Key points are not available for this paper at this time.
Response latencies to all questions in a computer-assisted telephone interviewing (CATI) survey of the 1990 Ontario election were recorded. Prior to the election, a random sample of respondents were asked about their voting intention, their leader preference, their party preference, and their customary party identification. After the election, respondents were called back to determine if they had voted and, if so, for what party. Models of vote choice demonstrated that the voting behavior of respondents who expressed their voting intention quickly in the preelection interview was much more predictable than the voting behavior of respondents who expressed their voting intention slowly. The accessibility of voting intentions was, in turn, related to conflict among determinants of vote choice. Moreover the accessibility of voting intentions was related to the accessibility of party identification, but only for respondents who did not experience a conflict in the election.
John N. Bassili (Sat,) studied this question.