My area of expertise is in public opinion and voting so I will focus on the measurement and description of polarization rather than its causes. From a public opinion perspective, one of the cleanest pieces of evidence for me is this graph of average attitudes on abortion (on a 1-4 scale, from the National Election Study) among self-declared Democrats, Independents, and Republicans:
On or about 1990, as a latter-day Virginia Woolf might say, American politics changed. I wouldn’t take the blip of the dotted line at 1990 very seriously–sampling variability and all that–but the general pattern in the graph above is real, and appear in all sorts of other data. In 1988 and before: zero correlations of partisanship with attitudes; since 1992, the correlations have been big and getting larger.
But, even now, we see polarization among some groups but not others. Here’s a graph made by Yair Ghitza, showing regression coefficients predicting abortion attitudes on party identification (with coefficients smoothed via a hierarchical model). For African Americans and Latinos, there has been a consistent near-zero correlation between partisanship and abortion attitudes. But among whites, the correlation has steadily increased since 1980.
And we can turn up the microscope even further by breaking down the whites into subgroups by income, education, or political information level:
Not only is the abortion/party relationship primarily driven by whites, it is substantially stronger among white elites–that is, people with high income, education, or levels of political information.
A lot more can be said about political polarization–indeed, entire books have been written on the topic–but perhaps these graphs give a sense of the value that can be gained by making comparisons for different groups of the population, and tracing those changes over time.
This is the latest post in our ongoing series on political polarization. The previous posts are listed below.