The Impact of ‘Third Party ’ Advertising in Canada: An Exploratory Ecological Analysis of the 2000 Election

This paper presents an exploratory ecological analysis of the effectiveness of third party intervention in the 2000 federal election campaign in Canada. It uses the total election-related expenses reported by interest groups in local riding contests as revealed in their official reports to Elections Canada. As such, the analysis ignores the impact of third party spending at geographic levels above the local constituency, and therefore it constitutes only a partial test for the impact of these efforts. However, it offers the first (to my knowledge) multivariate test for effects of constituency-specific third party spending. Preliminary analysis reveals that third parties intervened in a relatively modest number of constituencies during the 2000 campaign. In most cases, where they did intervene in the campaign the level of spending was reasonably modest – certainly in comparison to the total spending by candidates. Yet these interventions were more likely in settings where the race was eventually decided by a relatively small margin of the vote, suggesting that the decision of groups to intervene represented in part a rational calculation concerning the likelihood that their spending might alter the outcome of the local race. Multivariate ecological models show that third party spending significantly influenced the vote shares of Canadian Alliance and Bloc Québécois candidates in 2000, net of other socio-economic and political characteristics. Given the crudeness of the measure of third part