Economic performance and U.S. Senate elections: Reply

Major criticisms made by Chressanthis and Shaffer (C-S) are addressed: the authors' failure to include noneconomic variables in the regression model, the inclusion of open elections in the sample, and the use of variables other than the state unemployment rate as economic performance measures. Errors in C-S's interpretations are noted. Additional regression results are reported using modifications of the authors' original model. These suggest that the change (notrelative change as in C-S) in state unemployment rate is superior, and that the authors' model yields a stronger effect of state economic performance on election outcomes than does the C-S model.John Beck contributed helpful comments.