We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 10 November 2009
Given the Liberal party's dominant position in Canadian federal politics, it has been extremely difficult for the Progressive Conservatives to effect an electoral breakthrough and win enough parliamentary seats to form a government. In this century, the Conservatives have been in power for only 22 years. The last time the party formed a majority government was in the spectacular Diefenbaker victory of 1958 which gave the Conservatives 208 of 265 seats. After the demise of the Diefenbaker government in 1963, the Conservatives languished in opposition until 1979 when they won a plurality of seats and Joe Clark formed his short-lived government. It is the election of 1979 which provides the focus of analysis in this article, and especially the success of the Conservatives in supplanting the Liberals, although not with a majority government.
1 In this study a national probability sample (N=2,744) of the 1979 electorate was interviewed immediately after the election. Included in the survey were 1,295 respondents originally interviewed as part of the 1974 national election study. The 1974 survey design is described in Clarke, Harold D., Jenson, Jane, LeDuc, Lawrence and Pammett, Jon, Political Choice in Canada (Toronto: McGraw-Hill Ryerson, 1979Google Scholar), Appendix A, 397–400. Details regarding the 1979 survey design are available from the authors upon request. These studies were supported by the Social Science and Humanities Research Council of Canada. All analyses and interpretations are the responsibility of the authors.
2 Political Choice in Canada, passimGoogle Scholar.
3 Ibid., 148.
4 These figures are based on respondents’ recalled voting behaviour and party identification. Analyses of the 1974–1979 panel data also suggest considerable instability in both party identification and voting. For example, if one analyzes panel members voting in at least one of the 1974 and 1979 elections, 65 per cent have the same identifications in 1974 and 1979, 18 per cent have different identifications, 14 per cent have moved to or from the status of nonidentifier, and 3 per cent are nonidentifiers in both years. Sixty per cent report voting for the same party in 1974 and 1979, 22 per cent report switching their vote, and 18 per cent voted in only one of the two elections. Altogether, 48 per cent report stable party identifications and stable voting behaviour.
5 Political Choice in Canada, 309–16Google Scholar.
6 Ibid., 315.
7 Ibid., 343–48.
8 Conceptually, this treatment differs from a truly “permanent” pool of eligible voters to the extent that attrition due to death, disability or emigration must be taken into account in analyses extending over a substantial period of time. Between a pair of elections, however, such effects will be relatively small and can reasonably be disregarded in an examination of electoral turnover. Such effects should not be overlooked in analyses of long-term electoral trends. Butler and Stokes, for example, have demonstrated the importance of demographic variables in accounting for the decline of the Liberal party in Britain in the early part of this century and the concomitant rise of the class alignment in modern British politics. See David, Butler and Donald, Stokes, Political Change in Britain (2nd ed.; New York: St. Martin's, 1976Google Scholar), chaps. 7, 8 and 9.
9 Because the treatment here is confined to the 1974–1979 pair of elections, transient voters not voting in both of these elections are excluded from the analysis. For a general analysis of the attitudes and behaviour of transient voters see Political Choice in Canada, 309–16Google Scholar.
10 Chief Electoral Officer, Report, 1979 (Hull: Ministry of Supply and Services Canada, 1980Google Scholar), x.
11 The number of new voters is calculated from the Census of Canada 1976, v. 8, bulletin 8SD.1, catalogue 92–832, 1–3.
12 Chief Electoral Officer, Report, 1979, xiGoogle Scholar.
13 An increased turnout of 5 per cent among surviving 1974 voters would have brought an additional 646,000 of these voters to the polls in 1979, or 34 per cent of the 1.9 million increase in the total vote. Estimated from Census of Canada 1976, ibid., and Chief Electoral Officer, Report, 1974, ix.
14 See, for example, Allan, Kornberg, Clarke, Harold D. and Arthur, Goddard, “Parliament and the Representational Process in Contemporary Canada,” and Allan, Kornberg and Wolfe, Judith D., “Parliament, the Media and the Polls,” in Clarke, Harold D., Campbell, C., Quo, F. and Goddard, A. (eds.), Parliament, Policy and Representation (Toronto: Methuen, 1980Google Scholar), chaps. 1 and 3.
15 The feeling thermometer scales range from 0–100, with 50 explicitly designatedas the neutral point. See Political Choice in Canada, 406–07Google Scholar.
16 We must treat these results with some caution, however, for two reasons. First, for many respondents the score on one variable may be affected by the other: for example, if someone likes Trudeau that person might give the Liberal party a high score as well for that reason, thus causing Trudeau to appear similar to the party on the thermometer ratings. Second, a high score for a leader may not necessarily translate into votes for his party if the party itself is particularly disliked, or if the voter is making the decision without reference to leader images.
17 Political Choice in CanadaGoogle Scholar, chap. 12.
18 ibid., 243–44; Buter, and Stokes, , Political Change in BritainGoogle Scholar, chap. 13. The distinction between “position” and “valence” issues first appeared in Stokes, Donald E., “Spatial Models of Party Competition,” in Angus, Campbell, et al. (eds.), Elections and the Political Order (New York: Wiley, 1966), 170–71Google Scholar.
