Hostname: page-component-7479d7b7d-767nl Total loading time: 0 Render date: 2024-07-13T08:32:11.290Z Has data issue: false hasContentIssue false
  • English
  • Français

A Simultaneous Equation Model of Canadian Voting Behaviour*

Published online by Cambridge University Press:  10 November 2009

Keith Archer
Keith Archer
Affiliation:
University of Calgary
Get access Rights & Permissions [Opens in a new window]

Abstract

Recent research on voting in United States presidential elections has begun to disentangle long-term and short-term components of electoral choice. This has been achieved through the use of complex models involving instrumental variables and two-stage or three-stage least squares regression. These techniques are particularly appropriate to understanding the voting decision in Canada because of the short-term variability of partisan identifications. Modelling voting choice, using simultaneous equations and data from the 1979 Canadian National Election Study, it was found that the major attitudinal determinants of voting—party identification and attitudes towards issues and party leaders—were strongly related to one another and to the direction of vote whereas sociodemographic characteristics were only weakly related to political attitudes and behaviour. In addition, the strength of these variables may vary across the major parties in any given election.

Résumé

Grâce à des recherches récentes sur le vote durant les élections présidentielles des États-Unis, on a commencé à isoler les différentes composantes à court et à long terme du choix électoral. Cela a été réalisé grâce a l'emploi de modèles complexes comportant des variables constitutives et des calculs de régressions en deux ou trois étapes par la méthode des moindres carrés. Ces techniques conviennent particulièrement pour aider à comprendre le choix électoral au Canada, à cause de la variation à court terme des identifications à un parti. La modélisation de la représentation du choix électoral à l'aide d'équations simultanées et de données provenant de l'enquête sur les élections nationales canadiennes de 1979, a permis de montrer que les principales attitudes qui déterminent le choix électoral—identification à un parti et attitudes envers les problèmes de l'heure et les leaders de parti—étaient liées les unes aux autres et à la direction du vote. Les caractéristiques socio-démographiques, elles n'avaient qu'un faible lien avec les attitudes et le comportement politiques. De plus, on a trouvé que l'importance de ces variables peut différer selon les grands partis et pour chaque élection particulière.

Type
Research Article
Information
Canadian Journal of Political Science/Revue canadienne de science politique , Volume 20 , Issue 3 , September 1987 , pp. 553 - 572
Copyright
Copyright © Canadian Political Science Association (l'Association canadienne de science politique) and/et la Société québécoise de science politique 1987

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Downs, Anthony, An Economic Theory of Democracy (New York: Harper and Row, 1957), chap. 3;Google ScholarCampbell, Angus, Converse, Philip, Miller, Warren, and Stokes, Donald, The American Voter (New York: Wiley, 1960), chap. 2.Google Scholar

2 See, for example, Jackson, John, “Issues, forty Choices and Presidential Voting,” American Journal of Political Science 19 (1975), 161–86;CrossRefGoogle ScholarMarkus, Gregory and Converse, Philip, “A Dynamic Simultaneous Equation Model of Electoral Choice,” American Political Science Review 73 (1979), 1055–70;CrossRefGoogle ScholarPage, Benjamin and Jones, Calvin, “Reciprocal Effects of Policy Preferences, Party Loyalties and the Vote,” American Political Science Review 73 (1979), 1071–89;CrossRefGoogle Scholar and Frankl, Charles in and Jackson, John, “The Dynamics of Party Identification,” American Political Science Review 77 (1983), 957–73.CrossRefGoogle Scholar

3 Meisel, John, Working Papers on Canadian Politics (2nd ed.; Montreal: McGill-Queen's University Press, 1975), 67;Google ScholarJenson, Jane, “The Filling of Wine Bottles is Not Easy,” this JOURNAL 11 (1978), 437–46;Google Scholar and Clarke, Harold, Jenson, Jane, LeDuc, Lawrence, and Pammett, Jon, Political Choice in Canada (Toronto: McGraw-Hill Ryerson, 1979), chap. 5.Google Scholar

4 LeDuc, Lawrence, Clarke, Harold, Jenson, Jane, and Pammett, Jon, “Partisan Instability in Canada: Evidence from a New Panel Study,” American Political Science Review 78 (1984), 475.CrossRefGoogle Scholar

5 Data from the 1979 Canadian National Election Study were provided by the Inter-University Consortium for Political and Social Research (ICPSR). The data were originally collected by Harold Clarke, Jane Jenson, Lawrence LeDuc, and Jon Pammett. Neither the ICPSR nor the original collectors of the data bear any responsibility for analyses or interpretations presented herein.

