bloody religious wars that scarred the face of Europe. Moral certainty had produced a bloodbath lasting for decades—this was the background against which Locke wrote, and the background against which our Founding Fathers took his lead. We had excellent reasons to divide the realms of morality and legality, allowing for influence, but not straightforward dictatorship.

Seeking Common Ground

The battle cry that abortion is murder is just one form in which that founding wisdom of our nation is attacked. It is good that so few truly believe in it. It is bad that so many are deceived by it. There is another way. A recent Pew Poll found a desire for common ground:
Abortion continues to split the country nearly down the middle. But there is consensus in one key area: two out of three Americans (66%) support finding "a middle ground" when it comes to abortion. Only three-in-ten (29%), by contrast, believe "there's no room for compromise when it comes to abortion laws." This desire to find common ground extends broadly across the political and ideological spectrum.

Majorities of Republicans (62%), Democrats (70%) and political independents (66%) favor a compromise. So do majorities of liberals, moderates and conservatives. More than six-in-ten white evangelicals also support compromise, as do 62% of white, non-Hispanic Catholics.

Only one group expressed unwillingness to find a middle way. Two-thirds (66%) of those who support an outright ban on abortion say there should be no compromise. In contrast, two-thirds of those who want abortion to be generally available are ready to seek an accommodation.
Those who support an outright ban represent just 11% of population. Two thirds of that is roughly 8%. Every other group predominantly favors finding a consensus. It won’t be easy, especially with 30+ years of rightwing-sponsored culture wars behind us, and all the misinformation and disinformation that entails. There are a lot of myths and outright lies to undo. But the desire to find common ground is very much an affirmation of what liberal democracy is all about. It’s what sets us apart from the theocratic tradition that America broke with at its founding, and that animates our most threatening enemies today. It’s a good sign. And it’s a further sign that liberals are much more in tune with America’s mainstream than reactionary “movement conservatives” are.


Links To Previous Parts

Here again are links to the previous parts of this series:
    Part 1: Introduction. Overview of argument and data.
  • Part 2: GSS Spending shows conservative support for the welfare state, and high levels of cross-ideological agreement.
  • Part 3 shows shifts in party identification consistent with the historical record of race as the primary impetus for white Democrats shifting to the Republican Party.
  • Part 4 looks at the partisan shifts through the lens of religion.
  • Part 5A “The Big Picture—Parties, Abortion and Race Over The Years” began the 6-part look at abortion.
  • Part 5B A big-picture snapshot of abortion attitudes.
  • Part 5C Changes in abortion attitudes over three time-spans.
  • Part 5D: Abortion, party and ideology over three time-spans.
  • Part 5E: Abortion and Church Attendance

Thursday, August 24, 2006

The Rightwing Group Slander Of Liberals Refuted—Part 5E

Abortion and Church Attendance

This continues the abortion section of my argument (based primarily on data from the General Social Survey [GSS]) that it’s not liberals, but ultra-conservative movement conservatives who are far outside America’s mainstream. In contrast, ordinary conservatives and liberals agree much more often than not. This is the fifth of 6 sub-parts. [Links at end of post.] It examines attitudes towards abortion in light of church attendance.

While there are strong correlations between church attendance and abortion attitudes, these have tended to be relatively stable over time. The shifts in opinion are generally modest, and surprisingly uniform at the national level—particularly in contrast to the strong pattern of opinion that remains basically unchanged—regular churchgoers are far less supportive of abortion overall than all other groups, who are relatively close to one another. There are, however, two exceptions, which provide some significant insight.

We begin by looking at the ABThreat Scale:
On the AbThreat scale, nationwide across all time periods, 60.3% of regular churchgoers support abortion in all three cases, compared to 77.9% among frequent churchgoers, 85.3% among rare churchgoers and 86.3% among non-churchgoers. The gap between regular and frequent churchgoers is 17.6%, more than double the 8.4% gap between frequent churchgoers and non-churchgoers. This further undermines the claim that secular Americans are out of the mainstream. What’s more, the gap between regular churchgoers and everyone is growing. In the last timeframe, it was closing in on 20%.

Next we look at the AbAutonomy scale:
On the AbAutonomy scale, nationwide across all time periods, the figures for abortion in all four cases are 17.3% for regular churchgoers, 34.4% for frequent churchgoers, 44.3% for rare churchgoers and 51.5% for non-churchgoers. In this case, the 17.1% gap between regular and frequent churchgoers is exactly equal to the gap between frequent and non-churchgoers. (In the last timeframe, however, frequent churchgoers were closer to non-churchoers—a 15.9% gap—than to regular churchgoers—a 17.4% gap.)

However, the relative percentages again show regular churchgoers to be farther removed: 34.4% is 99% higher than 17.3%, but 51.5% is just 50% higher than 34.4%.

Furthermore, as Robert Friedman pointed out in an earlier comment on this series, church attendance appears to be overstated on opinion polls. This makes it quite likely that if we could weed out the false churchgoing reports, this group would diverge even further from the rest of America, since the false churchgoers are probably more similar to frequent or rare churchgoers in their attitudes.

