1 For a detailed analysis of the 1970 realignments, see Pinard, Maurice, “The Ongoing Political Realignments in Quebec,” in Quebec Society and Politics: Views from the Inside, ed. by Thomson, Dale C. (Toronto: McClelland and Stewart, 1973). 119–38.Google Scholar

2 For additional details, see Hamilton, Richard and Pinard, Maurice, “The Bases of Parti Québécois Support in Recent Quebec Elections,” this Journal 9 (1976), 3–26.Google Scholar Let us only mention that our estimates of the turnout (85 per cent) and of the support for each party are fairly accurate, even though 38 per cent of those who said they voted (N = 859) (or 32 per cent of the total sample [N = 1,012]) did not reveal their preference (see ibid.). Of those who did, 56 per cent said they voted Liberal, 32 per cent P.Q., 8 per cent Créditiste, 4 per cent Union Nationale, and 1 per cent others. The Liberal vote is overestimated by 1 per cent, the P.Q. vote by 2 per cent, while the Parti Créditiste vote is underestimated by 2 per cent and the U.N. by 1 per cent.

3 Like the poll of May 1973 reported in Table 1, earlier polls in 1972 gave an overall strength of 18 per cent to the Créditistes and led us to expect different developments in rural Quebec. But that strength vanished in the latter part of 1973. On these expectations at the time, see Pinard, Maurice, “The Ongoing Political Realignments,” as well as “La scission au sein du Ralliement Créditiste et ses conséquences électorales,” in Partis Politiques au Québec, ed. by Pelletier, Réjean (Montréal: Hurtubise HMH, 1976), 167–96.Google Scholar

4 A more refined analysis, replicating that of “The Ongoing Political Realignments.” would consider the outcome separately for the Montreal region and the rest of the province.

5 It should be stressed that in addition to the first two polls of Table 1, other polls prior to 1973 had also registered gains for the Créditistes, such as the one analyzed in ibid. and that conducted during the October 1972 federal election by Hamilton, and Pinard, (The Montreal Star and Le Soleil, October 25, 26, and 27, 1972).Google Scholar

6 See in particular Berelson, Bernard R., Lazarsfeld, Paul F. and McPhee, William N., Voting: A Study of Opinion Formation in a Presidential Campaign (Chicago: The University of Chicago Press, 1954),Google Scholar and Campbell, Angus, Converse, Philip E., Miller, Warren E. and Stokes, Donald E., The American Voter (New York and London: John Wiley and Sons, 1960).Google Scholar

7 Key, V. O. Jr., The Responsible Electorate: Rationality in Presidential Voting, 1936–1960 (Cambridge, Mass.: Harvard University Press, 1966).CrossRefGoogle Scholar

8 Converse, Philip E., Miller, Warren E., Rusk, Jerrold G. and Wolfe, Arthur C., “Continuity and Change in American Politics: Parties and Issues in the 1968 Election,” The American Political Science Review 63 (1969), 1083–1105.Google Scholar

