The Washington PostDemocracy Dies in Darkness

Americans have not become more politically polarized

(Alex Wong/Getty Images)

This is a guest post by Stanford political scientist Morris Fiorina.

The Pew Research Center recently released a report describing two decades of change in American public opinion. Much of the data—especially the striking graphical presentations—will find their way into college classrooms in the fall, including my own. But the numerous calls and e-mails I have received from political journalists testify to widespread misunderstanding of the report’s findings.

Although the report bears the title “Political Polarization in the American Public,” this is an inaccurate characterization of the findings. In common parlance polarization connotes a movement away from the center toward both extremes. This has not happened in the United States. If one thinks about polarization in partisan terms, one would expect to see an increase in the proportions of Democrats and Republicans and a decrease in the proportion of independents. But the American National Election Studies report that the distribution of American partisanship has been constant since the reelection of Ronald Reagan in 1984. Gallup had the proportion of independents at an all-time high in 2013.

If one thinks about polarization in ideological terms, one would expect to see a decline in moderates and an increase in liberals and conservatives. But the General Social Survey reports that the distribution of ideology in the United State has been stable since the early 1970s. With occasional small exceptions, “moderate” remains the modal category today just as it was in the days of Jimmy Carter.

If one thinks about polarization in terms of positions on specific policy issues, one would expect to see a decline in the center and a lumping up of people on the extremes. We do not have long time series of attitudes toward particular policy issues since they rise and fall on the national agenda, but on most issues, attitudes continue to cluster in the middle rather than lump up on the extremes.

In sum, we can argue about the size of the political center in the United States since the answer depends on various ways of measuring it, but whichever measure one chooses, the conclusion is the same: the country as a whole is no more polarized than it was a generation ago.

What has happened in the United States is not polarization, but sorting. Prior to the 1980s the Republican Party had a significant liberal wing and the Democrats a significant conservative wing. People of my vintage can remember liberal Republicans like Gov. Nelson Rockefeller of New York and senators like Jake Javits of New York, Hugh Scott of Pennsylvania, Charles Percy of Illinois, and Mark Hatfield of Oregon. No more. Similarly, the Democratic Party contained a slew of conservative southern governors, senators and representatives. In 1960 the greatest support for Civil Rights AND the greatest opposition to Civil Rights were both located in the Democratic Party, and in 1970 one would have been hard-pressed to say which party was more pro-choice. Today partisanship, ideology and issue positions go together in a way they did not in the mid-20th century. Issues and ideology used to cross-cut the partisan distribution, now they reinforce it.

But what about the data in the Pew Report?   The opinion distributions in the report appear to show a decline in the center and movement toward the poles. Appearances are misleading however. The Pew distributions show consistency, not extremity. The authors of the report are very clear about the difference (sidebar p. 21), but many commentators have conflated the two. The confusion of the two concepts is fundamental and underlies virtually all the misconceptions about the findings.

Pew constructs a liberal-conservative index from 10 policy questions. The items in the index are dichotomies: do you favor military strength or diplomacy; do you think blacks are responsible for their own condition or is racial discrimination responsible, etc. Most people do not fall 100 percent on one side or the other of such dichotomous choices. Rather, most would see some merit in both sides. A moderate Democrat, for example, might 65 percent believe that poor people need help but 35 percent worry about the incentive effects of welfare programs. She might 65 percent believe in diplomacy but 35 percent believe that military strength is important as well. If such a person responds to all the items in this moderately liberal way, she would fall into the extreme left category on the Pew scale, but her positions would not justify characterizing her as a left-wing extremist. Her score on the index reflects how consistently she takes liberal positions not the extremity of those positions.

Conversely, people in the middle of the Pew distributions are not necessarily centrists. Consider a stereotypical right-wing populist. Such a respondent espouses some extreme conservative positions — deport illegal immigrants, end programs for minorities, slash welfare, but also some extreme liberal positions — tax away business profits and heavily regulate their activities. This person will fall in the center (mixed) of the distribution even though he holds some extreme positions, but they are not consistently liberal or conservative.

In sum, the distributions in the Pew report do not run from extreme left through centrist to extreme right. They run from consistent liberal through inconsistent to consistent conservative. The report shows that ideological consistency in the American electorate clearly has increased. It does not show that extremity — polarization — has.

What the Pew report does, graphically and persuasively, is describe the progression of party sorting over the past 20 years. Political scientists continue to study the micro-processes underlying the sorting. Some people change their partisanship to match their ideology and issue positions, while some people do the opposite. Some older inconsistent voters die out of the electorate while some new voters choose the party that matches their ideology and issue positions. The net result is parties that are much more internally homogeneous than was the case a generation ago.

Most of the findings reported in the Pew study are reflections of party sorting. Increased antipathy toward the other party? If you are a conservative (liberal), there used to be people like you in the other party, so the other party wasn’t all bad. Now it is. Increased tendency to see the other party’s policies as a threat to the nation’s well-being? If you are a liberal and the other party’s policies have become increasingly conservative, or vice-versa, then those policies definitely look more misguided than they used to. Increased opposition to a president of the other party? If presidents govern to please the liberal and conservative bases of their parties, their agendas will be farther from the positions of members of the other party than they used to be.

As a side note, the Pew report precipitated a lively debate about the thesis of asymmetric polarization — that Republicans have moved far to the right while Democrats have not moved as far to the left, or have not moved at all. Critics of the thesis read the report as disproving the thesis. Democrats appear to show more movement leftward than Republicans moved rightward. Again, that is a misreading of the report. Democrats show a greater increase in consistency, not extremity, than Republicans. There is nothing surprising about that because Democrats were more badly sorted to begin with (see the Civil Rights example above).

Reading the Pew report it is hard not to feel some sympathy for the researchers who produced it. The report has been sucked into the maw of the polarization narrative despite the best efforts of the authors to qualify their findings. As noted above they are clear that their data measure consistency not extremity (although they slip up here and there as in the remark that “the center has shrunk” on p. 8, the section title on p. 18 and the figure title on p. 74). They point out that partisan sorting “tracks with increasingly polarized voting patterns in Congress, though to a far lesser extent” (p. 26) (my emphasis). They caution the reader (p. 7) that “These sentiments are not shared by all—or even most—Americans. The majority do not have uniformly conservative or liberal views. Most do not see either party as a threat to the nation. And more believe their representatives in government should meet halfway to resolve contentious disputes rather than hold out for more of what they want.”

Finally, nothing in what I have written here is meant to discount the importance of party sorting described in the report. There is nothing “mere” about it. On the contrary, as I have argued elsewhere, it is a fundamental cause of the gridlock and incivility that characterize contemporary politics. But contrary to the claims that “Americans are as polarized as Washington” or living in the “divided states of America, they are not. The unsorted and inconsistent middle still exists, but it has no home in either party.