Two Telling Shifts In Opinion

With only two exceptions, the shifts in opinion for all attendance groups for all options were within two percent of the average opinion shift for all groups. The first exception was regular churchgoers on the AbThreat scale, dropping 8.1% in support for abortion in all three cases—5.9% greater than the average 2.2% drop. The churchgoers’ level of support was already low—63.9% compared to the next-lowest level of 78.2%. It dropped to 55.8%, compared to the next lowest level of 75.1%. This is due to a 13.7% drop in the white South, compared to a 5.9% drop in the rest of the country:
As we can see, this drop was due to something surprising—the fact that white Southern regular churchgoers were about 8 points more supportive of abortions in this category than their counterparts outside the white South. This gap has now closed. And the closing of that gap produces far and away the largest change over this period of time.

It’s hard to believe that this change was caused by secular humanists—to put it mildly. But if this change is hard to square with the “blame liberal secular humanists” narrative, the next change is even more at odds with it. You see, it seems that they’re getting slightly more conservative.

The second exception was non-churchgoers on the AbAutonomy scale, decreasing in support for all 4 cases by 0.5%, compared to an average 4.9% increase in support—the exact opposite of what the conservative narrative would have us believe about those godless heathen secular humanists. This is entirely due to a 2.6% decrease outside the white South, compared to an average 3.6% increase—as gap of 6.2%:
Within the white South, they were relatively normal, increasing 1.1% more than the average, which was skewed sharply lower by the regular churchgoers:
The most plausible explanation for this shift is that the expanded ranks of new non-churchgoers are more like their parents or former selves, and less like the pre-existing non-churchgoers. But this is sheer speculation.

In contrast, the first shift seems easy to account for: the refrain equating abortion with murder has often been repeated in churches. This would account for reducing support for abortions resulting from rape, or threatening the life or health of the mother or the fetus: If abortion is murder, there are no extenuating circumstances.

This is, however, an extreme minority position, much like opposition to welfare state spending. We’ll consider the fringe nature of this position, among other things, in the last abortion post.


Links To Previous Parts

Here again are links to the previous parts of this series:
    Part 1: Introduction. Overview of argument and data.
  • Part 2: GSS Spending shows conservative support for the welfare state, and high levels of cross-ideological agreement.
  • Part 3 shows shifts in party identification consistent with the historical record of race as the primary impetus for white Democrats shifting to the Republican Party.
  • Part 4 looks at the partisan shifts through the lens of religion.
  • Part 5A “The Big Picture—Parties, Abortion and Race Over The Years” began the 6-part look at abortion.
  • Part 5B A big-picture snapshot of abortion attitudes.
  • Part 5C Changes in abortion attitudes over three time-spans.
  • Part 5D: Abortion, party and ideology over three time-spans.

Wednesday, August 23, 2006

The Rightwing Group Slander Of Liberals Refuted—Part 5D

Party, Ideology and Abortion

Introduction

This continues the abortion section of my argument (based primarily on data from the General Social Survey [GSS]) that it’s not liberals, but ultra-conservative movement conservatives who are far outside America’s mainstream. In contrast, ordinary conservatives and liberals agree much more often than not. This is the fourth of 6 sub-parts. [Links at end of post.]

We return to the abortion scales introduced in Part 5A, to both create a sort of snapshot encapsulation of the individual question data presented in the previous post, and to add the refinement of looking at political identity—party and ideology combined. Here, we will downplay the tables of numbers we’ve been using, leading with the visual impact of graphs.

* The AbThreat Scale *

In the first chart, of the AbThreat scale nationwide, by party and ideology (political identity) we see an initial condition with no real trend, out of which a sawtooth pattern emerges, and then develops a somewhat more distinct slope:



At first, independents and Republicans are slightly more supportive than Democrats in all three ideological orientations—a significant fact that is easily forgotten today. (See table below, first column.) Liberal independents are noticeably more supportive of abortion in all three cases (the red line) than are liberal Democrats. But the differences were slight, and the shape of the red curve shows no larger pattern. There is a hint of the saw-tooth pattern that’s about to emerge, but the differences are slight, and liberal Republicans are less supportive than moderate Republicans, making it impossible to see a sawtooth without projecting it into the data.

In the second time-period, the sawtooth is unmistakable—a sign of ideology taking precedence over party. Thus, it is not Democrats who are emerging as distinctly more pro-abortion on the ABThreat scale—it is liberals So much so, in fact, that there is an imperfection in the classic sawtooth form: The liberal independents still represent a high-point of support, which means that the highest points of the saw tooth not only fail to fall into line—they actually change direction. Indeed, with that one exception, there seems to be little difference among liberals, moderates and conservatives of different partisan orientations. All the conservatives have about the same level of support, as do all the moderates as well as the liberal Democrats and the liberal Republicans.

This sort of arrangement is quite common in American politics, because our politics has generally been much less ideological than that of other countries. This is also a reason why it’s not very effective for Democrats to “move to the center.” The independents closest to the Democratic Party are generally more liberal than conservative Democrats are. On the ABThreat scale they are dramatically more liberal, in fact.

This cross-cutting of party and ideology generally tends to have a conservative effect, preventing the sort of broad systemic change that liberalism advances. The protracted nature of the culture wars—with abortion as the most central and enduring manifestation—is partly just a manifestation of how American politics tends to work in general. Battles go on for a long, long time, because there are many other allegiances that work at cross-purposes. It’s precisely the opposite of the polarization we’re constantly told about. Of course, we are experiencing more polarization than usual. But that’s still not very much in the grand scheme of things. The rhetoric of polarization far exceeds the reality.

