Design of CES 1997

CES 1997 will employ four complementary methods of data collection (1) a campaign wave rolling cross-section survey, a post-election cross-section survey, and a mailback survey (2) semi-structured interviews with selected respondents (3) interviews with party strategists (4) media tracking of party advertising and TV news coverage.

The CES 1997 Pre-Election Rolling Cross-Section Survey

The campaign-wave survey will be designed to track dynamic influences on vote choice. The rolling cross-section design involves breaking the overall sample down into daily replicates so that all that distinguishes one replicate from another (within the range of sampling error) is the passage of time. This provides a remarkably powerful design for relating evolving campaign rhetoric and campaign events to opinion dynamics. Interviewing for the rolling cross-section will begin no later than four days after the election writs are issued and will terminate at the end of the last day before the election. The sample will consist of 3,800 randomly selected respondents. For the first time, the CES sample will include respondents from the Yukon and the Northwest Territories. The survey will employ computer-assisted telephone interviewing (CATI), thus allowing question-wording and question-order experiments to be incorporated into the design of the survey instrument. Interviews will be conducted in both official languages and will last approximately 35 minutes.

The CES 1997 Post-Election Survey

The post-election reinterview makes it possible to examine the impact of opinions and perceptions cross-sectionally, free of the confounding effects of campaign events. Interviewing for the post-election survey will commence on the day following the election, with the goal of reinterviewing as many of the campaign-wave respondents as possible as close to election day as possible. The post-election interviews will again employ CATI and will last approximately 20 minutes.

The CES 1997 Mailback Survey

A self-administered mailback questionnaire will be sent to all respondents who consent on completion of the post-election interviews. The questionnaire will contain approximately 100 items. The mailback survey is designed to measure voters' fundamental values and ideological beliefs.