INCUMBENCY, PARTISAN, AND SALIENCY EFFECTS IN CONGRESSIONAL ELECTIONS FROM 1956-1978

[摘要] The present state of knowledge concerning Congressional elections involving incumbents runs something like the following. After 1964 a great decline in marginality occurred within contested incumbency districts. This decline related primarily to changes in individual voting patterns, especially among members of the challenger's party, who since 1972 have defected to incumbents at rates equal to or greater than their votes for their own party's candidate. However, this increased pro-incumbency voting was not accompanied by an increased incumbent visibility advantage over challengers. Moreover, our one visibility measure seems to have been defective because voters recalling neither candidate still strongly supported the incumbent. But name recall saliency being the only individual level measure over the period, the reasons for the rise in incumbent advantage amongl voters have become anybody's guess. The best guess perhaps is that increased constituency service aby incumbents induced a deeper impact among those to whom they were visible.This dissertation asserts that while the decline of marginal incumbent districts is hardly to be questioned, many claims about individual level effects surrounding the incumbent phenomenon (if it is a phenomenon) are mostly mythical. Pro-incumbency voting has been severely overestimated. The incumbent-challenger saliency gap among voters has never been introduced and all evidences of name recall available in the SRC-CPS Surveys have not been tapped. In establishing a methodology providing continuity across surveys, we discover that individual level pro-incumbent voting has risen at a more moderate pace than presently claimed, that the incumbent-challenger name recall gap has opened up over time due primarily to a severe decline in challenger saliency, that after controlling for several factors, name recall saliency is closely related to the changes which have occurred, and that when all evidences of name recall are tapped, those recalling neither candidate demonstrate the expected patterns, i.e., they favor neither incumbent or challenger. Among both partisans and independents only those recalling both candidates demonstrate a clear extra-salient effect. The relationship between those recalling both cadidates to those recalling only the incumbent appears to be especially important to developments within the so-called ;;period of change;; in Congressional elections. Reasons for the pro-incumbent bias in the 1978 survey are also examined, going beyond the speculated nature of districts or voters sampled to focus on a proposed survey-induced bias.In conclusion, reasons for the decline of challengers, focussing on definite legislative acts rather than constituent service, are discussed with implications for politics, policy, and polity articulated.

[发布机构] Rice University

[效力级别] science