9 Ibid., 1096–99.

10 See, for instance, Repass, David E., “Issue Salience and Party Choice,” The American Political Science Review 65 (1971), 389–400CrossRefGoogle Scholar; Pomper, Gerald M., “From Confusion to Clarity: Issues and American Voters, 1956–1968,” The American Political Science Review 66 (1972), 415–28CrossRefGoogle Scholar; Richard W. Boyd, “Popular Control of Public Policy: A Normal Vote Analysis of the 1968 Election,” ibid., 429–49 (see also in the same issue the “Comments” by R. A. Brody and B. I. Page, and by J. H. Kessel, 450–58 and 459–65, and the “Rejoinders” by Pomper and by Boyd, 466–67 and 468–70); Page, Benjamin I. and Brody, Richard A., “Policy Voting and the Electoral Process: The Vietnam War Issue,” The American Political Science Review 66 (1972), 979–95.CrossRefGoogle Scholar A related debate concerned the lack or presence of ideological sophistication or belief systems in the American or British Electorate; on this, see in particular Campbell et al., The American Voter, chaps. 9–10; Converse, Philip E., “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. by Apter, David (Glencoe: The Free Press, 1964), 206–61Google Scholar; Butler, David and Stokes, Donald, Political Change in Britain (New York: St. Martin's Press, College Edition, 1971)CrossRefGoogle Scholar, chap. 12; Field, John Osgood and Anderson, R. E., “Ideology in the Public's Conceptualization of the 1964 Election,” Public Opinion Quarterly 33 (1969), 380–98CrossRefGoogle Scholar; Pierce, John C. and Rose, D. G., “Nonattitudes and American Public Opinion: The Examination of a Thesis,” The American Political Science Review 68 (1974), 626–49CrossRefGoogle Scholar (also Converse's “Comment” and Rose and Pierce's “Rejoinder” in ibid., 650–60 and 661–66); Achen, Christopher H., “Mass Political Attitudes and the Survey Response,” The American Political Science Review 69 (1975), 1218–31.CrossRefGoogle Scholar There is a problem with the original Berelson, Lazarsfeld and McPhee work in that the key questions demonstrating the widespread ignorance of the general population were asked in August, that is, before the beginning of the actual campaign. In two key passages (217–18 and 227), the authors indicate the levels of “ignorance” of candidate positions in their text and note the date of the questions in footnotes (e.g., “these figures apply to the early campaign period of August. Similar data for October… would almost certainly raise these estimates” [228]). In a brief popular text, this aspect of the research is summarized as follows: “In the Dewey-Truman election of 1948 the Taft-Hartley Act was hotly debated and the two candidates took clear and opposite stands (Dewey pro, Truman con); but one of the classic studies of voting behaviour showed that a full third of the voters had never heard of the Act at all, and that another third showed no interest in the controversy over it.” A footnote refers to the pages given above. This account however makes no mention of the date of the questions: the original authors’ cautionary statements, in short, have disappeared from view. There are some other difficulties with this summary; it misrepresents even these discussions. The original authors did not say the candidates took “clear and opposite stands.” They very clearly state the asymmetry of their respective presentations (“… their positions on the issues in the campaign were not equally clear…. Truman took a more straightforward and more aggressive position on these issues than Dewey…” [217]. And, “… there can be no question but that… Dewey's position was more amenable to misperception than Truman's…”). Neither of the pages cited, incidentally, contain the specific points made in this version, that is, the points about issue awareness or interest. On the following page, they do correctly cite the original authors with respect to the low levels of accurate public perception of the candidates’ positions on two key issues. Once again, however, the cautionary footnote statement is omitted. See Bone, Hugh A. and Ranney, Austin, Politics and Voters (New York: McGraw-Hill, 1963), 11, 12.Google Scholar For a more detailed criticism of the Berelson, Lazarsfeld and McPhee work see Hamilton, Richard, Class and Politics in the United States (New York: John Wiley, 1972), 49ff.Google Scholar

11 For another criticism of this failure, see Michael Margolis, “From Confusion to Confusion: Issues and the American Voter, 1956–1972,” The American Political Science Review, forthcoming.

12 See for instance the work of RePass, Pomper, Boyd, Page and Brody cited in footnote 10.

13 Campbell et al., The American Voter, chap. 8.

14 Butler and Stokes, Political Change in Britain, chaps. 10 and 11. Butler and Stokes combine the first two conditions just listed and add an additional one for issue-voting to be manifest not only among individuals, but also for it to affect the relative standing of the parties: the condition is that the distribution of opinions on the issue must be skewed; ibid., 173–79. See also Nie, Norman H., Verba, Sidney and Petrocik, John R., The Changing American Voter (Cambridge, Mass.: Harvard University Press, 1976), 97.Google Scholar On the logic of issue-voting, see ibid., 157–64.

15 RePass, “Issue Salience.”

16 Pomper, “From Confusion to Clarity.”

17 Boyd, “Popular Control of Public Policy.”

18 Converse et al., “Continuity and Change.”

19 Page and Brody, “Policy Voting.”

20 Margolis, “From Confusion to Confusion.” Margolis may in turn press his criticisms too far: among those meeting the conditions for issue-voting, there may well be an increase in such voting.

