Bartels: Partisanship and voting behavior, 1952-1996

Bartels. 2000. Partisanship and voting behavior, 1952-1996. AJPS.

In Brief

Objective: To analyze the extent of partisan voting in American elections from 1952-1996.

Conclusion: While partisan voting decreased from 1964-1976, the last six presidential elections have shown increased partisanship. The level of partisanship in presidential elections has increased significantly from 1976-1996 and has been accompanied by an increase at the congressional level.

Place in Literature: Against analysts like Wattenberg (1996) and Burnham (1989), who argue that partisanship has declined steadily since the 1960's.

Methodology: Regressions of ANES survey data on partisanship and voting behavior in period from 1952-1996.

Findings

For decades, political scientists have pointed to the fact that American voters were increasingly less likely to cast votes on the basis of partisanship. Bartels examines voting data from 1952-1996 and finds the following:

• The proportion of "strong party identifiers," after decreasing from 1964-1976 has increased significantly since then.
• While "identifiers" have decreased in the electorate as a whole since 1964, the proportion among voters has increased since reaching its low in 1976 and now matches the level of partisanship of the 1950's elections.
• While the decreasing partisanship trend among all eligible voters is significant, when only voters are taken into account, these differences disappear.

Discussion

Using probit regression to measure the effects of partisanship on votes, Bartels finds a steady increase from 1972 to 1996 and shows that the 1996 levels are 77% higher than in 1972 and 15-20% higher than the levels of the 1950's. More than any time within the past 50 years, voters are voting in correlation with their partisan preference. While this is based in part on changes in Southern voting patterns, Bartels notes that it is evident in a wide variety of subgroups in the electorate. He finds similar (but not as significant or steady) increases at the Congressional level.

While this measure has varied through elections, it has increased significantly from 1978-1996. Bartels provides evidence that this increase is due to an increased sense of differentiation among the parties in the era of Reagan, Newt Gingrich (and Bill Clinton). He also ties it to the changing voting patterns of Congressmen and other politicians in the wake of changes in the composition of parties. As the Democratic party in the South collapsed, there were increasingly strong distinctions between the parties. These changes have been noticed by NES respondents at increasing levels since 1980.