A Tale of Three Variables: Exploring the Impact of Alternative Measures of Presidential Approval On Congressional Voting

This study explores the impact of three distinct measures of public approval on congressional voting. Specifically, we test for the relative impact of a president's national approval, his partisan approval, and his district- (or state-) level approval on congressional support, measured at the level of individual members of Congress. Though we remain consistent with other arguments that hold presidential approval is likely to act as a background variable rather than a determinant of congressional voting, we maintain that theoretically more meaningful relationships between public opinion and congressional voting can be developed. Specifically, we argue that a more relevant test can be made which includes opinion measured closest to where that opinion is likely to matter to members, namely among electoral copartisans and district-level constituencies. We estimate various models (bivariate and integrated multivariate), and find strong support for our hypotheses that, to the extent members of Congress use public approval as a voting cue, they do in fact pay far more attention to partisan and constituency interests than they do to national opinion.