21 Nie et al., The Changing American Voter.

22 See, however, on the salience of issues in the 1968 Canadian election, Meisel, John, Working Papers on Canadian Politics (2nd enlarged ed.; Montreal: McGill-Queen's University Press, 1975), 14ff.Google Scholar Also Regenstreif, Peter, The Diefenbaker Interlude: Parties and Voting in Canada (Toronto: Longmans Canada, 1965)Google Scholar, chap. 4.

23 LeDuc, Lawrence, “Public Opinion and Minority Government,” paper read at the meeting of the CPSA, Toronto, 1974Google Scholar, and “The Majority Government Issue and Electoral Behaviour in Canada: 1965–1974,” paper read at the IPSA/ISA Conference on Elections in Complex Societies, Committee on Political Sociology, Kingston, 1975. See also the revision of the latter paper, “Political Behaviour and the Issue of Majority Government in Two Federal Elections,” in this issue of the Journal, 311–39.

24 Pammett, Jon, LeDuc, Lawrence. Jenson, Jane and Clarke, Harold, “The 1974 Federal Election: A Preliminary Report,” Occasional Papers No. 4, Department of Political Science, Carleton University, 1975Google Scholar, and LeDuc, L., Clarke, H., Jenson, J. and Pammett, J., “Partisanship, Political Interest, and Electoral Campaigns in Canada,” paper read at the IPSA Congress, Edinburgh, 1976.Google Scholar See also Pammett, et al. revision of the first of these papers, “The Perception and Impact of Issues in the 1974 Federal Election,” this Journal 10 (1977), 93–126Google Scholar, in which they more systematically examine the impact of inflation, leadership, and majority government in light of Butler and Stokes’ conditions for issue-voting. They find that none of these issues unequivocally meets all three conditions.

25 See Pinard, Maurice, “La rationalité de l'électorat: le cas de 1962,” in Quatre élections provinciales au Québec, 1956–1966, ed. by Lemieux, Vincent (Quebec: Presses de l'Université Laval, 1969), 187ff.Google Scholar This article reviews a number of instances in which nationalism was alleged to be the decisive issue. None of them, however, contained relevant evidence.

26 See “La rationalité de l'électorat.” For a recent statement of the widely-held opposite position, see Jones, Richard, Community in Crisis: French-Canadian Nationalism in Perspective (Toronto and Montreal: McClelland and Stewart, 1972), 130.Google Scholar The much-debated issue of nationalization was not found to be a salient one in Britain either. See Butler and Stokes, Political Change in Britain, 193.

27 Pinard, Maurice, The Rise of a Third Party (Montreal and London: McGill-Queen's University Press, 1975), 80–86.Google Scholar

28 “La rationalité de l'électorat,” 190.

29 For a brief summary of his research, see McRoberts, Kenneth, “Social Communication and Mass Nationalism,” Canadian Review of Studies in Nationalism 2 (1974), 148–51.Google Scholar But for a contrived argument rejecting McRoberts’ evidence, see Dion, Léon, Nationalismes et politique au Québec (Montréal: Hurtubise HMH, 1975), 144ff.Google Scholar

