Political Leapfrogging

From the 'Editor's Choice' section of the Nov. 26 Science Magazine:

Although there have been many discussions of the polarized nature of American politics, do the views of elected officials match the preferences of their electorate? Bafumi and Herron sought to answer this question by comparing a national opinion survey of American voters (the Cooperative Congressional Election Study; CCES) with legislator voting records of the 109th (2005–2006) and 110th (2007–2008) Congresses. In many cases, the CCES questions were similar to (or the same as) actual congressional roll call votes, which allowed for better comparison. By developing a linear scale bounded by representatives (or CCES respondents) who had taken consistently liberal or conservative positions, the authors found that members of Congress were more extreme than the voters they represented. The median member of the 109th House of Representatives was more conservative than the median American voter, but the median member of the 110th House of Representatives was more liberal. Thus, voting out one extremist usually led to replacement by someone equally extreme, but of the opposite party. The authors refer to this as “leapfrogging” because the moderate views of the median American voter are leapfrogged during the turnover. Although the turnover was similar in the Senate, overall it appeared to be more moderate.

The Article: Amer. Polit. Sci. Rev. 104, 519 (2010).