6 Endogenous variables are those whose values are determined by the model. In an ordinary least squares equation, they generally are known as dependent variables (that is, they represent the event[s] explained by the model). In a non-recursive system of equations, endogenous variables may be both dependent variables (that is, they may be predicted by the model), and they in turn may be used to predict the values of other variables. See, for example, Hanushek, Eric and Jackson, John, Statistical Methods for Social Scientists (New York: Academic Press, 1977), 225201328.Google Scholar For a further discussion and application, see Erlanger, Howard and Winsborough, Halliman, “The Subculture of Violence Thesis: An Example of a Simultaneous Equation Model in Sociology,” Sociological Methods and Research 5 (1976), 231–46;CrossRefGoogle Scholar and Wanner, Richard and Lewis, Lionel, “The Functional Theory of Stratification: A Test of Some Structural Hypotheses,” Sociological Quarterly 19 (1978), 414–28.CrossRefGoogle Scholar

7 Exogenous variables usually are referred to as independent variables. Their values are determined outside of the model and are taken as given. See Hanushek and Jackson, Statistical Methods for Social Scientists, 225–28.

8 Page and Jones, “Reciprocal Effects of Policy Preferences, Party Loyalties and the Vote,” 1077–82; Markus and Converse, “A Dynamic Simultaneous Equation Model of Electoral Choice,” 1059.

9 Page and Jones, “Reciprocal Effects of Policy Preferences, Party Loyalties and the Vote,” 1073. Technically, individuals vote directly for a slate of electors from their state who are part of the electoral college and who are pledged to vote for the presidential candidate receiving a plurality of votes from their state.

10 Clarke, Harold, Jenson, Jane, LeDuc, Lawrence and Pammett, Jon, “Voting Behaviour and the Outcome of the 1979 Federal Election: The Impact of Leaders and Issues,” this JOURNAL 15 (1982), 530–38.Google Scholar

11 Clarke et al., Political Choice in Canada, chap. 5.

12 Sniderman, Paul, Forbes, H. D., and Melzer, Ian, “Party Loyalty and Electoral Volatility: A Study of the Canadian Party System,” this JOURNAL 7 (1974), 268–88;Google ScholarJenson, Jane, “Party Loyalty in Canada: The Question of Party Identification,” this JOURNAL 8 (1975), 543–53. See also Richard Johnston, “The Reproduction of the Religious Cleavage in Canadian Elections,” this JOURNAL 18 (1985), esp. 102–05.Google Scholar

13 Meisel, Working Papers on Canadian Politics, chap. 2; Clarke et al., Political Choice in Canada, chap. 4. An interesting explanation for the persistence of this cleavage in the electorate despite its virtual irrelevance among the parties can be found in Johnston, “The Reproduction of the Religious Cleavage in Canadian Elections,” 99–113.

14 Schwartz, Mildred, Politics and Territory (Montreal: McGill-Queen's University Press, 1974);Google ScholarEngelmann, Frederick and Schwartz, Mildred, Political Parties and the Canadian Social Structure (Scarborough: Prentice-Hall, 1967)Google Scholar; and Alford, Robert, Party and Society: The Anglo-American Democracies (Chicago: Rand McNally, 1963).Google Scholar

15 Clarke et al., Political Choice in Canada, 122–24.

16 16 John Zipp, “Social Class and Canadian Federal Electoral Behavior: A Reconsideration and Elaboration” (unpublished doctoral dissertation, Duke University, 1978), chap. 3.

