Two weeks ago, the Pew Research Center released a sprawling report on the relative politics of Americans, spawning a cottage industry of interpretative articles and posts sorting out just what “Political Polarization in the American Public” says about political polarization in the American public. The problem is, well, it doesn’t say much about it at all. As Stanford political scientist Morris Fiorina expertly explains at the Washington Post‘s Monkey Cage, Pew didn’t measure polarization, that is to say, political extremes. Rather, they measured ideological consistency, according to a 10-question scale of their own construction:

Items in the Ideological Consistency Scale

Pew constructed a scale whereby respondents were scored -1 for a “liberal” response, and +1 for a conservative response, then all the scores were added up to obtain the political consistency scale. A perfectly consistent conservative would score a 10, a perfectly consistent liberal, a -10. Thus, as Fiorina notes, “people in the middle of the Pew distributions are not necessarily centrists.” A “middle” score of 0 would be obtained by a mix of answers, none of which have their intensity or extremity measured, and so would indicate heterodoxy, not moderation. I suspect few TAC readers (or writers), for instance, would score perfect 10’s.

To see what this means in practice, look at the following graphic, frequently reproduced in the coverage of the Pew Report:

Understandably, many see the separation between the two sides, and assume that Americans are moving further apart on issues. What the graphic actually shows, however, is that Americans are becoming more ideologically consistent with their co-partisans. Republicans are becoming more consistently Republican, and Democrats more consistently Democratic. As Fiorina describes,

Prior to the 1980s the Republican Party had a significant liberal wing and the Democrats a significant conservative wing. … 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. … The net result is parties that are much more internally homogeneous than was the case a generation ago.

While some may breathe a brief sigh of relief that Americans, on average, generally occupy the same middle ground as they did a generation ago, these sorting changes may in fact be much more politically pernicious than mere extremism. As Noah Millman has discussed periodically, the decline of the pro-life Democrat and the pro-choice Republican has not necessarily been beneficial for either political cause. Instead, it binds the fortunes of abortion policy to nearly every other political issue, so that abortion policy is dependent on the political state of economics, or foreign policy debates. And with less heterogeneity within the parties, there is less reason for politicians to try to cross party lines in the first place: there’s no one waiting for them on the other side.

As the report notes, most Americans are not very ideologically consistent, and heterogeneity is still widespread. But because of the partisan sorting of the past decades, Fiorina says, “The unsorted and inconsistent middle still exists, but it has no home in either party.”