A few weeks ago I noted that Congressional polarization has been the highest in 120 years. I briefly discussed a couple of explanations and noted that I would discuss each one in further detail. Before beginning with academic explanations, let me first discuss a journalists explanation via an academic scholar. Evan Thomas, in a Newsweek article, uses the work of Markus Prior (Princeton) to explain that electoral polarization is the consequence of “post-broadcast” democracy. How? Evan Thomas makes the following argument:
In 1970, at about 6:30 p.m. at least two or three nights a week, about half the country could be found watching the evening news on one of the three major networks. The broadcasts tended to be fairly sober-minded, on-the-one-hand, on-the-other-hand presentations…But then, in the 1980s and ’90s, came cable TV and the Internet. Before long, viewers had scores of channels to choose from, or they could abandon TV altogether and entertain themselves online. Prior estimates that about half the viewers of the evening news wandered away to watch entertainment—sports, movies, reality TV, whatever.
The people that wandered away (and this is where Evans makes a major causal leap) were the moderate voters. The moderate voters not only stopped watching the news, but stopped participating in the electoral process. So Evans’ solution, “[T]hey will have to switch off the Xbox or click away from the Home Shopping Network or ‘Girls Gone Wild’ and go out and vote.”
How can we test Evans’ theory of electoral polarization?
Evans implicitly makes the claim that voters are more extreme than non-voters. So, are voters more extreme than non-voters? Fiorina, in Whatever Happened to the Median Voter, notes,
Combining party and ideology reveals somewhat more evidence of polarization between voters and nonvoters. Republican (Democratic) party identifiers who vote are increasingly more conservative (liberal) than those who do not. In sum, electorates are slightly more partisan and more ideological than the country as a whole, but as documented by numerous empirical studies (most recently, Teixeira (1992: ch. 3) the differences are not major.
So the hunt continues…