30 Nationalist attitudes, though very important, were probably less crucial in the 1970 election than in the 1973 one. In particular, dissatisfaction with the U.N. administration turned out to be a stronger determinant of the vote. See Lemieux, Vincent, Gilbert, Marcel et Blais, André, Une élection de réalignement: l'élection générale du 29 avril 1970 au Québec (Montréal: Editions du Jour, 1970)Google Scholar, chap. 7. The role of nationalism in the period from 1969 to 1973 has been attested to by a number of studies; see Jenson, Jane and Regenstreif, Peter, “Some Dimensions of Partisan Choice in Quebec, 1969,” this JOurnal 3 (1970), 308–17Google Scholar; Daniel Latouche, “The Independence Option: Ideological and Empirical Elements,” in Quebec Society and Politics, 179–96; Carlos, Serge et Latouche, Daniel, “La composition de l'électorat péquiste,” in Le processus électoral au Québec: les élections provinciales de 1970 et 1973, ed. by Latouche, D., Lord, G. et Vaillancourt, J.-G. (Montréal: Hurtubise HMH, 1976), 187–211Google Scholar; see also Serge Carlos, E. Cloutier et D. Latouche, “Le choix des électeurs en 1973: caractéristiques sociales et orientation nationale,” in ibid., 213–34. Notice, however, that while previous discussions in the literature usually involved the assessment of the impact of issues and ideology relative to such other short-term factors as the candidates’ popularity, the assessment of the incumbent party's administration, economic and other grievances, as well as long-term forces such as partisan loyalty and the socio-economic characteristics of the voters, the latter group of studies has reduced the debate to a simple alternative between nationalist issues and ideology vs. socio-economic characteristics; they attribute to the former a much greater impact, while Pinard, allegedly, would stress the latter. (For Pinard's actual position which is along the lines of the literature in general, see “Le Rationalité de l'électorat.”) The first three works just cited also purport to challenge Pinard's position on the role of nationalism. Since they do not deal with any specification of the conditions under which a nationalist issue will or will not affect the vote, they are actually missing the point completely. The Carlos-Latouche piece besides makes serious misrepresentations of the Pinard argument. Their citation drops a key word, so as to have him making an invariant as opposed to a conditional argument. Yet the absence of impact of the nationalist issue in 1962 had explicitly been presented as not invariant: “We are not asserting that nationalism… or ethnic reactions cannot exert major effects… on the electoral behavior of masses, but simply that the occasions when this happens are fewer than it is generally believed” (“La rationalité de l'électorat,” 188). The Carlos-Latouche piece also attributes a position to Pinard with respect to 1970 conditions for which they have no basis whatsoever. Having thus transformed the conditional into the absolute, it is no problem for them to find that he is “wrong.” Such argumentation simply fails to come to grips with the prior evidence or with the argument.

31 In this survey, only 8 per cent of the non-French minority answered “don't know” or undecided on this issue, an indication of widespread familiarity. As for the intensity of feelings, it was found in the lead author's study of the independence movement carried in 1970 that among the non-French who had an opinion on the issue of separatism (that was 89 per cent of that subgroup, dividing into 3 per cent favourable and 86 per cent unfavourable) no less than 74 per cent said that their opposition to the separation of Quebec was “extremely strong” and 72 per cent said that their opinion on the matter could change “only with a great deal of difficulty.” Given the clarity of the stands taken by the parties and, in particular, the strong relationship observed elsewhere between intensity of feelings and accuracy of perception of party positions (see RePass, “Issue Salience,” 394), it can reasonably be assumed that the perception of party positions in this instance was also extremely accurate.

32 The Union Nationale and Parti Créditiste received 2 per cent each (N = 64). Three polls taken during the electoral campaign (see references to polls I, II, and IV in Table 1) similarly gave to the Liberals 88, 88 and 84 per cent of the non-French vote, respectively, with 9, 3 and 8 per cent to the P.Q., and 2, 9 and 9 per cent for the other two parties combined. Note the similarity with blacks in the U.S.A. who voted more than 90 per cent Democratic in 1964 and 1968. (See Boyd, “Popular Control of Public Policy,” 444.) In both cases, we are dealing with groups who feel very seriously threatened by the opposite party.

33 According to three polls reported by Lemieux, 64, 72, and 76 per cent intended to vote Liberal and 5, 16, and 3 per cent Parti Québécois; 23, 13, and 20 per cent preferred the other two parties, which was on the whole more than twice as large a percentage as in the 1973 election (from Vincent Lemieux et al., Une élection de réalignement, 60).

34 The Social Research Group study done before the 1960 provincial election indicated that 50 per cent of the non-French-speaking population of Montreal intended to vote Liberal as against 40 per cent for the National Union and 10 per cent for others (N = 60). (Recomputed from the SRG report, “Les Electeurs Québécois,” Montreal, 1960, 47. Mimeographed.)

