A Feeling Person’s Game Affect and Voter Information Processing and Learning in a Campaign

We have developed and carried out a complex experiment using a dynamic information board (Lau and Redlawsk, 2001) in which subjects experienced a primary election campaign under varying conditions of emotional involvement and manipulation. In addition to manipulating potential emotional responses to candidates, we have developed measures to help us understand the implications of these emotional responses for voter information search, acquisition, and learning. Here we report on two basic analyses of these data. First, we want to look at the extent to which subjects can accurately recall their emotional responses to candidate information after the fact, and second we wish to establish just how “sticky” initial evaluations of candidates are in the face of countervailing information. We demonstrate that post-hoc measures of affective reactions are good, but not perfect, approximations of affect as it would be reported if measured immediately. The overall level of anxiety, whether through the anxiety threat manipulation or through congruency manipulation, increased the rate at which subjects incorrectly recalled having felt something when they did not report feeling that way immediately after viewing the information. Moreover, there were high levels of errors of recalling felt affect for all subjects. Finally, we find that affect has a non-linear effect on voting – by manipulating subject-candidate agreement, we are able to demonstrate a clear bias against rational updating.