In the third time-period, the sawtooth remains, again a bit imperfectly. But the partisan slant just hinted at before is also present—the first clear sign that ideological polarization is becoming partisan as well, though not to the same extent. Support has dropped significantly among both liberal and conservative Republicans, while liberal Democrats have pulled even with liberal independents.



Thus, in this stage, we finally have a rather mild version of the pattern that Ender claimed for the early 1980s in the comment that first instigated this series. For the first time, liberal Democrats are no longer more conservative on abortion than liberal independents. But this result at this late date clearly refutes Ender’s thesis that liberal Democrats were the engine driving this change.

Another way to look at this same data is to rearrange it, grouping liberals together on the left, subdivided into Democrats, Independents and Republicans, followed by moderates, and then conservatives. The first grouping was by party, then ideology. This is a grouping by ideology, then party. The emergence of the sawtooth in the first view is reflected in the emergence of definite slope in this second view:



The slope that emerges above still has an interruption, when we get to moderate Republicans, who are more pro-choice than moderate independents. The slope is also relatively modest. That’s to be expected. After all, there just isn’t that much difference in views. There is overwhelming support for abortion in the cases on the AbThreat scale.

* The AbAutonomy Scale *

To get a view of where real divisions lie, we have to turn to AbAutonomy. Here is what the party/ideology view looks like:



[Corresponding Figures in Table Below]

Here we see a similar sort of progression as we did with AbThreat, but we appear to begin in the middle of the process, where a sawtooth pattern has already emerged, and we end up farther along in the process, with a much more pronounced downward-slope. At the same time, the different curves grow closer together—a reflection of the large-scale pattern noted above, that the extreme positions increased slightly at the expense of already-much smaller numbers taking middle positions We begin with the liberal independent peak higher than the liberal Democrat peak on all four lines. Even the liberal Republican peak is higher at the top line (supporting abortion in 1 or more cases), though only by a whisker. Moderate independents and Republicans are nearly dead-even, about five points above moderate Democrats, and the same pattern holds for conservatives as well—though with a larger Democratic gap.

The overall result is that independents are most liberal, followed by Republicans, then Democrats—a strikingly different constellation than we have today. Of course, this does not automatically mean that the parties as a whole were aligned this way. The balance of liberals, moderates and conservatives was quite different in the two parties. Yet, as we’ve already seen above, this pattern generally did hold, even though conservatives were a far more significant block among Republicans.

The second time-frame still has independents as the most liberal, but Democrats and Republicans are much closer together. Liberal Democrats are clearly more liberal than liberal Republicans, but moderates and conservatives rank about the same.

The most significant change in the third time-frame is the sharp plunge among conservative Republicans. The percent supporting all 4 options drops five percent, while the range from total support to supporting just one option also tightens five percent, resulting in a total drop of ten percent for those supporting one or more options. Since conservative Republicans also continued to expand their share of the population, this constitutes a significant development. At the same time, moderate Republican support for abortion in all four cases increased about five percent, increasing the cumulative support for fewer options., but with diminishing impact at the “one or more” level. The only other change approaching it is that moderate and conservative Democrats both gained about five percent in support of abortion, pretty much across the boards.



If we flip our perspective to ideology-then-party, we see a familiar pattern: a gradual progression towards a consistent downward slope, which is interrupted by moderate Republicans, who are more pro-choice than moderate independents:



The initial state is more structured, a slight variant on the familiar sawtooth pattern, while the third state has a more pronounced slope than ABThreat had—a reflection of the fact that there’s a greater diversity of opinion among the four questions on the AbAutonomy scale, and greater polarization as well.

* AbAutonomy & The White South *

Finally, if we look at the white South, we find a slightly different pattern, from beginning to end, but the rest of the country shows a pattern almost identical to that of the country as a whole.

White South / Rest of Country / Nationwide:



The white South has sharper contrasts in general, due to lower levels of support amongst all groups except for liberal independents and Democrats. The lines are closer together as well, a further sign of polarizing, since it means that fewer people take intermediate positions between supporting or opposing abortion in all four cases:



This is a crucial point. Because of the concentration of conservative Republicans in the white South, opinions there appear more monolithic. But there is actually more polarization within the white South than there is outside of it. We saw this already in the figures for the individual questions. Lower levels of agreement equate with more polarization. And the white South had lower levels of agreement on every question in the last time-frame. This is yet another indication that it is conservative Republicans who are driving polarization, not liberal Democrats.

Links To Previous Parts

Here again are links to the previous parts of this series:
    Part 1: Introduction. Overview of argument and data.
  • Part 2: GSS Spending shows conservative support for the welfare state, and high levels of cross-ideological agreement.
  • Part 3 shows shifts in party identification consistent with the historical record of race as the primary impetus for white Democrats shifting to the Republican Party.
  • Part 4 looks at the partisan shifts through the lens of religion.
  • Part 5A “The Big Picture—Parties, Abortion and Race Over The Years” began the 6-part look at abortion.
  • Part 5B A big-picture snapshot of abortion attitudes.
  • Part 5C Changes in abortion attitudes over three time-spans

Tuesday, August 22, 2006

The Rightwing Group Slander Of Liberals Refuted—Part 5C

Abortion: Changes Over Three Time-Spans

Introduction

This continues the abortion section of my argument (based primarily on data from the General Social Survey [GSS]) that it’s not liberals, but ultra-conservative movement conservatives who are far outside America’s mainstream. In contrast, ordinary conservatives and liberals agree much more often than not. This is the third of 6 sub-parts. [Links at end of post.] Abortion is very important, as it represents the first long-lived social wedge issue supplanting the role of race, which has since slipped into the background. It is only such issue that has robust GSS polling from the early 1970s to date.