17 See Johnston, “The Reproduction of the Religious Cleavage in Canada,” 107–12, fora discussion of the importance of political context. A classic statement of the importance of context can be found in Lazarsfeld, Paul, Berelson, Bernard, and Gaudet, Hazel, The People's Choice (New York: Columbia University Press, 1944).Google Scholar

18 For an examination of this issue with respect to perceptions of provincial parties, see Elkins, David, “The Structure of Provincial Party Systems,” in Elkins, David and Simeon, Richard (eds.), Small Worlds: Provinces and Parties in Canadian Political Life (Toronto: Methuen, 1980).Google Scholar

19 Archer, Keith, “The Failure of the New Democratic Party: Unions, Unionists and Politics in Canada,” this JOURNAL 18 (1985), 362–64.Google Scholar

20 Kornberg, Allan and Archer, Keith, “A Note on Quebec Attitudes Toward Constitutional Options,” Law and Contemporary Problems 45 (1983), 76–82.Google Scholar

21 Fiorina, Morris, Retrospective Voting in American National Elections (New Haven: Yale University Press, 1981)Google Scholar; see also Page, Benjamin, Choices and Echoes in Presidential Elections (Chicago: University of Chicago Press, 1978).Google Scholar

22 See also Kornberg, Allan, Mishler, William and Clarke, Harold, Representative Democracy in the Canadian Provinces (Scarborough: Prentice-Hall, 1982), chap. 7.Google Scholar

23 To take an extreme example, suppose Trudeau was rated 100, Clark rated 0, and Broadbent rated 50. A weighted score for Broadbent, calculated as (leader C - [leader A + leader B + leader C] / 3) = Broadbent rating, would yield the same results as would the evaluations of Trudeau = 50, Clark = 50, and Broadbent = 50, despite the fact that the two individuals rated the candidates very differently. Although using the single-candidate thermometer rating is not statistically better than the weighted score, it may be inferentially superior.

24 Respondents were asked whether: (a) they were satisfied or dissatisfied with their economic condition and lives more generally, and (b) government had an effect on their condition. Responses were scored as follows: very satisfied = 2; fairly satisfied = 1; a little dissatisfied = -1; very dissatisfied = -2. These scores were multiplied by government's perceived impact responses, which were scored as follows: A great deal = 2; some = 1; not much = 0. The resulting index ranges from +4 to -4. A comparable index was generated for general life satisfaction, and the two measures were combined to yield an overall index with scores ranging from +8 to -8. For a previous application of this scale see, Kornberg, Allan, Clarke, Harold and Stewart, Marianne, “Federalism and Fragmentation: Political Support in Canada,” Journal of Politics 41 (1979), 889–905.CrossRefGoogle Scholar

25 The Blishen scale usually is interpreted as being at least an ordinal measure, and typically is treated as if it were interval level, and therefore appropriate for regression analysis. For a discussion of the construction of this index, see Blishen, Bernard and McRoberts, Kenneth, “A Revised Socioeconomic Index for Occupations in Canada,” Canadian Review of Sociology and Anthropology 13 (1976), 71–80.CrossRefGoogle Scholar

26 Clarke et al., “Voting Behaviour and the Outcome of the 1979 Federal Election,” 528; Clarke, Harold, Jenson, Jane, LeDuc, Lawrence and Pammett, Jon, Absent Mandate: The Politics of Discontent in Canada (Toronto: Gage, 1984), chap. 5.Google Scholar

27 To maximize the number of cases analyzed in the model, I have substituted the mean value for missing data, the effect of which has been to “flatten” the regression line. Although this procedure weakened the model overall, the eight-fold increase in the number of cases used in the analysis greatly increased our confidence in the reliability of results. Interestingly, in all but a few instances, the relative strength of the coefficients has been unaffected. Note that the explained variance (R-square) presented at the bottom of each table corresponds to the “vote” equation.

28 Archer, “The Failure of the New Democratic Party,” 362.

29 LeDuc et al., “Partisan Instability in Canada,” 481; Kornberg et al., Representative Democracy in the Canadian Provinces, 114–22.