Document Type


Peer Reviewed


Publication Date


Journal/Book/Conference Title

American Political Science Review

DOI of Published Version


Start Page


End Page



This article explores two dimensions of public evaluations of presidential candidates on the basis of open-ended survey questions from 1952 to 1980. The first dimension looks at whether citizens evaluate candidates on the basis of policies, performance, or strictly candidate attributes; the second examines the time perspective of these assessments, that is, whether they are retrospective or prospective. It is found that incumbents have been judged primarily on the basis of retrospective performance, challengers on prospective policy, and candidates running in nonincumbent races on prospective performance. Throughout the period from 1952 to 1980 both policy and performance considerations have become increasingly related to the vote. Except for 1964, performance has outweighed policy as a predictor of the vote, with an emphasis on retrospective evaluations whenever an incumbent runs for reelection and on prospective assessments in nonincumbent races. The 1964 case provides the best example of a policy mandate, with the 1972 election also fitting the pattern to a lesser degree. The data for the 1980 election, however, fail to support the claim of a mandate for Reagan's policy stands.

Journal Article Version

Version of Record

Published Article/Book Citation

American Political Science Review, 79:2 (1985) pp. 359-372.


Copyright © 1985 American Political Science Association. Used by permission.