This item is by Alan Abramowitz, who is Alben W. Barkley Professor of Political Science at Emory University and a member of the TDS advisory board.
In a recent article posted at The Democratic Strategist I presented evidence that conservatism had a negative influence on the performance of Republican incumbents in recent U.S. Senate elections. In this article I examine whether liberalism had a similar effect on the performance of Democratic incumbents. I collected data on all contested Senate races involving incumbents in five elections between 2000 and 2008. A total of 69 races involving Republican incumbents and 61 races involving Democratic incumbents were included in the analysis.
Overall, Democratic incumbents were much more successful than Republican incumbents in these five elections. Only 6% of Democratic incumbents (4 of 61) were defeated compared with 25% of Republican incumbents (17 of 69). This occurred despite the fact that Republican incumbents generally represented states that more strongly supported their party’s presidential candidates than did Democratic incumbents. On average, Republican presidential candidates received 54% of the vote in Republican incumbents’ states while Democratic presidential candidates received an average of only 50% of the vote in Democratic incumbents’ states.
The difference between the success rates of Democratic and Republican incumbents was especially striking in states in which the incumbent’s party was at a disadvantage or enjoyed only a marginal advantage based on the presidential vote. In strongly Democratic and Republican states, those in which the vote for a party’s presidential candidate was at least five points above the party’s national vote share, almost no incumbents lost: 94% of Republican incumbents (33 of 35) and 100% of Democratic incumbents (23 of 23) were reelected. However, in high risk or marginal states, those in which the vote for a party’s presidential candidate was below or less than five points above the party’s national vote share, 44% of Republican incumbents (15 of 34) were defeated compared with only 11% of Democratic incumbents (4 of 38).
In order to explain the difference in the performance of Republican and Democratic incumbents, I performed separate multiple regression analyses of the results of elections involving each party’s incumbents. The dependent variable in each regression analysis is the incumbent’s share of the major party vote. The independent variables are the vote share for the presidential candidate of the incumbent’s party in the state and the ideology of the incumbent, as measured on the DW-NOMINATE scale, as conservatism for Republicans and liberalism for Democrats. (Dummy variables for each election year were also included to control for the national political climate at the time of each election but the coefficients for these variables are not included in the table). The results of the regression analyses are displayed in Table 1.
The findings in Table 1 once again show that among Republican incumbents, conservatism had a substantial and statistically significant negative influence on vote share after controlling for the strength of the Republican Party in each incumbent’s state and the national political climate at the time of each election. In contrast, among Democratic incumbents, liberalism had only a small and statistically insignificant influence on vote share.

These results suggest that one of the reasons why Republican incumbents from marginal and Democratic leaning states suffered an exceptionally high rate of defeat is that some of them were too conservative for the states that they represented. In these five elections, 58% (7 of 12) of strongly conservative Republican incumbents (those with above average conservatism scores) from marginal or Democratic leaning states were defeated compared with only 36% (8 of 22) of less conservative Republican incumbents from these states.