This post takes a look at the data across three timeframes in order to understand the changes that have taken place. While the anti-choice movement has produced little net change in attitudes, it has re-arranged the structure of such attitudes—particularly in the white South. Polarization has increased, but agreement still outweighs disagreement between parties and between ideological orientations.

Little Change In Net Support

In the last post, we concluded with a look at abortion views on 7 questions that have been asked repeatedly as part of the General Social Survey since the 1970s. Levels of agreement were high—ranging from 77.9% to 93.8%. In the white South, it went even higher—up to 95.2%. This overall view, however, masks considerable change over the three time periods we’ve used to break down the data (1972-1984, 1985-1993 and 1998-2004), as we’ll see in a moment. First, however, it’s important to note there was relatively little net change. As mentioned previously in this series, the two sets of questions are quite distinct, and can be used to create distinct indices—AbThreat and AbAutonomy, each measuring the number of cases for which an individual supports the right to choose an abortion.

We will use these scales more in future parts. But for now, we’ll just use them for a quick peak to establish the lack of dramatic change. If we look at the results for all three time frames (both nationwide and divided between the white South and the rest of the country) we find shifts of only a few percentage points, which are mostly indicative of increased polarization on the AbAutonomy scale, and a very slight increase in the low level of complete opposition on the AbThreat scale.

First, the ABThreat Scale:
The only notable shift is a slight increase in those supporting abortion in no case, with a comparable decline in those supporting abortion in all three cases. The “no case” increase is 2.5% in the white South, 1.0% elsewhere.

Yet, those supporting abortion in “no case” on the AbThreat scale remain a tiny fraction of the population—just under 1/12th, to be precise. This small fraction is the real maximum size of the anti-choice hardcore. This is the maximum number of those convinced that abortion is murder. Anyone who makes an exception for rape, birth defects, or health of the mother clearly does not believe that abortion is murder—regardless of what they may otherwise say. They believe that abortion is, at worst, homicide, which includes the category of justifiable homicide.

There is, of course, good reason to think that most of that 8% actually doesn’t think that abortion is murder. Few of them, for example, would seriously suggest executing women who get an abortion. This is further evidence that abortion as a political issue is significantly deceptive, however passionately individuals may feel about it.

Things are a little more complex with the AbAutonomy scale—both extremes gain slightly at the expense of the middle:
Again, the shifts are greater in the White South. Those supporting abortion in all 4 cases increased 6.4% in the white South, compared to 3.6% in the rest of the country. Those opposing abortion in all 4 cases increased 5.1% in the white South, compared to 2.2% in the rest of the country.

Polarization Changes Over Time

But if the levels of support and opposition remained fairly stable, the levels of polarization did not. For that, we return to the question-by-question view, beginning with the first time-frame.



And in ratios:



As we can see, the beginning state of agreement was around 95% for the AbThreat set (around 97% in the white South) and 85% for three questions in the AbAutonomy set, with the fourth at 77.9%. In ratios, this translates into the 20-1 range for the AbThreat set, and 6-1 for three of the questions in the AbAutonomy set—very high levels of liberal-conservative agreement. Few married couples agree that often on anything:

Things begin to change significantly in the second time-frame:



In ratios:



At this point, just one measure in the AbThreat set remains about 90, although agreement remains higher for all three in the white South. Agreement levels for the AbAutonomy set are now incredibly uniform, as agreement for the three questions that were around 85% fall to the level of the previous outlier, around 77%. Agreement remains higher in the white South.

In the third time-frame, agreement continues to erode, but the pattern changes as well:



And in ratios:



In the AbThreat set, a sharp split appears. While the highest level of agreement (for ABHLTH) drops 3 percent, support for the other two questions in the AbThreat set drops by twice that much. In the white South, they dropped by about 10 points—10.5% for ABDEFECT and 9.4% for ABRAPE. As a result, the level of disagreement in the white South now exceeds that in the rest of the country. In the rest of the country, agreement decreased by 4-5%.

In the AbAutonomy set, agreement plummeted 11-12% in the white South for all four questions, resulting in agreement levels in the mid-to-high 60s (about 2-1). In the rest of the country, polarization grew less than half as fast, from 2.8 to 5.7%, ending up in the 71-72% range for all four questions. For the whole country, agreement fell 5.0 to 7.6%, ending up in the 70-71% range for all four questions.

The changes discussed are summarized in the following table:



We can clearly see the greater increase in polarization in the white South. It holds for all seven questions. However, that was not the case for the transition between the first and second time-frames, when the opposite pattern prevailed for five questions. The white South actually became less polarized on ABANY, while polarization grew less quickly than it did for the rest of the country for the other AbAutonomy questions, and for ABDEFECT. All this was wiped out by the extraordinary jump in polarization in the white South in following transition, from the second to the third time frames, when all five of those questions increased in polarization by double digits.

This pattern of change is reminiscent of the pattern of change in partisan alignments. And, indeed, there appears to be an obvious reason: the explosive growth of conservative Republicans in the white South. However, as we’ve already seen, the interactions between ideology and party ID can be complicated. Which is what we turn to next.

