Empirical foundation of valence using Aldrich–McKelvey scaling

This paper uses data from the 2004 pre-election survey of American National Election Study to test empirically different ways incorporating a valence parameter into Downsian utility function. We call particular attention problem interpersonal incomparability responses liberal-conservative scale, and use Aldrich–McKelvey’s pathbreaking method obtain accurate distances between respondents candidates, key regressors. find that function most supported by empirical evidence, intensity function, is one which permits make better predictions for presidential election. also consider counterfactual analyses wherein we if Bush, candidate with highest valence, has dominant strategies would have insured him majority popular vote. According theory, it known does not such distribution voters in policy space too heterogenous. Nevertheless, show sufficiently homogenous Bush strategies.