1 Task Force on Canadian Unity, A Future Together (Ottawa: Supply and Services Canada, 1979), 104–06Google Scholar. For additional discussion see Irvine, William P., Does Canada Need a New Electoral System? (Kingston: Queen's University Institute of Intergovernmental Relations, 1979)Google Scholar; Elton, David and Gibbins, Roger, Electoral Reform: The Need is Pressing, The Time is Now (Calgary: Canada West Foundation, 1980)Google Scholar. This list is not exhaustive. For a more skeptical appraisal see Courtney, John C., “Reflections on Reforming the Canadian Electoral System,” Canadian Public Administration 23 (1980), 427–57.CrossRefGoogle Scholar

2 Transcript of an address by the Rt. Hon. Joe Clark to the 11th Annual Leadership Conference sponsored by the Centre for the Study of the Presidency (Ottawa, Oct. 19, 1980), mimeo, 13.

3 Meisel, John, The Canadian General Election of 1957 (Toronto: University of Toronto Press, 1957),258–62.Google Scholar

4 Meisel, John (ed.), Papers on the 1962 Election (Toronto: University of Toronto Press, 1964).Google Scholar

5 T. Peterson and I. Avakumovic, “A Return to the Status Quo: The Election in Winnipeg North Centre,” ibid., 91-106.

6 Morris Davis, “Did They Vote for Party or Candidate in Halifax?” ibid., 19-32.

7 Davis, Morris, “Ballot Behaviour in Halifax Revisited,” Canadian Journal of Economics and Political Science 30 (1964), 538–58.CrossRefGoogle Scholar

8 Davis, Morris, “A Last Look at Ballot Behaviour in the Dual Constituency of Halifax,” Canadian Journal of Economics and Political Science 32 (1966), 366–71.CrossRefGoogle Scholar

9 Though these were excluded from the analysis (which was addressing a different question) Courtney and Smith reported 1,075 who voted both federally and provincially, and who voted a single-party ticket provincially, 120 who split their votes provincially and 259 who were nonvoters in either or both elections. While we cannot infer exactly how many provincial voters there were among the last 259, among those who voted at both the federal and the provincial election, 18 per cent split their provincial vote. This is about the same as the percentage who changed their vote across levels. Courtney, John C. and Smith, David E., “Voting in a Provincial General Election and a Provincial Byelection,” Canadian Journal of Economics and Political Science 32 (1966), 344.CrossRefGoogle Scholar

10 Cunningham, Robert, “The Impact of the Local Candidate in Canadian Federal Elections,” this JOURNAL 4 (1971), 287–90.Google Scholar

11 Kornberg, Allan and Mishler, William, Influence in Parliament: Canada (Durham: Duke University Press, 1976), 88–89Google Scholar. For other studies stressing the importance of constituency service as a proclaimed objective, see among others: Hoffman, David and Ward, Norman, Bilingualism and Biculturalism in the Canadian House of Commons (Ottawa: Queen's Printer, 1970)Google Scholar, chaps. 4 and 5; Clarke, Harold and Price, Richard G., “Freshman MPs' Job Images,” this JOURNAL 13 (1980), 583–606.Google Scholar

12 Ibid.

13 Clarke, Harold D., Price, Richard G. and Krause, Robert, “Constituency Service among Canadian Provincial Legislators,” this JOURNAL 8 (1975), 520–42.Google Scholar

14 Price, Richard et al., “The Socialization of Freshman Legislators,” in Pammett, Jon and Whittington, Michael (eds.), Foundations of Political Culture (Toronto: Macmillan, 1976), 211–38.Google Scholar

15 In the October 1974 election, only 7 of 635 constituencies showed clear candidate impact. See Steed, Michael, “The Results Analyzed,” in Butler, David and Kavanagh, Dennis (ed.), The British General Election of October 1974 (London: Macmillan, 1975), 342–45.Google Scholar

16 Ferejohn, John A., “On the Decline of Competition in Congressional Elections,” American Political Science Review 71 (1977), 166–76;CrossRefGoogle ScholarCover, Albert D., “One Good Term Deserves Another,” American Journal of Political Science 21 (1977), 523–41.CrossRefGoogle Scholar Quite apart from the growth of political independence in the electorate, legislative parties in the United States have always been much weaker, and legislative committees much stronger than in Canada. Thus the potential for action by elected representatives is much larger in the United States. On this, see Kornberg, Allan and Wolfe, Judith D., “Parliament, the Media and the Polls,” in Clarke, Harold D., et al. (eds.), Parliament, Policy and Representation (Toronto: Methuen, 1980), 35–40.Google Scholar

