1 Clarke, Harold D., Jenson, Jane, LeDuc, Lawrence, Pammett, Jon H., Political Choice in Canada (Toronto: McGraw-Hill Ryerson, 1979), 291–97Google Scholar. In this essay the terms canvassing and contacting are used interchangeably, even though it is appreciated that the two may have different connotations. In particular, canvassing connotes party activity pertaining to the identification of its support, both actual and potential, and therefore is mostly a prelude to “getting out the vote” on election day. (Of course, it also serves other purposes as well, such as providing feedback for the campaign managers.) On the other hand, individuals may be “contacted” without providing such information to the parties, for example, simply by receiving party literature in their mail boxes. However, the available data do not lend themselves to such subtle distinctions and, hence, the alternative usage of both terms.

2 The data set in question is the 1974 Canadian national election study, carried out by Clarke and his associates and made available through the Inter-University Consortium for Political and Social Research. Neither the original investigators nor the Consortium bear any responsibility for the analysis or interpretations presented here.

3 Apart from Clarke, et al., only Morton has written on the topic in a scholarly way. His interest was with the impact of canvassing efforts by the NDP in the 1967 Ontario election campaign. However, working only with election return data, the knowledge of the party's efforts in different constituencies and, importantly, lacking the ability to control for pertinent variables, Morton was unable to reach any firm conclusions (Morton, Desmond, “The Effectiveness of Political Campaigning: The NDP in the 1967 Ontario Election,” Journal of Canadian Studies 4 [1969], 21–33)CrossRefGoogle Scholar. By contrast, somewhat more attention has been paid to the impact of the local candidate in Canadian elections, a topic that obviously could be related to the one here. However, it is possible to separate the two analytically since, for example, the candidate may gain or lose votes independently of canvassing efforts. To ensure that the separation occurred empirically, affect towards the local candidate, that is the candidate qua candidate, was controlled for when assessing canvassing effects. This included the canvassing effect associated with the local candidate. On the subject of the impact of the local candidate see, for example, Wilson, John, “The Myth of Candidate Partisanship,” Journal of Canadian Studies 3 (1968), 21–31Google Scholar, and Cunningham, Robert, “The Impact of the Local Candidate in Canadian Federal Elections,” this JOURNAL 4 (1971), 287–90;Google Scholar for a study with a somewhat different slant on the subject, see William P. Irvine, “Does the Candidate Make a Difference? The Macro-Politics and Micro-Politics of Getting Elected,” a paper delivered at the annual meeting of the Canadian Political Science Association, Halifax, 1981.

4 Some of the more prominent American studies include (in chronological order): Gosnell, H. F., Getting Out the Vote: An Experiment in the Stimulation of Voting (Chicago: University of Chicago Press, 1927);Google ScholarEldersveld, Samuel J. and Dodge, Richard. W., “Personal Contact or Mail Propaganda? An Experiment in Voting Turnout and Attitude Change,” in Katz, Daniel, et al., (eds.), Public Opinion and Propaganda (New York: The Dryden Press, 1954), 532–42;Google ScholarEldersveld, Samuel J., “Experimental Propaganda Techniques and Voting Behavior,” American Political Science Review 50 (1956), 154–65;CrossRefGoogle ScholarKramer, Gerald H., “The Effects of Precinct-Level Canvassing on Voting Behavior,” Public Opinion Quarterly 34 (1970), 560–72;CrossRefGoogle ScholarBlydenburgh, John C., “A Controlled Experiment to Measure the Effects of Personal Contact Campaigning,” Midwest Journal of Political Science 15 (1971), 365–81;CrossRefGoogle Scholar and Bartell, Ted and Bouxsein, Sandra, “The Chelsea Project: Candidate Preference, Issue Preference, and Turnout Effects of Student Canvassing,” Public Opinion Quarterly 37 (1973), 268–75.CrossRefGoogle Scholar British studies include: Butler, David and Stokes, Donald, Political Change in Britain (New York: St. Martin's Press, 1969), 421–27Google Scholar, especially, 421–24, and Bochel, J. M. and Denver, D. T., “Canvassing, Turnout and Party Support: An Experiment,” British Journal of Political Science 1 (1971), 257–69CrossRefGoogle Scholar. Of related interest, see Taylor, A. H., “The Effect of Party Organization: Correlation Between Campaign Expenditure and Voting in the 1970 Election,” Political Studies 20 (1972), 329–31CrossRefGoogle Scholar, and Denver, D. T. and Hands, H. T. G., “Marginality and Turnout in British General Elections,” British Journal of Political Science 4 (1974), 17–35CrossRefGoogle Scholar.

