Abramowitz, Alan I. 2010. The Disappearing Center: Engaged Citizens, Polarization, and American Democracy. New Haven, CT: Yale University Press.Google Scholar
Aldrich, John H. 2011. Why Parties? A Second Look. Chicago, IL: University of Chicago Press.Google Scholar
Bawn, Kathleen, Martin Cohen, David Karol, Seth Masket, Hans Noel, and John Zaller. 2012. “A Theory of Political Parties, Groups, Policy Demands and Nominations in American Politics.” Perspectives on Politics 10 (3): 571–597.Google Scholar
Cook, Rhodes, Alice V. McGillivray, and Richard M. Scammon. 2010. American Votes 28: Election Returns by State. Washington, DC: CQ Press.Google Scholar
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper & Row.Google Scholar
Erikson, Robert S., Gerald C. Wright, and John P. McIver. 2006. “Public Opinion in the States: A Quarter Century of Change and Stability.” In Public Opinion in State Politics, edited by J. Cohen. Palo Alto, CA: Stanford University Press.Google Scholar
Fiorina, Morris P., with Samuel J. Abrams. 2009. Disconnect: The Breakdown of Representation in American Politics. Norman, OK: University of Oklahoma Press.Google Scholar
Fiorina, Morris P., with Samuel J. Abrams and Jeremy C. Pope. 2011. Culture War? The Myth of a Polarized America. 3rd ed. New York, NY: Longman.Google Scholar
Goren, Paul. 2012. On Voter Competence. New York, NY: Oxford University Press.Google Scholar
Layman, Geoffrey C., and Thomas Carsey. 2002. “Party Polarization and ‘Conflict Extension’ in the American Electorate.” American Journal of Political Science 46 (October): 786–802.CrossrefGoogle Scholar
Layman, Geoffrey C., Thomas M. Carsey, John C. Green, Richard Herrera, and Rosalyn Cooperman. 2010. “Activists and Conflict Extension in American Party Politics.” American Political Science Review 104 (2): 324–346.CrossrefWeb of ScienceGoogle Scholar
McCarty, Nolan, Keith T. Poole, and Howard Rosenthal. 2006. Polarized America: The Dance of Ideology and Unequal Riches. Cambridge: MIT Press.Google Scholar
Miller, Gary, and Norman Schofield. 2003. “Activists and Partisan Realignment in the United States.” American Political Science Review 97 (2): 245–260.Google Scholar
About the article
Benjamin Highton is a Professor of Political Science at UC Davis who specializes in public opinion and voting behavior. A former APSA Congressional Fellow, he joined the Davis faculty in 1999. *Special thanks to Matthew Buttice, Ronald Rapoport, and Walter Stone for advice and comments on various aspects of this project.
Published Online: 2013-02-09
Among the best, all of which were exceedingly accurate in their predictions, were Simon Jackman (http://elections.huffingtonpost.com/2012/romney-vs-obama-electoral-map), Drew Linzer (votamatic.org), Nate Silver (http://fivethirtyeight.blogs.nytimes.com/), and Sam Wang and Andrew Ferguson (http://election.princeton.edu/).
Throughout, I focus on the margin of victory/defeat, which is the Democratic percentage of the vote minus the Republican percentage of the vote. The data from 2008 and earlier years are from Cook et al. (2010). The data for 2012 are from the Atlas of U.S. Presidential Elections (http://uselectionatlas.org/), downloaded on 11/14/2012. The resulting correlation is barely influenced by the extreme values for the District of Columbia (top right corner of Figure 1). With the DC included, the correlation is 0.983. With DC excluded, the correlation is 0.976.
In light of the seminal work of Anthony Downs (1957) that provides a model predicting party ideological and policy convergence in a two-party system, the existence of party divergence and its increase over time are especially interesting and have drawn scholarly attention. Explanations tend to focus on party activists, interest groups, and a host of features relating to long-term changes in the nature of American society (Miller and Schofield 2003; McCarty, Poole, and Rosenthal 2006; Schofield and Miller 2007; Aldrich 2011; Fiorina, Abrams, and Pope 2011; Bawn et al. 2012).
Importantly, the implication is not that voters care about cultural issues more or that cultural issues have been “primed” or become more salient – though both may be the case. Rather, if voters simply follow the underlying rule of responding to relative policy proximity, then change in behavior will be evident. Changes in voting behavior can be caused by changes in candidate issue locations even if voter preferences and their decision rules remain unchanged (Fiorina, Abrams, and Pope 2011).
Table A1 in the Appendix reports missing data rates by state. Forty states are missing none of the items, and six others are missing only one. Nebraska and Nevada are each missing the 21 GSS items. The District of Columbia has missing values for roughly one quarter of the items – the PSES and the 2006 CCES items. The largest amount of missing data is observed for Alaska (all the NAES items) and Hawaii (all the NAES items and one GSS item), each of which is missing a bit more than half the items.
The software and documentation are available on Gary King’s website (http://gking.harvard.edu/amelia/). Each instance of state-item missing data is imputed 500 times. For each missing item, I compute the mean value across the 500 imputations and substitute that value for the missing value in the subsequent empirical analyses. The reason for the large number of imputations (typically scholars rely on as few as five) is to assess the effects of the imputation process on the estimation of state issue locations. As demonstrated in the Appendix (Figure A1), they are minimal.
The issue measures are reversed coded from the how they are presented in Figure 2 so higher scores indicate more liberal positions. With a dependent variable of Obama vote margin, the estimated effects are expected to be positive.
All of the estimated issue effects across all of the estimated models can quite confidently be distinguished from zero (p<0.05).
The entries for 1972 are the correlations (with and without the District of Columbia) between the outcomes in 1972 and those in 1968. The entries for 1976 are for those between 1976 and 1972. And so on.
The assumption that state issue locations are fixed over the 40-year period (1972–2012) is probably incorrect, but it may be a reasonable approximation in light of Erikson et al. (2006), which finds substantial stability in the ideological positions of states (in an absolute sense and relative to the stability of state party identification). Further, Miller and Schofield (2003, 2008) treat state issue locations as fixed over a much longer (a century) period of time.
If one believes that the importance of cultural issues to voters has remained relatively constant, then the increase in cultural effects is due to a steady increase in the distance between the Democratic and Republican presidential candidates’ locations on cultural policy.
If one regresses the effect of state economic positions on an election year counter, there is an estimated increase of less than one percentage point per election suggesting hardly any long-term linear trend in the influence of economic policy locations on state presidential margins.
To be precise, the division of states is not between those that are predicted to be Democratic and those that are predicted to be Republican in a given year. The division is between those that are predicted to be more Democratic than the average state in a given year and those that are predicted to be more Republicans than the average state in a given year. This approach provides a common frame of reference across election years, which is akin to “normalizing” the presidential vote and treating differences in overall vote margins across years as resulting from short-term forces.
For each election year the cutting lines are based on the average of the six model estimates. When lines represent a group of elections in Figure 5 the lines are based on the average of the yearly estimates.
Apart from changes among and between party elites and others in the “political class” it is possible that the calculus of voters has changed, too. But as noted earlier the changes in the correlates of outcomes identified in this essay would be observed with changing parties and unchanging voters (Fiorina Abrams, and Pope 2011).