35 As of this writing (August 1976), polls indicate a high degree of dissatisfaction towards the Quebec Liberal party among the non-French, particularly on language and related issues. And only about half of those who express a preference say they would support the Liberals if an election were held today. (From the Gallup poll conducted for La Presse, May 1976. We are indebted to the direction of La Presse for making the report of this poll available to us.)

36 The 1970 figures come from Regenstreif's poll, which came closest to the actual results, though overestimating the P.Q. vote and underestimating the Liberal one. (This is the source also for the U.N. and P.C. figures cited above.) See V. Lemieux et al., Une élection de réalignment, 60.

37 Actually, the proportions answering “don't know” (thus possibly unaware) among French Canadians have varied somewhat over time and have been as low as 11 per cent in one study in 1970 (see Table 8). In this study, those who gave such an answer (21 per cent among French Canadians) were asked if they would be tempted to be for or against the separation of Quebec if they had to make a decision. Nine per cent were then willing to choose an option (5 per cent for and 4 per cent against), leaving only 12 per cent still undecided or unaware. Further, when the latter were asked if they ever had been in favour of separation, an additional 5 per cent answered positively or negatively, implying awareness about the issue. This leaves a maximum of only 7 per cent as unaware, some of which are no doubt also undecided or unwilling to reveal their opinion.

38 In this study, 79 per cent of the French Canadians in favour of separation and who reported a vote said that “the problem of the independence or separation of Quebec” had been “very important in their choice for a party” (N = 130). Moreover, in the 1970 Quebec study of independence cited before, only 21 per cent of the French Canadians in favour of independence said their opinion on the matter could change “very” or “fairly easily”; 40 per cent answered “with some difficulty”; and 36 per cent answered “with a great deal of difficulty.” Finally, the data collected by Clarke and his colleagues for a study of the 1974 federal election indicate that 40 per cent of the French-speaking Quebeckers in favour of independence felt “very strongly” about the issue, while another 51 per cent felt “fairly strongly”; only 9 per cent felt “not very strongly.”

39 To cite but one study among many, a poll carried out by CROP during the 1970 provincial election campaign revealed that 79 per cent of the French-speaking respondents mentioned one of the two economic problems listed (unemployment, 51 per cent, and cost of living, 28 per cent) as the most urgent one to be solved after the election, while only 8 per cent mentioned one of the two ethnic problems (status of French language in Quebec, 2 per cent, and the malaise in the relationship between Quebec and the rest of Canada, 5 per cent). Eleven per cent mentioned other problems. (Recomputed from data kindly supplied to the authors by CROP.) See also the Social Research Group study for the B. and B. Commission. “A Study of Interethnic Relations in Canada,” Montreal, 1965, 21, 23.Google Scholar Dittoed; and a poll during the 1974 federal election, La Presse (Montreal), June 15, 1974, A6. On the centrality of such concerns elsewhere, see Hamilton, Class and Politics, chap. 3, and also Cantril, Hadley, The Pattern of Human Concerns (New Brunswick: Rutgers University Press, 1964).Google Scholar

40 It was found that among those opposed to separation the assumed economic costs of independence were by far the most important reason for their opposition. See Pinard, Maurice, “La dualité des loyautés et les options constitutionnelles des québécois francophones,” in La nationalisme québécois à la croisée des chemins. Collection Choix (Québec: Centre québécois de relations internationales, 1975), 87–90.Google Scholar

41 In this study no less than 75 per cent of the French Canadians opposed to separation and who reported a vote said that the problem of separation had been “very important” in their vote (N = 275).

42 In Meisel's 1968 federal election study, 68 per cent of the French Canadians in Quebec said they were opposed to separation; these divided between 14 per cent “slightly opposed” and 54 per cent “strongly opposed.” In the 1970 study of independence, it was found that among French Canadians in Quebec opposed to separation, only 8 per cent said their opinion on the matter could change “very” or “fairly easily”; 33 per cent answered “with some difficulty” and as many as 56 per cent answered “only with a great deal of difficulty.” In another poll carried out in October 1972, a group of 63 per cent among French Canadians in Quebec said they would be disappointed if Quebec were to become independent in a year or two; among them, more than half (59 per cent) said they would be “extremely disappointed,” while 30 per cent said they would be “somewhat disappointed” and only 11 per cent said “vaguely disappointed.” (For this set of data, see Maurice Pinard, “La dualité des loyautés,” 78.) Finally, in Clarke et al.'s study of the 1974 federal election, the French Quebeckers opposed to separation divided into 64 per cent feeling “very strongly” about it, 30 per cent “fairly strongly,” and only 6 per cent “not very strongly.” These unusually skewed distributions among those opposed to separation provide unambiguousx indications of the strength of their feelings on the issue. Indeed, it is surprising to find that the strength of the feelings of those opposed to separation is greater than that of the feelings of those in favour of it (compare the data in this footnote with that in footnote 38).