5 Such a conclusion was sounded most clearly in Kramer (“The Effects of Precinct-Level Canvassing”) and in Denver and Hands (“Canvassing, Turnout and Party Support”). For their part, the parties and party strategists also believe that reinforcement is the most important element in canvassing. See, for example. Gargrave, Anthony and Hull, Raymond, How to Win an Election (Toronto: Macmillan, 1979)Google Scholar, especially chap. 2, and Simpson, Dick, Winning Elections: A Handbook in Participatory Politics (Chicago: The Swallow Press, 1972)Google Scholar. especially chap. 6.

6 Kramer, “The Effects of Precinct-Level Canvassing.” His results showed few differences when several parties canvassed as opposed to one.

7 In this sense, the incorporation of the competitive context is tantamount to a testing of a specific question: is the impact of canvassing more pronounced in competitive areas? It is also a question clearly in need of testing, for only Crotty has addressed it empirically. To be sure, he did find that “party effort” did produce higher dividends in competitive or marginal areas, but his results were based only on aggregate data. Equally important, campaign activities comprised only part of his “party effort” measure, a composite index which also included leadership and organization items (Crotty, William, “Party Effort and Its Impact on the Vote,” American Political Science Review 65 [1971], 439–50.CrossRefGoogle Scholar) The incorporation of the competitive context is important for other reasons as well, namely, to ensure that certain effects are controlled for in the examination of canvassing-voting relationships. For example, in noncompetitive constituencies there is a necessity to take into consideration the fact that a party already dominant in an area may obtain its large share of the vote virtually by definition and independently of its activities. A more subtle variation on this notion has been offered by Cutright. He pointed out that in some areas voter choice could be due to the “result of living in a relatively homogeneous political environment”—the product of a socio-psychologically based trend toward conformity in voting preferences (Cutright, Phillips, “Measuring the Impact of Local Party Activity on the General Election Vote,” Public Opinion Quarterly 27 [1963], 372–86CrossRefGoogle Scholar, at 373). Controlling for the competitive context would at least partially adjust for this “majority effect,” assuming, of course, that it exists.

8 Cutright, Phillips and Rossi, Peter H., “Grass Roots Politicians and the Vote,” American Sociological Review 23 (1958), 171–79CrossRefGoogle Scholar, and “Party Organization in Primary Elections,” American Journal of Sociology 24 (1958), 262–69:Google ScholarKatz, Daniel and Eldersveld, Samuel J., “The Impact of Local Party Activity Upon the Electorate,” Public Opinion Quarterly 25 (1961), 1–24CrossRefGoogle Scholar: Cutright, , “Measuring the Impact of Local Party Activity”: Raymond Wolfinger, “The Influence of Precinct Work on Voting Behaviour,” Public Opinion Quarterly 27 (1963), 387–98:Google Scholar and Crotty. “Party Effort and Its Impact on the Vote.”

