Two recent essays arrive at seemingly opposite assessments of trends in Americans' ideology. First, Larry Bartels wrote an op-ed in the Washington Post (9/30/13) asserting that Americans are more conservative than at any time since 1952. Second, a Gallup Organization report (1/10/14) trumpeted that pooled polls found a larger percentage of Americans identified themselves as liberals in 2013 than at any time since 1992.

(I wish I had Gallup-type data going back to the early 1950s, but, alas, efforts to locate such were fruitless.)

Which of these claims is closer to the mark? (Hint: only minimal change in Americans' ideology can be documented.) And what is ideology's role among the public? (Hint: Americans' ideological outlooks are more complex than either Bartels or Gallup allows.)

Bartels' measure of ideology is much more complex than Gallup's, which was a single question: "[h]ow would you describe your political views - [ROTATED] very conservative, conservative, moderate, liberal, or very liberal?" Bartels uses James Stimson's "policy mood index" which is "derived from responses to a wide variety of opinion surveys involving hundreds of specific policy questions on topics ranging from taxes and spending on environmental regulation to gun control." Although Stimson "tracks liberal policy mood," Bartels flips the measure to explore a "conservative policy mood."

Using that approach to tapping Americans' "policy mood," Bartels reports that, contrary to those who have claimed that Obama's election and re-election indicate Americans' political opinions have shifted leftward, Bartels asserts that, "[i]n reality, Stimson's data show, the public was already more conservative than usual in 2008, and a good deal more conservative by 2012."

Gallup's measure of ideology has already been described. Gallup uses it to track trends in Americans' ideology from 1992 to 2013. Jeffrey M. Jones, who wrote Gallup's report, correctly notes that the percentage of the public identifying as "liberal" or "very liberal" (23) is higher than at any time in the last 21 years.

Jones also reports that, in the early- to mid-1990s, the percentage of Americans calling themselves some kind of liberal was 16-17%. Moreover, he observes that the percentages consistently identifying themselves as liberals did not rise to 20-22% until around 2005 and thereafter.

Jones also writes that the percentage of the public labeling themselves as some kind of conservative has ranged from the upper 30s to the low 40s, thus outstripping liberal self-identification. (In 2013, the percentage of Americans identifying themselves as conservatives was 38.)

So, has the American public become more liberal lately, or is the conservative "mood" rising?

In a limited sense, both stories have merit. But, a much more compelling perspective on ideology in America -- if we're satisfied with a surface perspective -- is one of constancy. (I can't assess Bartels' data, because he provides no specifics. Given the venue, that doesn't surprise.)

Jones' graphic presentation of trends in ideology between 1992 and 2013 shows very minimal change -- from year-to-year or from decade-to-decade. At most, annual shifts in percentages identifying with the three ideological categories -- liberal, moderate, conservative -- are two percentage points. If we look at shifts between 2012 and 2013, for example, each of the three categories changes by one percentage point.

If all we had were Gallup's data, we should probably tread lightly. But, additional data buttress the claim that ideological self-identification has changed very little for many years.

Almost every year between 1996 and 2012, except 1997, the Pew Research Center asked random samples of the American public: "[d]o you think of yourself as conservative, moderate, or liberal?" In 1996, 40% of the public identified themselves as conservatives, 40% said they were moderates, and 20% claimed to be liberals. In 2012, the percentages of the three categories were 39, 37, and 23. With one or two exceptions in the 1990s, most years saw shifts of one or two percentage points.

These data enable the Pew Research Center's conclusion that, "[d]espite electoral swings in recent elections, the fundamental ideological breakdown of the American public has shifted little in recent years."

To reiterate, if ideology can be captured by a single query, Pew's data underscore the assertion that Americans' ideological identifications have changed very little over the last three decades. Moreover, whether one looks at Gallup or Pew Research Center polls, self-identification as conservative has outweighed liberal identification by as much as 2-to-1 or as little as 3.5-to-2.

Claims that the "Obama Era" represents a profound shift to the left at the American grassroots ring hollow.

We come at last to that infamous question, what difference does it make? (Thank you, Hillary Clinton.)

In 1964, Philip Converse published a seminal essay, "The Nature of Belief Systems in Mass Publics." In that paper, which has been repeatedly confirmed over the last 50 years, Converse reported that only a tiny portion of the American public held wide-ranging, internally consistent policy opinions. Converse also noted that the associations between ordinary citizens' policy opinions were far weaker than the correlations between comparable dispositions expressed by congressional candidates. In short, a very large percentage of the public is "innocent of ideology." (The "innocence of ideology" phrase comes from Paul Sniderman's 1993 compendium on American public opinion.)

It would be mistaken, however, to think that we can jettison the notion of ideology when discussing American politics.

First, as researchers have repeatedly shown, political elites are motivated by ideology, and any discussion of political cleavages between Democrat and Republican elites should take ideological matters into consideration.
Second, ideology can still tell us something about public opinion among Americans. However, matters are more complex than Gallup's report, the Pew Research Center's data, and even Bartels' op-ed imply.

In 1967, Lloyd Free and Hadley Cantril published The Political Beliefs of Americans, which was a comprehensive analysis of American public opinion in the mid-1960s. They reported that many Americans were "ideological conservatives" and "operational liberals." These people simultaneously opposed big government in the abstract, but favored a plethora of specific governmental programs.

The notion that Americans are ideological conservatives and operational liberals has been confirmed by Albert Cantril and Susan Davis Cantril in Reading Mixed Signals (1999), and by Christopher Ellis and James Stimson in Ideology in America (2012).

All this sends "mixed signals" to Republicans and conservatives. GOP candidates who are conservative don't confront an overly hostile public (although they must remember that only two-fifths of the public think of themselves as conservative). More important, conservative candidates for public office have to do more than just rail against big government. Some way has to be found to wean people off unnecessary "nanny state" programs, while shoring up programs that contribute to the polity's health.

The "trick" is to distinguish between nanny statism and legitimate government functions.

Conservatives ought to acknowledge that government has its uses... which are nicely articulated in the Constitution's Preamble, namely, "to form a more perfect Union, establish Justice, insure domestic Tranquility, provide for the common defence, promote the general Welfare, and secure the Blessings of Liberty to ourselves and our Posterity..."