In 3 studies, subjects were asked to vote on all positively described candi
date vs. an opponent with a 1-sentence negative. In Study 1, the all-positi
vely described candidate was voted for more often, was rated higher on 10 o
f 13 semantic differentials, and generated more positive and less negative
thoughts. In Studies 2 and 3, the positively described candidate was Voted
for more, was rated higher on the semantic differentials, and generated mor
e positive and less negative thoughts in the regular negative condition. In
the condition in which an independent nonpartisan committee declared the n
egative groundless, there was no significant difference in any dependent va
riables. These theories have important theoretical implications related to
Johnson-Cartee and Garramone's theories.