Links To Previous Parts

Here again are links to the previous parts of this series:
    Part 1: Introduction. Overview of argument and data.
  • Part 2: GSS Spending shows conservative support for the welfare state, and high levels of cross-ideological agreement.
  • Part 3 shows shifts in party identification consistent with the historical record of race as the primary impetus for white Democrats shifting to the Republican Party.
  • Part 4 looks at the partisan shifts through the lens of religion.
  • Part 5A “The Big Picture—Parties, Abortion and Race Over The Years” began the 6-part look at abortion.
  • Part 5B: A big-picture snapshot of abortion attitudes.

Monday, August 21, 2006

The Rightwing Group Slander Of Liberals Refuted—Part 5B

Abortion: The Big Picture

After a long hiatus, I’ve resumed my argument (based primarily on data from the General Social Survey [GSS]) that it’s not liberals, but ultra-conservative movement conservatives who are far outside America’s mainstream. In contrast, ordinary conservatives and liberals agree much more often than not. In the second of 6 parts devoted to abortion, I look squarely at the data on agreement and disagreement, which is not the same as consensus, since it includes how much liberals take the conservative position as well the reverse. This is, in effect, a measure of how much polarization there is, not where majorities lie.

Some of the data presented comes from 7 questions that were asked repeatedly over the years. A detailed look at how these numbers have changed over time will be the subject of the next post in this series. But first, a simple “snapshot” post, to get a feel for the conceptual terrain.

Refresher on Agreement and Disagreement

The concept of polarization can readily be quantified for any question we have cross-tabs for. The method is straightforward. Either one of two complementary methods can be used—tabulating the agreement or the disagreement. With spreadsheets, it is trivial to do both. Explaining how and why they work is only a bit more complicated.

So let’s look at an example, and I’ll explain:



We have three conceptually identical examples. I’ll use the last for my example.

Agreement in any row is simply the minimum of the columns being compared. Here we are measuring liberal/conservative agreement and disagreement. (We could do the same with party ID, race, gender, anything we have crosstabs for.) The minimum of 32.6 and 11.8 is 11.8. The minimum of 28.4 and 23.6 is 23.6. And so on. The total agreement is the total of the column of all the agreement figures for all the options.

Disagreement is slightly more complex, for a simple reason: when two people disagree, their disagreement shows up twice: once in the category chosen by the first person, and once in that chosen by the second. For this reason, the sum of the column will have to be divided by two to get the proper figure for total disagreement. Disagreement for each row is the difference between the columns being compared. The difference between 32.6 and 11.8 is 20.8. The difference between 28.4 and 23.6 is 4.8. And so on. And, as already explained, the total disagreement is half the total of the column of all the disagreement figures.

The two total figures, agreement and disagreement, must add up to 100%—give or take rounding errors, which we have in this case: the total is 100.1%. The agreement ratio is the amount of agreement divided by the disagreement.

With that out of the way, we are now ready to put this concept to use.

Abortion: The Big Picture

On the macro scale—across the whole population—abortion views have changed remarkably little over the course of 30 years. However the composition of those views in the population has changed considerably, and this has resulted in increased polarization—but still, far less polarization than is commonly assumed. Furthermore, while there were substantial shifts in abortion views by political identity, there was virtually no shift by church attendance.

The GSS has 18 substantive questions on abortion. Of these 11 were asked once or just a few times in the early 1980s. They provide a snapshot of that time, with agreement levels ranging from the high 90s down to 72.3% nationwide, and 67.2%—just over 2/3rds—in the white South.

In terms of percentage agreement the results are as follows:



In terms of agreement/disagreement ratios, the results are as follows:



None of these figures is indicative of polarization anything like what we see at the Presidential level, which is around 0.4 to 1. That’s a level at which more than two out of three liberals and conservative disagree with on another. The lowest figure on the entire table above is roughly the reverse of that: 2.0 to 1, meaning that two out of three liberals and conservatives agree with one another.

Another seven questions have been asked repeatedly over the years. As mentioned in the previous post, these fall into two distinct groups. The first is a set of three questions dealing with threat situations—rape, threat to the mother’s health, or a strong chance of serious birth defects. The remaining four deal with choices that clearly reflect a woman’s right to autonomy—abortion for any reason, if not married, if married but wants no more children, and if low-income. In the discussion that follows, we will look at these both as groups of related questions, and collectively as separate abortion scales—AbThreat and AbAutonomy.

Considering the entire period as a whole, agreement levels among the AbThreat set of questions are remarkably high—the high 80s and low-to-mid 90s. Interestingly, the highest agreement level comes in the white South—an exception to the overall rule that the white South is otherwise more polarized than the rest of the nation. Agreement among the AbAutonomy set is significantly lower, though still above 75% (3-1), and closer to 80% (4-1) overall. There is surprisingly little variation among all four questions.



In terms of agreement/disagreement ratios, the results are as follows:



This overall view shows remarkably low levels of polarization. Interstingly, the levels of polarization are lower for the white South.

However, as we’ll soon see in Part 5C, (“Abortion: How The Picture Changes Over Time”), this combined data from the 1970s to date masks considerable internal change over the three time periods we’ve used to break down the data (1972-1984, 1985-1993 and 1998-2004).

