¹ Kornberg, Allanet al., “Some Differences in the Political Socialization Patterns of Canadian and American Party Officials: A Preliminary Report,” this JOURNAL, II, no. 1 (March 1969), 74Google Scholar; Winham, Gilbert R. and Cunningham, Robert B., “Party Leader Images in the 1968 Federal Election,” ibid., III, no. 1 (March 1970), 51–5Google Scholar; Kamin, Leon, “Ethnic and Party Affiliations of Candidates as Determinants of Voting,” Canadian Journal of Psychology, XII (Dec. 1958)Google Scholar, and in Courtney, John C., ed. Voting in Canada (Scarborough, Ont. 1967), 191–8Google Scholar; Scarrow, Howard A., “Three Dimensions of a Local Political Party,” in Meisel, John, ed. Papers on the 1962 Election (Toronto, 1964), 53–67Google Scholar; Engelmann, F. C. and Schwartz, M. A., Political Parties and the Canadian Social Structure (Scarborough, 1967), 225Google Scholar; Meisel, John, “Religious Affiliation and Electoral Behaviour: A Case Study,” Canadian Journal of Economics and Political Science, XXII, no. 4 (Nov. 1956), 481–96CrossRefGoogle Scholar, and in Courtney, Voting in Canada, 145–61; Anderson, Grace M., “Voting Behaviour and the Ethnic-Religious Variable: A Study of a Federal Election in Hamilton, Ontario,” Canadian Journal of Economics and Political Science, XXXII, no. 1 (Feb. 1966), 27–37.CrossRefGoogle Scholar

² One reason for this may lie in the fact that researchers wish to generalize to as large a population as their resources and sample will allow. Therefore the research effort is spread over many ridings. Consequently, the sample size for each individual riding is usually too small to be sensitive to individual differences in attractiveness among candidates. The present study, which focuses on three ridings, provides information about local candidate influence within a limited geographical area. The generality of the findings must await further research in other (probably limited) geographical areas.

³ Hanham, H. J., Elections and Party Management (London, 1959), 199Google Scholar, cited in Davis, Morris, “Ballot Behaviour in Halifax Revisited,” Canadian Journal of Economics and Political Science, XXX, no 4. (Nov. 1964), 554.Google Scholar

⁴ “Did They Vote for Party or Candidate in Halifax?” in Meisel, Papers on the 1962 Election, 21. The method of arriving at this figure is interesting. Straight ticket voting indicates party voting; while ticket splitting indicates candidate orientation; Davis’ method probably leads him to underestimate the importance of local candidates.

⁵ For more detail on the sampling method, see Winham and Cunningham, “Party Leader Images,” 38–9.

⁶ The obvious drawback is that concentration on any three ridings renders the findings questionable on the grounds of uniqueness. With a national sample, on the other hand, local candidate influence can be compared to party leader, party identification, ethnicity, religion, region, or class variables. However, in most national samples there are too few respondents from any single riding to compare the relative strengths of various local candidates. Consequently, inter-candidate variation cannot be ascertained, and the question as to whether one candidate rather than another makes a difference cannot be answered.

⁷ Previous studies indicate party identification to be much stronger than shown here; see fn. 1 above.

⁸ Had there been no significant deviations from the overall mean, the necessary conclusion would have been that the influence of local candidates, while important in pulling votes, is randomly distributed. Although different candidates may appeal to different people, the pull of any one candidate is no stronger or weaker than the pull of any other candidate. Consequently, for electoral victory one nominee is just as good as any other.

⁹ Those party identifiers who intend to vote for their party's candidate are removed from the analysis. Underlying this procedure rests the assumption that evaluation of the various candidates is highly dependent upon the party identification of the voter; that is, party identifiers perceive more favourably candidates of their own party than candidates of other parties.

¹⁰ It can be argued that the personal attractiveness of incumbents is underestimated by this method of calculating candidate influence. The incumbent has created party identifiers of his personal followers, thereby reducing his score on influencing independents and those who cross party lines both absolutely by cutting down numbers and relatively by increasing the size of the identifier group. Especially salient in this study is the case of John Munro, an MP generally conceded to have considerable candidate attractiveness.

¹¹ This 10 per cent minimum candidate pull needs further testing in various geographical areas across several elections. The 10 per cent figures does not seem high for the appeal of Trudeau likely depressed the salience of local candidates generally in the 1968 election.