9 Katz, and Eldersveld, , “The Impact of Local Party Activity.” 6. 8Google Scholar.

10 Kramer, “The Effects of Precinct-Level Canvassing,” 564.

11 Ibid., 561.

12 Butler, and Stokes, , Political Change in Britain, 422Google Scholar.

13 Clarke, et al., Political Choice in Canada, 296Google Scholar.

14 Ibid.

15 The authors apparently included telephone canvassing, a form of personal contacting, in the category of impersonal contacting, and the same procedure was used here. Simply, there did not seem to be any possibility of further differentiating within this category as the following percentages—generated in response to a single question—suggest: 87.7 by literature only, 2.4 by telephone only and 9.8 by both. In any case, additional analysis in the present study indicated that eliminating those contacted by telephone from this category did little to alter the results.

16 Clarke et al., Political Choice in Canada, 296.

17 Ibid., Table 9.12, 293.

18 Ibid., 294, 296.

19 Kramer, , “The Effects of Precinct-Level Canvassing,” 568, 571Google Scholar.

20 Another, simpler, test took advantage of the fact that interviews were carried out over many weeks following the 1974 election—9.9 per cent within three weeks, 42.4 per cent between four and six weeks, 39 per cent between seven and nine weeks, 4.9 per cent between ten and twelve weeks and 4.0 per cent later than that. It was assumed that if extensive and systematic “forgetting” occurred, it would most likely do so the more remote in time the interview was from the election. This was not the case at all. Correlations between the various contacting items and the week of interview (coded as above) were effectively zero in all instances.

21 These results are available from the author upon request.

22 In addition to Clarke et al. see, especially, Eldersveld and Dodge, “Personal Contact or Mail Propaganda?”

23 The full set of results is available from the author upon request.

24 A few constituencies in 1972 were three-way competitive contests; that is, the margin of victory separating the winner and the third-place finisher was nine percentage points or less.

25 The independent variables used here—transformed into dummy variables in all instances—included religion, language, partisanship (including intensity), measures of party-related issue saliency and affect towards the party leaders, local candidates and the parties themselves. Political interest, not normally utilized in analyses of vote direction, was also brought into the regression equations. This was done in the wake of findings by Clarke, Jenson, LeDuc and Pammett that the variable correlated modestly with reports of impersonal contacts but not with contacts by candidates or party workers. About this, they said: “It seems likely that contact in person with workers or candidates was more memorable to all voters and therefore reported by most people with whom it took place. Literature, on the other hand, was likely delivered to most people, but read in detail only by those with some degree of political interest” (Political Choice in Canada, 295, emphasis in original). In other words, political interest might have some bearing on the degree of underreporting of some types of campaign contacts. However, when it was incorporated into the tests described in the section on validity, it had only a marginal impact. Nevertheless, to be on the safe side, it was decided to control for its possible effects by entering it into the regression equations for vote direction.

26 One might also speculate about the relevance of national competitiveness for the NDP result in the probable loss category. Here the coefficient for that party is the weakest of the three. Perhaps in this instance, it is the conjunction of constituency and national noncompetitiveness that mitigates against any strong canvassing effect.

27 This should not be taken to mean that canvassing necessarily has a strong impact on the (constituency) outcome. The results did not show strong recruitment and conversion effects, especially in competitive areas. At the same time, the patterns that did emerge tend to emphasize what can be considered to be status-quo kinds of situations; namely, significant reinforcement effects and, possibly, limited recruitment and conversion effects for the two older parties, but in ridings where they had already dominated. In the normal course of events, such effects would operate to help the same parties continue to win. On the other hand, it would be a mistake to discount totally the relevance of the findings for outcomes; in the extreme, the failure to hold onto past supporters or to win over new ones would be quite disastrous for a party. This would occur in the context of one election with widespread, extraordinary voting shifts and in the context of several elections with opposition parties increasingly eroding the lead of the dominant party.

28 The results, as they may elucidate party strategy, are also limited because not all the possible canvassing-voting relationships were considered here. Indeed, the combinations are numerous. For example, the impact of, say, Liberal canvassing on Conservative voting in NDP-dominated areas. The comparatively less elaborate analysis undertaken here posed enough challenges in terms of a full understanding; to study these added complications at this juncture would have been premature.