CANDIDATE PERCEPTION AMONG NOMINATION ACTIVISTS - A NEW LOOK AT THE MODERATION HYPOTHESIS

The moderation hypothesis, that ideologically moderate candidates in a two-party contest are more likely to win elections than extreme candi dates, is both prevailing wisdom and a major conclusion of spatial the ories of elections. We examine the hypothesis using perceptual data fr om samples of caucus attenders in the 1984 and 1988 presidential nomin ation campaigns at both the individual and the cross-candidate levels of analysis. In both analyses, we find qualified support for the moder ation hypothesis. In explaining individuals' perceptions of the four n ominees' electability, we find a modest effect of perceived proximity to the American voter for all candidates, save George Bush. Other vari ables, such as party, candidate affect, and nomination chances are str onger predictors, and the candidate's perceived abilities on TV rivals moderation in its impact. In our comparative analysis of some 20 cand idates, we find that Ronald Reagan and John Glenn were very significan t outliers in what is otherwise a strong relationship between moderati on and electability. We also find that performance on TV is a strong a nd significant predictor of electability. We conclude that the moderat ion hypothesis holds up reasonably well for the large majority of cand idates, but that other candidate factors are also important to include in any complete assessment of candidate electability in November.