¹ The sample design for the 1974 National Election Study was a multi-stage, stratified, cluster sample of 2.562 respondents, weighted by province to allow for some systematic oversampling of several of the smaller provinces. Items dealing with economic and life satisfaction, political efficacy, political participation, campaign activity, media exposure, party-leader images, some specific issue questions, and some special purpose sequences were asked of one of two random half-samples of respondents only. A more detailed discussion of the construction and design of this sample may be found in LeDuc, L., Clarke, H., Jenson, J. and Pammett, J.. ‘A National Sample Design’, Canadian Journal of Political Science, VII (1974), 701–8.CrossRefGoogle Scholar

² See, for example, Easton, David, A Systems Analysis of Political Life (New York: Wiley, 1965).Google Scholar

³ See, for example, Patterson, S. C., Wahlke, John and Boynton, G. R., ‘Dimensions of Sup port in Legislative Systems’, in Kornberg, Allan, ed., Legislatures in Comparative Perspective (New York: David McKay, 1971), pp. 282–313Google Scholar; Loewenberg, Gerhard, ‘The Influence of Parliamentary Behavior on Regime Stability: Some Conceptual Clarifications’, Comparative Politics, 111 (1971), 177–200CrossRefGoogle Scholar; Boynton, G. R. and Loewenberg, Gerhard, ‘The Development of Public Support, for Parliament in Germany, 1951–59’, British Journal of Political Science, 111 (1973), 169–89CrossRefGoogle Scholar; and Jewell, M. C. and Kim, C. L., ‘Sources of Support for the Legislature in a Developing Nation: The Case of Korea’, Comparative Political Studies, VIII (1976), 461–89.CrossRefGoogle Scholar Much of the theory that underlies their analyses derives from a paper originally presented by John Wahlke to the Seventh World Congress of the International Political Science Association (1967) and subsequently revised and published as ‘Policy Demands and System Support: The Role of the Represented’, British Journal of Political Science, 1 (1971), 271–90.CrossRefGoogle Scholar

⁴ Easton, David, ‘A Re-Assessment of the Concept of Political Support’, British Journal of Political Science, V (1975), 435–57CrossRefGoogle Scholar and ‘Theoretical Approaches to Political Support’, Canadian Journal of Political Science, IX (1976), 431–48.Google Scholar

⁵ According to Easton, ‘The proposal is that we…abandon the distinction between types of support. We could then consider support simply as a single homogeneous phenomenon subjected to one set of measures.’ ‘A Re-Assessment’, p. 443.Google Scholar

⁶ See, inter alia, Loewenberg, , ‘The Influence of Parliamentary Behavior on Regime Stability’, p. 184Google Scholar; and Davidson, R. H. and Parker, B. R., ‘Positive Support for Political Institutions: The Case of Congress’, Western Political Quarterly, XXV (1972), 600–12, especially p. 602.Google Scholar

⁷ Attempts to operationalize and measure concepts such as regime support raise a variety of complex conceptual and methodological issues. Because these issues are complex, they are open to a variety of interpretations and no single operational formulation of a concept such as support is likely to please everyone. Undoubtedly, our own attempt will fare no differently. However, the scholarly controversy is certainly not a bad thing. Given the status of political science as a science, it may well be a good thing.

⁸ On the concept of generalizing or spread of affect see Deutsch, M. and Krauss, R., Theories in Social Psychology (New York: Basic Books, 1965), pp. 86–7Google Scholar: Miller, N. E. and Dollard, J., Social Learning and Imitation (New Haven: Yale University Press, 1941)Google Scholar; and Mower, O. H., Learning Theory and behavior (New York: Wiley, 1960).CrossRefGoogle Scholar

⁹ See, inter alia, Clarke, Harold D., Price, Richard and Krause, Robert, ‘Constituency Service Among Canadian Provincial Legislators: Basic Findings and a Test of Three Hypotheses’, Canadian Journal of Political Science, VIII (1975), 520–42.CrossRefGoogle Scholar

¹⁰ Epstein, Leon D., ‘A Comparative Study of Canadian Parties’, American Political Science Review, LVIII (1964), 46–59.CrossRefGoogle Scholar

¹¹ When this paper was written there were at least three organized separatist movements in Western Canada. These movements are based upon the conviction that historically the Prairie provinces and British Columbia have been economically disadvantaged by tariff policies that protect Ontario industry, and by transportation policies (particularly the structure of freight rates) that favour the provinces of Ontario and Quebec.

