Conflicted Conservatives By Christopher Ellis page4

The degree to which Americans systematically identify with an ideological label that does not reflect their policy beliefs is significant only inasmuch as it affects political choices. Moral and conflicted conservatives, unlike consistent liberals and conservatives, are “in play” electorally. Consistent ideologues behave as expected: In 2004, 91 percent of liberals voted for John Kerry, 96 percent of conservatives for George Bush. But 36 percent of moral conservatives reported voting for Kerry and 64 percent for Bush. And conflicted conservatives—who hold operational beliefs similar to those of consistent liberals, but differ in self-identification—split exactly evenly: 49 percent for Bush, 49 percent for Kerry.

Given that these groups are unsure how to align their own ideology and operational beliefs, it is possible that moral and conflicted conservatives are less likely to understand which candidate is, in fact, more conservative. Only 68 percent of moral conservatives and 66 percent of conflicted conservatives, for instance, placed George Bush to the right of John Kerry on a standard liberal-conservative scale (compared to 95 percent of consistent conservatives). Even more interesting, these citizens are considerably more likely to vote for the candidate they perceive to be more conservative—even if that perception is incorrect. Of the moral and conflicted conservatives who placed George Bush to the ideological right of John Kerry, 74 percent voted for Bush. But of those who considered Kerry more conservative, only 15 percent voted for Bush.

Much more remains to be done, of course, but the suggestion is that the political impact of one’s ideological self-identification may be real, even if that identification is derived outside of operational politics. The degree to which operational or ideological messages are prominent in political discourse, for example, may affect the decision-making calculus of conflicted conservatives. When specific policies are at issue, liberals win. When ideological symbols are prominent, conservatives win. The real power of ideology for many may not be its role in structuring political choices or its helping to understand how one’s own beliefs correspond with those of political elites. Rather, it may be the label itself that matters. The power of the ideological terms are not lost on political candidates—even the most liberal work to convince voters that they share “conservative” values, take a “conservative” approach to governance, and so on.

In addition, the “conservative” label likely has a particularly important appeal for the large portion of the population that is religiously orthodox, but largely uninterested in politics and operationally liberal on most issues. Candidates can occasionally gain electoral advantage by mobilizing morally and religiously conservative voters around some social issue of the time (same-sex marriage, for example). But individual issues such as these generally have little staying power, and the number of religiously orthodox voters who will care a great deal about any particular issue is quite small. Most religiously conservative Americans, for example, are likely to oppose gay marriage on principle. But finding a large number of religiously conservative voters who care enough about gay marriage (or any other particular issue) to have the issue drive their vote choice (especially if they are economically disadvantaged and operationally liberal on most other issues) is another matter.

But if politicians can connect traditionalist positions on these issues to a broader message of conservatism, then it is possible to mobilize a large group of religiously conservative voters, not around transient political issues of limited import to most, but rather under the guise of protecting a conservative value system that people both identify with and care deeply about. Instead of attempting to persuade doctrinally conservative voters using explicitly religious appeals (which make many Americans, even devout ones, uneasy), candidates can use the ideological term “conservative” as a way to show support for the traditional values of religiously conservative citizens.

In the context of survey research, the notion that many conservatives are identifying as such because of affinity for the term itself (and a dislike of its opposite, “liberal”) suggests that more research into the nature of responses to the ideological self-identification question itself is warranted. Many elite liberals, for example, have turned to the label “progressive” as a way to explain their general political values without having to call themselves “liberal.” This analysis suggests that asking respondents a “conservative-progressive” self-identification question instead of a “conservative-liberal” one would not only produce a different distributions of response, but would disproportionately affect the responses of certain types of citizens.

More subtly, the degree to which the ideological self-identification question is framed by its lead-in in an explicitly political way (“In politics today…”) as opposed to taking a general approach (“Generally speaking…”), or the context in which the question is asked (that is, whether it is asked before or after questions about specific issues, or near questions about religious commitment) may affect responses in a consequential way.

There are many loose ends in this account, and many other questions that are raised. But the broad idea is that the many Americans who call themselves conservative, but really are not in any political sense, are quite important to American electoral politics. Unlike consistent liberals and conservatives, they do not constitute the ideological and partisan bases of the major parties. But unlike nonideological “moderates,” they can be influenced by both operational and ideological appeals. As a result, the role of ideological self-identification in both micropolitical choices and macropolitical outcomes may be more interesting and subtle than is often assumed.

Christopher Ellis is a Ph.D. candidate at the University of North Carolina–Chapel Hill. This study is based on work supported by the National Science Foundation under grant SES-0118451, and the University of Michigan. Any opinions, findings and conclusions or recommendations expressed in these materials are those of the author and do not necessarily reflect the views of the funding organizations.