Cable news and social media portray US politics as beset by partisanship and political polarization. But do most Americans take ideologically extreme positions? In new survey research of almost 300,000 people, Anthony Fowler examines Americans’ likely real political positions. He finds that nearly 3/4 of Americans have views that correspond with a single ideology, and of those, most are moderate. These moderates, he writes, also drive the selection and accountability of political candidates.

Americans, we are often told, are polarized and hyper-partisan. To the extent that people appear moderate or independent in surveys, political scientists often assume that these individuals are unsophisticated, uninformed, ideologically naive, secretly partisan, or ideologically cross-pressured. But what if there are reasonably informed and sophisticated people with real views that just happen to be somewhere in the middle?

In new research with Seth Hill, Jeffrey Lewis, Chris Tausanovitch, Lynn Vavreck, Christopher Warshaw, we revisit the finding that in surveys, most Americans give a mix of liberal and conservative responses to policy questions. One possible explanation for this is that people aren’t paying attention to the survey questions, or they don’t have real views, so they answer as-if randomly. Another possibility is that people have idiosyncratic preferences that aren’t well summarized by a left-right ideological dimension. And a third possibility we explore is that people’s positions are ideological and well summarized by a single, left-right dimension, but they happen to be somewhere in the middle. For example, such a person might want to raise the federal minimum wage, but not to $15/hour. They might want some federal legal protections for abortion, but they don’t think it should be legal in all circumstances. And overall, these people would likely be well to the right of Democratic House Speaker Nancy Pelosi but well to the left of Senate Minority Leader Mitch McConnell.

Determining people’s real policy positions

We study the stated policy positions of more than 280,000 respondents across dozens of policy questions, and for each respondent, we determine which possibility is most consistent with each person’s pattern of responses. Specifically, we implement a mixture model that assumes that there are three types of respondents—inattentives who answer as if randomly, Conversians (named after the political scientist Philip Converse) who have idiosyncratic preferences not well summarized by a single ideological dimension, and Downsians (named after the economist Anthony Downs) who have preferences that are well summarized by a single ideological dimension. From each pattern of responses across questions, we estimate the probability that each respondent fits into each of those categories, and we estimate each respondent’s ideological position conditional upon them being a Downsian.

Photo by Karsten Würth on Unsplash

The intuition for how we do this is that Downsians should be more likely to select the conservative response to a policy question when the conservative response is more common among other Downsians. For example, a Downsian moderate will likely give a liberal answer to a question about whether we should abolish the federal minimum wage, but they will likely give a conservative answer to a question about whether we should raise the federal minimum wage to $20/hour. Alternatively, a Conversian might have an extreme, liberal view on minimum wage and an extreme, conservative view on abortion, while an inattentive respondent is equally likely to give any pattern of responses.

We estimate that approximately 6.5 percent of American survey respondents are inattentive, 20.7 percent are Conversian, and 72.8 percent are Downsian. In other words, approximately three-in-four Americans have views that are well summarized by a single ideological dimension. Furthermore, among those Downsians, most are close to the ideological middle. Therefore, most Americans have real policy preferences that just happen to be moderate.

Figure 1 below shows how our classifications and ideological estimates correlate with political behaviors and characteristics of interest. In each panel, we plot an outcome of interest across our estimated ideology scores (where the mean is 0 and the standard deviation is 1), and we separately show results for Downsians (black), Conservians (dark gray) and inattentives (light gray).  As expected, ideologues (i.e., Downsian liberals and conservatives) are more likely than others to vote, donate to candidates, and consume political news. However, Downsian moderates and Conversians also participate in politics (albeit to a lesser extent), and they are more likely than other groups to change the party they support between elections. Therefore, this latter group may be more electorally consequential than previously appreciated.

Figure 1 – Electoral Attitudes and Behavior across Types of Respondent

Kernel regressions of electoral attitudes and behaviors across estimated ideologies (mean = 0, standard deviation = 1) for Downsian (black), Consversian (dark gray), and inattentive (light gray) respondents in the 2016 CCES.

The moderate majority

To improve our understanding of electoral selection and accountability, we study how our survey respondents voted in congressional elections. Previous research has shown that ideologically moderate and more experienced candidates perform better in elections, and we replicate these findings. Interestingly, we find that these effects are almost entirely driven by Downsian moderates and Conversians. Although the ideologues participate in politics at higher rates, electoral selection and accountability are driven by the non-ideologues who are somewhere in the middle.

Accounts of American politics that focus on extremists and partisans are missing the often silent but electorally consequential moderate majority. The US electorate is much more moderate than you would think if you watched cable news or paid close attention to social media, and scholars and politicians should pay closer attention to the middle of the political spectrum.

• This article is based on the paper, ‘Moderates’, in American Political Science Review.
• Please read our comments policy before commenting.
• Note: This article gives the views of the author, and not the position of USAPP – American Politics and Policy, nor the London School of Economics. 
• Shortened URL for this post: https://bit.ly/3SeUIyO