Still, two things will also become clear: First, even at the most extreme levels that have appeared in recent years, polarization is still far below the levels seen in presidential elections. This is compatible with the recently reported polling data that most people want to some sort of common ground position developed on abortion. Second, the pattern of increasing polarization will not support the claim that liberal Democrats are driving the process.

Links To Previous Parts

Here again are links to all the previous parts of this series:
    Part 1: Introduction. Overview of argument and data.
  • Part 2: GSS Spending shows conservative support for the welfare state, and high levels of cross-ideological agreement.
  • Part 3 shows shifts in party identification consistent with the historical record of race as the primary impetus for white Democrats shifting to the Republican Party.
  • Part 4 looks at the partisan shifts through the lens of religion.
  • Part 5A “The Big Picture—Parties, Abortion and Race Over The Years” began the 6-part look at abortion.

Sunday, August 20, 2006

The Rightwing Group Slander Of Liberals Refuted—Part 5A

The Big Picture—Parties, Abortion and Race Over The Years.

After a long hiatus, I continue my argument (based primarily on data from the General Social Survey [GSS]) that it’s not liberals, but ultra-conservative movement conservatives who are far outside America’s mainstream. In contrast, ordinary conservatives and liberals agree much more often than not. Originally planned as a single post on abortion, it has grown so long that I’m presenting it in 6 sub-parts. Abortion is very important, as it represents the first long-lived social wedge issue supplanting the role of race, which has since slipped into the background. It is only such issue that has robust GSS polling from the early 1970s to date.

With decades of organizing, the anti-choice movement has managed to produce very little net change in the balance of attitudes, which remain closely divided, depending on how questions are asked. It has done a great deal to re-arrange the structure of such attitudes, and the ways they play out in the electoral arena. Polarization has increased since 30 years, but the end result falls far short of the extreme that is commonly supposed.

Backstory Refresher

The original claim that started this all was that liberals are to blame for Ann Coulter’s repeated calls for murder. Their alleged “extremism” around social issues in the early 1980s, it was claimed here, drove moderates and conservatives out of the Democratic Party and into the GOP. As a result, “Coulterization of the GOP is your own doing.”

In contrast, I have argued—using good, old-fashioned reality-based data—that [beginning in Part 1] polarization around social issues is vastly exaggerated, and that [in Part 3] race was the primary impetus for the growth of conservatism in the Republican Party, and exodus from the Democrats. This showed up first in votes for President—as far back as the Dixiecrat candidacy of Strom Thurmond in 181948, and then trickled down the ballot, slowed by the fact that Southern Democrats generally reflected the views of Southern whites, rather than those of the national party. Social wedge issues did play a role, of course—but it was primarily that of covering for race, facilitating, consolidating, rationalizing and moralizing the racially-motivated abandonment of the Democratic Party that centered heavily in the white South.

The primary proof was found in comparing shifts in political identification (from 1972-1984 to 1994-2004) in the white South with the rest of the country. Within the white South, consolidating the figures for liberals, moderates and conservatives within each party, Republicans gained 14.8%, and Democrats lost 14.0%, a dramatic net shift of 28.8%. (One gets slightly different figures using party ID figures directly, due to rounding errors, and people who don’t identify themselves ideologically: Republicans gained 14.1%, Democrats lost 14.7%, for am identical net shift of 285.8%) Outside the white South, by the first measure, Republicans gained just 1.8%, while Democrats lost 4.2%--a much more modest net shift of 6.0%, less than ¼ the shift in the white South. By the second measure, Republicans gained just 1.6%, while Democrats lost 5.3%--a next shift of 6.9%.

This still left the Democrats with a massive advantage outside the white South: 37.7% to 24.5% by the first measure, 37.5% to 24.3% by the second. Inside the white South, however their 42%/23% advantage in 1972-1984 eroded to a 27.9%/37.3% disadvantage, approaching a complete reversal, by the first measure. By the second, the results were nearly identical: a drop from 42.0%/23.2% to 27.3%/37.3%.

The evidence from look at abortion introduces secondary proof---the lack of any sudden, massive increase in social liberalism within the Democratic Party in the early 1980s. Indeed, the Democratic Party membership remained more conservative socially than the Republican Party until the late 1980s. Liberalism on abortion did not increase dramatically within the Democratic Party until the 1990s, as a result of, and reaction to twenty years of anti-abortion activism that centered strongly in the Republican Party.

For quick reference, here are links to all 4 previous parts of this series:
  • Part 1: Introduction. Overview of argument and data.
  • Part 2: GSS Spending shows conservative support for the welfare state, and high levels of cross-ideological agreement.
  • Part 3] shows shifts in party identification consistent with the historical record of race as the primary impetus for white Democrats shifting to the Republican Party.
  • Part 4] looks at the partisan shifts through the lens of religion.



The Plan of Part 5

Part 5 in this series consists of 6 parts. Before describing them, I need to quickly explain two different sorts of measure that we’ll be using. First are primary measures of the levels of support for and opposition to abortion rights among different groups. These are important for establishing when and where shifts in support or opposition occurred. Second are levels of agreement and difference in the primary measures between groups. These are important for establishing the degrees of polarization.

