Sabato's Crystal Ball

Ideology in the American Public

Alan I. Abramowitz, Senior Columnist October 1st, 2009

facebooktwittergoogle_plusredditpinterestmail

Is the United States today a center-left nation or a center-right nation? There is no question that Barack Obama’s victory in the 2008 presidential election and the Democratic gains in the 2006 and 2008 congressional elections dramatically changed the ideological make-up of America’s political leadership. On almost every major domestic and foreign policy issue, Barack Obama is well to the left of his predecessor in the White House, George W. Bush. And it is equally clear that the ideological makeup of Congress has shifted significantly to the left since 2006. But did this leftward shift among our nation’s political elite reflect a similar leftward shift among the American public?

Despite the results of the last two elections, some conservative pundits and politicians have argued that there has been no leftward shift in ideology among the American public. They claim that the United States was and is a center-right nation. In their view, the Republican Party did not lose the 2006 and 2008 elections because it was too conservative but because it abandoned its conservative principles during the Bush years. It follows that to revive its fortunes the Party should not move to the center, as some moderate Republicans have suggested, but go back to those conservative principles.

One piece of evidence that is frequently cited by those who claim that the U.S. remains a center-right nation is that when Americans are asked to describe their political views, they continue to prefer the conservative label to the liberal label by a wide margin. Thus, a recent Gallup Poll found that more than twice as many Americans described themselves as conservative (38 percent) as liberal (18 percent). Moreover, 39 percent of the respondents in this poll said that their political views had become more conservative in recent years while only 18 percent said that their political views had become more liberal.

This preference for the conservative label is nothing new, of course. It has existed since at least 1972 when the American National Election Study (ANES) began asking respondents in its biennial election-year surveys to place themselves on a 7-point liberal-conservative scale. In that survey and every subsequent ANES survey, self-identified conservatives outnumbered self-identified liberals by a wide margin.

Despite the popularity of the conservative label, however, many polls have found that Americans frequently take liberal positions on specific policy issues such as health care. In the 2008 ANES, for example, 42 percent of Americans placed themselves on the conservative side of the 7-point ideology scale while only 28 percent placed themselves on the liberal side. But in the same survey, 51 percent of Americans placed themselves on the liberal side of a similar 7-point scale measuring support for government-sponsored health insurance while only 37 percent placed themselves on the conservative side. More recently, a CBS/New York Times Poll in September 2009 found that almost two-thirds of Americans supported a government-run insurance option as part of a plan to reform health care in the U.S. It is findings like this that have led some liberal political commentators to argue that the United States is now a center-left nation.

So who is right? One way of answering this question is to compare the effects of ideological identification and policy preferences on Americans’ political behavior. For example, we can find out whether one of these measures of ideology has a stronger influence on voting decisions. Fortunately, an October 2008 Time Magazine poll provides an unusual opportunity to compare Americans’ ideological self-identification with their positions on a wide variety of specific policy issues. In this survey, approximately 1,000 likely voters were asked to place themselves on a five-point liberal-conservative scale and to give their opinions on ten policy issues ranging from abortion and gay marriage to the war in Iraq and health insurance. On each issue, respondents were asked to place themselves on a zero-to-ten scale with zero indicating strong opposition and ten indicating strong support for a specific position.

In the Time poll, as in almost every other national survey since the 1970s, self-identified conservatives outnumbered self-identified liberals by a wide margin: 16 percent of likely voters described themselves as very conservative, 26 percent as somewhat conservative, 29 percent as moderate, 20 percent as somewhat liberal, and 9 percent as very liberal.

When asked about their views on specific policy issues, however, these likely voters took the liberal side more often than they took the conservative side. On six of ten issues–abortion, health insurance, the war in Iraq, regulation of financial institutions, government assistance to homeowners threatened by foreclosure and global warming–a majority or plurality of respondents came down on the liberal side. On three issues–gay marriage, offshore drilling, and business tax cuts–a majority or plurality of respondents came down on the conservative side. On one issue, federal bailouts for financial institutions, it was not possible to identify liberal and conservative positions.

Opinions on seven of the ten policy issues were correlated strongly enough that it was possible to combine them into a liberal-conservative issues scale. The issues included in the scale were abortion, gay marriage, health insurance, offshore drilling, global warming, the war in Iraq, and regulation of financial institutions. Scores on this scale ranged from 0 for those who took the most conservative position on all seven issues to 70 for those who took the most liberal position on all seven issues. I then collapsed the scale by combining scores of 0-10, 11-20, 21-30, 31-40, 41-50, 51-60, and 61-70 to form a seven-point scale with 1 the most conservative score and 7 the most liberal score.

Figure 1.



Figure 1 displays the distribution of likely voters on the collapsed seven-point issues scale. In contrast to the ideological identification scale, this issues scale has a liberal tilt with 46 percent of likely voters on the liberal side of the scale (5-7) vs. 37 percent on the conservative side (1-3) and 16 percent in the middle (4). But the more important question is which one of these measures of ideology had a stronger influence on presidential candidate choice?

Both the ideological identification scale and the issues scale were strongly related to candidate preference. However, the issues scale had a substantially stronger relationship with the vote than the identification scale–the correlation (Pearson’s r) was .77 for the issues scale versus .62 for the identification scale. Ninety-three percent of those on the liberal side of the issues scale voted for Barack Obama while 92 percent of those on the conservative side of the scale voted for John McCain. Moreover, the partial correlation between the ideological identification scale and the vote while controlling for both the issues scale and party identification was a meager .01. In contrast, the partial correlation between the issues scale and the vote while controlling for both ideological identification and party identification was a substantial .39. This is especially impressive given the powerful influence of party identification on candidate choice.

Figure 2.

The strength of the relationship between the liberal-conservative issues scale and the presidential vote can be seen in Figure 2, which displays the distribution of likely voters on the 7-point issues scale by candidate preference. There is very little overlap between the policy preferences of Obama and McCain voters: 81 percent of Obama voters were on the liberal side of the scale and only 6 percent were on the conservative side; in contrast, 76 percent of McCain voters were on the conservative side of the scale and only 7 percent were on the liberal side.

Implications

Despite the continued popularity of the conservative label among Americans, on many specific issues public opinion has a liberal tilt. Moreover, it is policy preferences that seem to matter most in determining candidate choice. However, while it may be useful to know that the policy preferences of the electorate had a slightly liberal tilt in 2008, at least on the issues included in the Time survey, it is probably more politically significant that Democratic and Republican voters were deeply divided on these issues.

These findings provide further evidence of the extraordinary level of partisan polarization in the American electorate today. This is important because Democratic and Republican officeholders are primarily responsive to the preferences of their own electoral constituencies and those preferences are decidedly liberal for most Democratic officeholders and decidedly conservative for most Republican officeholders. So regardless of which party is in power, supporters of the opposing party are bound to be very unhappy with the direction of public policy. It is not surprising, therefore, that recent polls have found over 80 percent approval of Barack Obama’s performance among Democrats, but less than 20 percent approval among Republicans.