43 See RePass, “Issue Salience,” 394.

44 Nie et al., The Changing American Voter, esp. chap. 4.

45 In 1970, while the majority was opposed to independence, only a minority was satisfied with the incumbent Union Nationale government. In 1973, the majority opposed to independence was also largely the majority satisfied with the Liberal government. (Polls gave the proportion of people satisfied with the government on the eve of the election in 1970 to be 15 per cent, and in 1973 56 per cent.)

46 Nie et al., The Changing American Voter, 157–64.

47 As in almost all research of this kind, the implication here is of course that of a causal relationship, that people's stand on this issue led to their electoral choice. We are conscious that only an association is being established and that causality is not thereby demonstrated. In particular, the causal link could be running in the opposite direction, revealing “rationalization,” “persuasion” or “projection” (see Brody and Page, “Comments”). But this is indeed a constant problem in almost every study of issue-voting, with the analysts making the causal inferences and the commentators, as in the case of Brody and Page, playing the critical role. (The latter incidentally were, however, less rigorous in their own research! See Page and Brody, “Policy Voting and the Electoral Process,” esp. 990–93.) We think that while causality cannot for the most part be directly established, our study contains many indications supporting the inference. First, our most direct piece of evidence relates to the time sequence between the two variables which can be established for part of the sample. Among the 130 respondents favouring independence and who expressed a vote preference (see first line of Table 2), 36 indicated that they had been for the independence of Quebec for more than five years, that is, before the P.Q. was created in October 1968. and first contested an election in April 1970. Of these, no less than 97 per cent voted P.Q. in 1973. Among those who had been in favour of independence for less than two years, only 37 per cent voted P.Q. in 1970, while 94 per cent of those who had been in favour for more than five years already did so in 1970 (N = 35 and 33 respectively). Second, while the P.Q. was only created in 1968, as just mentioned, the separatist issue has been part of the public debate since the early sixties. If rationalization of one's issue position in line with one's party preference was the main process involved in the relationship, a sudden and substantial increase in the popular support of separatism should have accompanied the growth of the P.Q. Data to be presented below (Table 8) show that this is not the case, the support for Quebec's independence on the whole continuing its regular and slow growth, despite the later and faster growth of the P.Q. Third, we do not deny the existence of a partial reversal of the causal sequence; P.Q. support seems to have led in some instances to persuasion or conversion, particularly during that party's electoral campaign, as is established below. Fourth, with 73 per cent of the French Canadians reporting that the issue of independence had been “very important” in their vote decision, it is difficult to argue that our basic relationship reflects mainly rationalization. Fifth, a large part of the results presented below and in particular those based on turnover analysis (changes in party preferences) are simply inconsistent with a causal inference opposite to the one we are making (see especially Tables 3 and 7 below). Finally, and above all, given the salience of the independence issue and the strength of feelings toward it, it is simply difficult to make a sound argument for the opposite causal inference. Feelings about the issue are most of the time too strong for positions on it to be simply rationalizations of party preferences made on other grounds.

48 The respective N's. for those in favour of independence, are 33, 84, and 13. For those against it, they are 29, 166, and 82.

49 See Hamilton and Pinard, “The Bases of Parti Québécois Support.”

50 When unemployment rather than financial situation was considered (in a manner similar to Table 5, second panel), we found that those from families with unemployment were always more likely to vote P.Q., but that the percentage differences decreased from 42 to 1 per cent as one moved from the low education group in Montreal to the high education one outside of Montreal.