¹² Empirical research and conventional wisdom both have argued that federalism and regionalism strongly affect the operation of the Canadian political system and political behaviour in Canada. See, for example, a recent account of the impact of federalism on Canadian politics, Smiley, Donald V., Canada in Question: Federalism in the Seventies (Toronto: McGraw-Hill, 1976).Google Scholar On the existence of regional political cultures in Canada see Schwartz, Mildred A., Politics and Territory (Montreal: McGill-Queen's University Press, 1974)CrossRefGoogle Scholar and Simeon, Richard and Elkins, David J., ‘Regional Political Cultures in Canada’, Canadian Journal of Political Science, VII (1974), 397–437.CrossRefGoogle Scholar

¹³ See Easton, , ‘A Re-Assessment,’ pp. 455–6.Google Scholar

¹⁴ On this regression technique see Nie, N., Hull, C., Jenkins, J., Steinbrenner, K. and Bent, D., Statistical Package for the Social Sciences, 2nd edn. (Toronto: McGraw-Hill, 1975), pp. 345–6.Google Scholar

¹⁵ For a recent discussion of the use of standardized versus unstandardized coefficients see Kim, Jae-On and Mueller, Charles W., ‘Standardized and Unstandardized Coefficients in Causal Analysis: An Expository Note’, Sociological Methods and Research, IV (1976), 423–37.CrossRefGoogle Scholar

¹⁶ For a detailed explanation of this analytic technique see Sonquist, John A., Baker, Elizabeth and Morgan, James N., Searching for Structure (Ann Arbor, Mich.: Institute for Social Research, 1973).Google Scholar

¹⁷ Moreover, despite evidence of interaction effects in the data, the AID and regression analyses largely agreed in their estimates of the variance in regime support explained by the several predictor variables – 31–8 per cent and 32·5 per cent respectively.

¹⁸ Region is a dummy variable with Ontario residency the basis of comparison.

¹⁹ Kornberg, Allan, Canadian Legislative behavior (New York: Holt, Rinehart and Winston, 1967), p. 73.Google Scholar

²⁰ By way of illustration, during interviews taken in a study of the 28th Parliament (1968–72), a number of MPs representing metropolitan areas indicated that constituents write and sometimes even telephone to complain about matters that are primarily of concern to local governments (i.e., inadequate garbage collection, the failure to keep grass and trees properly trimmed in public parks, and so forth). See Kornberg, Allan and Mishler, William, Influence in Parliament: Canada (Durham: Duke University Press, 1976), Chap. 5, on communications from constituents.Google Scholar

²¹ In Canada, an additional factor may be the MP's inability to process requests because of a lack of staff assistance.

²² See Kornberg, and Mishler, , Influence in Parliament: Canada, p. 192.Google Scholar

²³ Easton, , ‘A Re-Assessment’, p. 437.Google Scholar

²⁴ The majority of francophone Canadians who do not live in Quebec reside in three provinces – New Brunswick, Ontario and Manitoba.

²⁵ Respondents in the 26–34 category had average support scores of 57·5, those 35–44 had scores of 61·6, and those who were 45–54 years old had average scores of 66·2.

²⁶ For example, Liberal party identifiers in the under 25 age cohort had average scores of 62·8, those in the 26–34 category had scores of 63·2 whereas those in the 35 and over category had average scores of 73·1. The scores for Parti Québécois identifiers in these age categories were, respectively, 43·2, 47·8 and 50·8.

²⁷ It will be recalled that the explanatory importance of these variables derives primarily from findings of voting and political participation studies carried out in the United States during the 'forties and 'fifties, a period during which the use of multivariate analytic techniques was limited because the required technology was not well developed.

²⁸ Relatedly, Miller has demonstrated that electoral support for the Scottish National Party in the October 1974 British election was correlated with ‘trust in government’, a concept that obviously is analogous to regime support. See Miller, William L., ‘The Connection Between SNP Voting and the Demand for Scottish Self-Government’, European Journal of Political Research, v (1977), 83–102.CrossRefGoogle Scholar

²⁹ For a number of reports on the phenomenal rise of the Parti Québécois see, inter alia, Hamilton, Richard and Pinard, Maurice, ‘The Bases of Parti Québécois Support in Recent Quebec Elections’, Canadian Journal of Political Science, IX (1976), 3–26CrossRefGoogle Scholar; Latouche, Daniel, ‘Quebec’, in Bellamy, David et al. , eds., Provincial Political Systems (Toronto: Methuen, 1976), 19–30Google Scholar; Cuneo, Carl J. and Curtis, James, ‘Quebec Separatism: Analyses of Determinants Within Social Class Levels’, Canadian Review of Sociology and Anthropology, XI (1974), 1–29Google Scholar; and Posgate, Dale and McRoberts, Kenneth, Quebec: Social Change and Political Crisis (Toronto: McClelland and Stewart, 1976), pp. 189–203.Google Scholar

³⁰ The scale included items such as whether respondents talked about politics, tried to convince others, and carried out election related activities.

³¹ Thus, respondents who voted in the 1974 federal election had average regime support scores of 63·5 whereas those who had not had scores of only 58·6 (F= 15·49, P ≤ 001). With respect to voting over an extended period of time, we found that respondents who said they voted in all federal elections in which they were eligible had average scores of 63·8; those who had voted in most elections had scores of 61·9; those voting in only some elections had average scores of 58·2; and those who said they never had voted in federal elections had scores of 50·7 (F= 13·97, P ≤ 001). Finally, respondents who were on the low end of a Guttman scale of political participation had scores of 56·8 whereas those in the middle of the scale had scores of 62·5. Those on the high end of the scale had scores of 64·5 (F = 2·44, P ≤ 05).

³² Easton, , ‘A Re-Assessment’.Google Scholar