19 Political Choice in Canada, chap. 8.
20 Respondents were asked “Now I would like to ask you some more specific questions about the recent [1979] federal election. What, in your opinion, was the most important issue in that election?” Respondents who cited an issue then were asked three follow-up questions: (a)“How do you feel about the issue?” (b)“Which party is closest to you on this issue?” and (c)“How important was that issue to you in deciding how to vote in the election?” These respondents then were questioned about whether or not another election issue was important to them and the remainder of the sequence [(a)-(c)] was repeated for those citing such an issue.
21 Ibid., 246.
22 This assumes voters consider inflation and other issues in isolation from one another. However, if voters perceive inflation as being tied to other issues (for example, levels of unemployment), then depending upon how these issues are presented to the electorate in the context of a specific election campaign, inflation may take on positional qualities. On this point see Alt, James E., The Politics of Economic Decline: Economic Management and Political Behaviour in Britain Since 1964 (New York: Cambridge University Press, 1979), 8–12.Google Scholar
23 A Gallup survey on the mortgage deductibility question found 74 per cent in favour, 15 percent against, and 11 percent with no opinion. Canadian Institute of Public Opinion, The Gallup Report (December 27, 1978).
24 Additional evidence in support of this conclusion is provided by individual-level comparisons of leader popularity. Pairwise comparisons of voters thermometer ratings of Trudeau and Clark show that 53 per cent preferred the former, 38 per cent the latter, and 9 percent rated them equally. For Trudeau and Broadbent, comparable percentages are 48 per cent, 38 per cent, and 13 per cent respectively. Indicative of Clark's relative unpopularity is the finding that when compared to Broadbent, 48 per cent of the electorate gave higher thermometer scores to the NDP leader, whereas Clark was favoured by 38 per cent (15% rated them equally).
25 The question asked was“In deciding how you would vote in the recent 1979 federal election, which was the most important to you: the party leaders, the candidates here in this constituency, orthe parties taken as a whole?” For those citing party leaders or local candidates the follow-up question was:“When you say that_____was the most important to you, are you thinking of this perso's personal qualities or this perso's stand on certain issues?” For persons mentioning party the follow-up question was: “When you say that party was the most important to you, are you thinking of the part's general approach to government or its position on certain issues?” Those in either question sequence who cited“issues” were asked:“Which issues are you thinking of specifically?
26 These percentages are very similar to those obtained in the 1974 survey. In 1974, 49 per cent (58% of the“party leader,” 48% of the“local candidate” and 43% of the“parties as a whole” streams) of the respondents cited an issue basis for their answers regarding the“most important factor” in their voting choices, and 51 percent selected the“personal qualities” or“general approach” options.
27 A useful discussion of the problems such effects pose for analyzing voting behaviour is Richard A. Brody and Benjamin I. Page, “The Assessment of Policy Voting.” American Political Science Review 66 (1972), 450–58Google Scholar.
28 Political Choice in Canada, chaps. 10 and 11.
29 Ibid., 343–48.
30 This method of multiple regression analysis is known as the analysis of commonalities. For a brief nontechnical discussion of this technique see Nie, N.Verba, S. and Petrocik, J., The Changing American Voter (Cambridge: Harvard University Press, 1976), 303Google Scholar, note 8.
31 In 1974, issue effects were greater than those of party leaders among flexible-high interest partisans in the analyses of voting for all three parties. Leader effects were greater than issue effects among flexible-low interest partisans in every case (Political Choice in Canada, 346). In this respect, the 1979 results appear to be a mirror image of those for 1968. In that election, leader effects consistently exceeded those for issues among all categories of flexible partisans (ibid., 354–55).
32 Multivariate stepwise logit analyses yield results very similar to those produced by the regression analyses. The logit procedure shows that the impact of short-term leader and issue forces was almost wholely confined to flexible partisans; that issue effects were invariably greater than those for leader images among all political interest categories of flexible partisans: and that leader effects were, relatively speaking, greater in analyses where Liberal rather than Conservative or NDP voting was the dependent variable. For a description of the logit procedure used see the SAS Supplemental Library Use's Guide, 1980 edition (Cary, North Carolina: SAS Institute, 1980), 83–102Google Scholar.
33 Political Choice in Canada, 373–80Google Scholar.
34 Ibid., 367.
35 For data on the partisanship of various age cohorts see Lawrence LeDuc,“Sources of Long-Term and Short-Term Electoral change in Canada,” paper presented at the Conference on Critical Electoral Forces: Changing Mass Politics in Advanced Industrial Societies, Florida State University, Tallahassee, Florida, May 23–24, 1980, 11–15.
36 Political Choice in Canada, 374–75Google Scholar.
37 By region, the percentage of respondents mentioning one of these issues as most important was: Atlantic (13), Quebec (25), Ontario (12), Prairies (15), BC (12).
38 Of course, the pattern is partly accounted for by the combination of the greater salience of this issue in Quebec and the strong tendency of Quebec voters to support the Liberal party. However, more detailed analyses show that the Confederation issues were basically favourable to the Liberals in all parts of the country.
39 George, Perlin, The Tory Syndrome: Leadership Politics in the Progressive Conservative Party (Montreal: McGill-Quee's University Press, 1980)Google Scholar.