17 Fiorina, Morris P., “The Case of the Vanishing Marginals,” American Political Science Review 71 (1977), 180.CrossRefGoogle Scholar

18 Cover, “One Good Term,” 5; Ferejohn, “On the Decline of Competition,” 169-72.

19 In a 1972 survey in Britain, 76 per cent of all constituents were unaware of any outside interest or affiliation of their MP and 47 per cent were unable to name their MP. Note, however, that these proportions are taken two years after the 1970 election and two years before the 1974 elections. Ivor Crewe, “Electoral Reform and the Local M.P.,” in Finer, S. E. (ed.), Adversary Politics and Electoral Reform(London: Anthony Wigram, 1975), 331.Google Scholar

20 By way of comparison 31 per cent of all respondents in a 1972 British survey felt unable to judge whether their MP was doing a good job (Ibid.).

21 These figures are obtained by adding the constants in each party's candidate column to the coefficient for the dummy variable for that party's identifiers. For example, the Progressive Conservative “base” of 25 points is obtained by adding the intercept of 12.55 to the Progressive Conservative coefficient of 12.82.

22 These results are analyzed in Crewe, “Electoral Reform and the Local M.P.” 317-42. While the last result might be understated because of the lapse of time since the last election, we would not expect the recollection of “helpfulness,” a noncampaign activity, to be similarly distorted.

23 Candidates, their parties and leaders were rated by respondents on “feeling thermometers” which ask them to indicate the degree of warmth they feel towards these objects on a 100-point scale.

24 Some of these questions were asked of only a random half-sample of the respondents, and failure to answer any one question excluded a respondent from the regression. For this reason, constituents of NDP and Social Credit MPs have had to be eliminated from Table 5 because of a small effective number of cases. Coefficients in Table 5 and in the discussion following all have statistically significant associated F-statistics, except where noted.

25 The beta coefficient measures the impact of a standard deviation unit change in the independent variable and allows us to compare effects for party and leader ratings, which are measured on a 100-unit scale, with that of “helping” measured as a dichotomy.

26 These figures are calculated by applying the beta coefficients to the standard deviations for the relevant variables.

27 This is worded cautiously for two reasons. First, it cannot be demonstrated that a superior or inferior result is entirely due to the candidate. It might be due to unmeasured but systematic effects. Secondly, in many (but not all) cases, a superior performance by one candidate is associated with an inferior performance by another. There is, therefore, some double-counting in Table 7 and it would be both invidious and presumptuous to assign blame. Moreover, there is internal evidence to suggest that many of these superior or inferior performances are, in fact, random, and subject to regression effects. Of 81 superior performances for some party in 1979, 50 showed negative residuals in 1980. Of 76 inferior performances in 1979, 49 had turned into positive residuals in 1980. In the interest of simplicity, these qualifications are omitted from the text but are important to keep in mind.

28 From the regression equations reported in Table 6, it is possible to estimate a “normal” voter shift from the province, educational level, homeownership rate and other characteristics of each constituency. Adding this normal shift to the 1974 vote for all leading contestants in each constituency allows us to decide which party would have won the constituency had all shifts been normal.

29 This figure is obtained by adding the overall constant for the 1979-1974 (dash to be read as “minus” not “to”) Progressive Conservative shift (–.037) to the coefficient for the Quebec dummy variable (–.095) for a basic decline of 13.2 points.

30 It is difficult to assess the extent of MP helpfulness from the survey data. For the 1979 survey, 96 constituencies (using 1974 boundaries) were sampled with, typically, between 14 and 51 respondents per constituency, except for Prince Edward Island constituencies which had only 3 to 6 respondents. In all, 12 constituencies had more than 39 respondents, 48 had more than 29 and 14 had fewer than 20. The most active MP was credited with service by 14 of his 29 constituents. Because of the low numbers, the constituency proportions must be taken with some caution but, on their face, the data indicate that only 8 of the 96 MPs reached as many as one-third of their constituents. This is the same proportion found for incumbents contesting the 1974 election and indicates the magnitude of the task facing any MP whose party is losing popularity.

31 Jon Pammett, “Public Evaluations of the Canadian Federal System, 1974-80,” paper delivered at the Canada-India Workshop on Federalism, Hyderabad, India, 1981.