After the introductory material, Part 5A looks at primary measures. In the main section (“The Big Picture—Parties, Abortion and Race Over The Years.”), we see that partisan shifts in opinion did not occur suddenly in the early 1980s, as the conservative narrative claims. Indeed, it was not until after 1987 that Democrats remained reliably more pro-choice than Republicans. We contrast views on abortion with underlying attitudes on race, to emphasize how differently the two dynamics played out, and how clearly race has played a more constant, and leading role. (The abortion measures we use are two scales—AbThreat and AbAutonomy—which are explained below.)

These findings are sufficient to refute the claim that an early 1980s influx of liberals into the Democratic Party was the cause of increased polarization. They clearly show that the impact of race set in well before that time period, and was overwhelmingly focused in the same part of the country. They are perfectly consistent with the vast sweep of historical evidence that race drove the realignment of the parties in the South, not abortion.

The remaining posts go beyond simply refuting the rightwing myth, to developing a more nuanced understanding of how the changes actually took place over time.

Part 5B (“Abortion: The Big Picture”) moves on to looking at agreement and disagreement, starting with refresher to explain how the concept works in terms of a concrete example. After that, we look take a high-level look at the 18 substantive abortion questions asked by GSS. These fall into two distinct categories. First, we look at the 11 questions which were asked once or a few times in the early 80s, but not long enough to establish significant trends. Then we take our first direct look at 7 individual questions that were asked repeatedly from the early-to-mid 1970s to the present. We look at the results for the entire 30+ year timespan taken as a single whole.

These 7 questions fall into two distinct groups. The first is a set of three questions dealing with threat situations—rape, threat to the mother’s health, or a strong chance of serious birth defects. A combined measure of views on these three questions is called the AbThreat scale. It ranges from support for all three to opposition to all three. The remaining four questions deal with choices that clearly reflect a woman’s right to autonomy—abortion for any reason, if not married, if married but wants no more children, and if low-income. A combined measure of these views is called the AbAutonomy scale, ranging from support for all four to opposition to all four.

Part 5C (“Abortion: How The Picture Changes Over Time”) is a detailed look at those 7 questions over time. We examine them both as groups of related questions, and collectively as the two abortion scales—AbThreat and AbAutonomy. We examine them across three time periods (1972-1984, 1985-1993, and 1994-2004) and in two demographic divisions—the white South and the rest of the country.

Part 5D (“Party, Ideology and Abortion”) reintroduces the issue of party, and looks at how party and ideology interact in terms of abortion attitudes.

Part 5E (“Abortion and Church Attendance”) examines the role of church attendance. While regular church attendance is strongly associated with opposition to abortion, there was very little change over the years in the strength of this correlation, except for the AbThreat scale, most notably in the white South.

Finally, Part 5F (“Abortion And Political Narratives”) sums up the findings combines them with some other recent data and relates them to the larger issue of political narratives. It stresses the degree to which those driving polarization from the right represent precisely the sort of out-of-touch, anti-American fringe that they falsely accuse liberals of being. This is deeply significant in itself, but also quite timely, since exactly the same sort of dynamic is playing out in the attempted demonization of Ned Lamont, and the roughly 60% of the American people who agree with him. What this data shows is that Lamont’s candidacy is emblematic of a chronic, enduring condition in American politics, not just the acute crisis that has created the electoral opening.

We now turn to the body of Part 5A.

The Big Picture—Parties, Abortion and Race Over The Years

The lesson here is pretty simple: race has been a fairly constant means to distinguish between the parties since the Civil Rights Era, but abortion has shifted dramatically—not in the early 80s or before, when conservative revisionists would have it, but in the late 80s and early 90s, just around the time that Soviet Union disappeared as a bogey man, and “Operation Rescue” erupted as a magnet for drawing together anti-choice forces.

Let’s start off with looking at abortion, observe the pattern it reveals, and then look at race for a comparison. We’ll use two abortion scales that will be explained more fully below. The first one, AbThreat, combines three questions that involve abortion in threatening circumstances—rape, birth defect, or serious threat to the pregnant woman’s health. The second one, AbAutonomy, combines four questions that are concerned with abortion as function of a woman’s right to make fundamental life-choices, rather than have them made for her. We have two charts presenting these scales on a year-by-basis—one for the white South, the other for the rest of the country.

First, the white South:
The colors follow a simple scheme: red/orange=anti-choice; blue/purple=pro-choice; gray=margin less than 1.0%; darker colors=stronger margins. As they help to highlight, Republican Party members as a whole were more supportive of abortion rights through 1987. The resurgence of Republican support in 1990 (for AbThreat) and 1991 (for AbAutonomy) was probably nothing more than a statistical anomaly—much like the earlier periods of higher Democratic support—since the number of people in these samples is small enough that random fluctuations are to be expected. At any rate, the most it can signal is some lingering transition period. From 1993 on, the margins of Democratic support are overwhelming.

If we look at the rest of the country, the pattern is similar, but with even less ambiguity. The Republican edge in supporting abortion rights ends sharply in 1987, never to reappear:
The conservative narrative that first sparked this series claims that it was extreme social liberalism in Democratic Party in the early 1980s that drove conservatives out of the party, and created the current highly polarized political culture. We’ve already seen—in Part 3 of this series—that no such mass influx of liberals occurred. The above table shows that—in the white South, at least—there was more anti-abortion sentiment among Republicans than among Democrats. If anyone was out of step at this time, it was the anti-abortion forces that flocked to support Ronald Reagan in the GOP primaries in 1976 and 1980. Their views would not become clearly more typical of the Republican Party until the very end of his presidency in 1988.

In contrast, let’s look at a situation where views of party members are in harmony with the party and party activist position. We use a scale created from four different explanations for why blacks have a lower socio-economic status (SES) than whites. Two answers offered internal reasons—essentially blaming blacks for their own condition—lack of intelligence and lack of will. Two answers offered external reasons—discrimination, and lack of educational opportunity. The scale combines all four answers—counting a negative to an internal reason the same as affirming an external reason. Those who give all four external reasons score “1;” three external and one external score “2;” two of each score “3;” one external and three internal score “4;” all four internal score “5.” Table measures the results in two ways: First, by the percentage giving more external reasons than internal ones (scoring “1” or “2”); and, second, by the mean answer given.

From the time of the Carter presidency, in the white South, Democrats are more likely to assign external causes. There were just 4 years when this wasn’t so, and the differences then were less than 1% in one year, and less than 2% in a second. The remaining two years could simply be statistical anomalies. But whatever they are, they are far from constituting a trend:
In the rest of the country—which, of course, includes blacks and other minorities, most of whom are Democrats—the differences are even more pronounced:
I am not arguing that these tables alone prove that racial attitudes are the cause of shifting allegiances. There is plenty of historical data—narrative as well as statistical—to support this view as well, which is precisely the sort of data that’s needed to distinguish between correlation and causality. The tables alone can only support that view, which they do. They show an historical correlation of party and racial attitudes. The historical data shows the causality.

In contrast, the abortion scale tables do not support the view that abortion was the cause of shifting party allegiances. The shifts in allegiance went against the correlations. This is what happens when a previously minority position takes over a party.

We can see another striking example of this drawing on another data source, the American National Election Study [ANES], which began in 1948, and asked about school integration beginning in 1962. [An interface for accessing both GSS and ANES data can be found here.]

This gives us a very good time series to look at how Democratic Party civil rights activism—which really got serious in 1963, with the Civil Rights Act—impacted party attitudes. This longer view goes back 15 years farther than the SES table above, to when Southern Whites in the Democratic Party were more anti-black than Republicans.
The pattern here is clearly the same as that seen with the abortion scales—a party’s position is directly opposed to the majority of its members, which takes some years to come around. In the column tracking the decline of Republican support, we also see the effects of mass defections to the GOP by the more racist elements of the Democratic Party.

Conclusion

The post that triggered this series claimed that liberal extremism drove the process of political polarization, leading to the state where a leading conservative such as Ann Coulter has repeatedly called for murder. Specifically, it was claimed:
It is my understanding that the two parties have been moving further and further apart on several very important fronts. The cultural front encompasses many diverse issues from illegal immigration to God in the public square to what it means to fight a war to gay rights to affirmative action. The economic front includes taxation, social programs, welfare, etc.

In my view the Democrats have been moving in the counter-cultural direction since at least the 1980s. The liberal/progressive wing of the party has grown rapidly and politicians took notice. Conservative democrats either adapted and changed, were tossed out in favor of more liberal candidates, or switched to GOP. Democratic base became dominated by the generation that went to Woodstock. Liberals who control the Democratic Party's current agenda (and comprise somewhere around 25% of US population) differ greatly from the rest of the American public on all the cultural issues I listed above.

Liberals overwhelmingly support homosexual lobby's agenda, affirmative action, and abortion on demand. They are much more secular than the rest of the Democrats and indeed the rest of Americans....

An average American has noticed the ascendancy of the Liberal movement. An average conservative has taken it as a threat to the American way of life....

American conservatives sense real danger to their value system and they do not need Ann's rhetoric to know what is going on. She just provides a sense of validation for their beliefs. We have been polarized long before Ann and Republican spinmeisters came along. Liberals (right or wrong) through their counter-cultural agenda polarized us. The blame is misplaced.


We’ve already seen that this narrative is filled with myths and errors. The liberal wing of the Democratic Party did not grow rapidly in the 1980s, forcing out conservatives. It was the conservative wing of the Republican Party that grew rapidly, and it did so primarily in the white South, where it was the “extremism” of the Civil Rights Movement—and the Democratic Party finally enforcing the 13th, 14th and 15th Amendments—that drove them into the GOP.

In this post, we’ve seen that conservative and moderate Democrats in the white South were leaving the Democratic Party in the early 1980s for a party whose members were more pro-abortion than the Democrats. This directly contradicts the narrative above. That’s because race—not abortion, gay rights or any other “social issue”—was driving the shift.

These social issues were important, of course—primarily as a means of rationalization. Those who had been on the wrong side of history, opposed to racial justice, needed a way to recapture the stance of moral superiority that they had previously claimed simply by the color of their skin. The “defense of innocent life” was a perfect vehicle for this. What’s more, leaders of the religious right specifically identified abortion as an issue to use to create a mass grassroots movement in direct imitation of the civil rights movement. Of course, many of those attracted to this politics were not racially motivated. But the origins of the movements they joined most definitely were.

Links To Previous Parts

Here again are links to all 4 previous parts of this series:
    Part 1: Introduction. Overview of argument and data.
  • Part 2: GSS Spending shows conservative support for the welfare state, and high levels of cross-ideological agreement.
  • Part 3] shows shifts in party identification consistent with the historical record of race as the primary impetus for white Democrats shifting to the Republican Party.
  • Part 4] looks at the partisan shifts through the lens of religion.