51 “The Bases of Parti Québécois Support,” esp. 12–13.

52 Carlos and his associates have argued that with regard to the Liberal voters, satisfaction with the government was more important in determining their vote than the independence issue, while it was the other way around for P.Q. voters. (See Serge Carlos et al., “Le choix des électeurs en 1973,” 233.) This goes against the argument of this paper. There are, however, problems with these authors’ interpretation. First, in their data, the overall effect, considering all parties, is greater in the case of the independence issue than in the case of the “satisfaction issue.” Or, to put it differently, the independence issue pulled more voters toward the P.Q. (and away from other parties) than dissatisfaction pulled voters away from the Liberals (and towards other parties). Indeed, one striking finding from their data is that as many as 49 per cent of the voters dissatisfied with the Liberal government, but rejecting independence nevertheless voted Liberal in 1973 (ibid.). Moreover, this stronger impact of the independence issue holds despite the weakness of their measure of independence options (to be discussed below) which, among other things, leads to a weaker association of these attitudes with the vote than is the case with our data. Finally, one of our arguments is that the independence issue led to a polarization between the Liberals and the P.Q. Dissatisfaction, to the extent that it existed, just worked the other way around, toward depolarization, as will be discussed shortly.

53 For practically identical results, see the Poll, Blais-Lemieux-Renaud in Le Devoir, 23 October 1973, 5.Google Scholar

54 Whether or not that fear is justified is of course another, separate and distinct question, one that would demand a completely different line of enquiry or speculation. Given the definition of the situation, the choices made by these voters are clearly “rational,” that is to say, they have chosen the appropriate (or at least what appeared to be the appropriate) mean for the realization of their ends.

55 Let us add that the P.Q. retained practically all its independentist voters of 1970, whatever their image of the Liberals, while the Liberal party retained 95 per cent of the anti-separatists who agreed with the statement, but only 69 per cent of those who disagreed (N for the Liberals = 137 and 13, respectively). The reader should be cautioned, however, about some small N's in some of this analysis.

56 See Butler and Stokes, Political Change in Britain, 175–77.

57 A few other polls done in recent years have revealed higher proportions in favour of independence, but they are not directly comparable. The results of these polls, for all Quebec citizens, are as follows:

(The two 1972 polls are CROP—La Presse polls carried during the federal election campaign of 1972; La Presse, October 7 and 28, 1972. The 1973 poll is that of Carlos, Cloutier and Latouche, carried out during the 1973 provincial election campaign; Carlos et al., “Le choix des électeurs en 1973,” 223. The 1974 poll is from IQOP—La Presse, November 14, 1974.) Note that the results obtained here with regard to the proportions in favour of independence are about 10 per cent higher than the results for the corresponding years in Table 8. The differences seem to be due to the questions employed which used the expression “independence of Quebec” or “political independence of Quebec,” rather than the expression “separation of Quebec” or “separation of Quebec” together with “independence,” as in all the questions cited with the data in Table 8. (The questions here: 1972A and B: “Personally are you for or against the political independence of Quebec from the rest of Canada”; 1973: “How do you feel towards the idea of Quebec's independence? Do you strongly approve, approve, disapprove, strongly disapprove?”; note however a slightly different rendition of that last question in French: “Comment réagissez-vous face à l'indépendance du Québec ? Etes-vous complètement d'accord, etc.”; 1974: “Personally, are you for or against the independence of Quebec?”) Why is it that questions using only the word “independence” yield higher proportions in favour of independence than questions using the word “separation”? One reason seems to be that there are problems of validity with regard to the former questions. In particular, some respondents seem to simply equate “independence of Quebec” with “provincial autonomy,” as in the saying that “Quebec should be independent of Ottawa in its own decisions.” (Hence our preference for the expression “to form an independent country,” which we have also used, as reported below.) It should be noted in this regard that already, in the first study of separation carried out by the Social Research Group in 1963, it was found that while only 13 per cent were in favour of “the separation of Quebec from the rest of Canada,” as many as 26 per cent considered “the independence of Quebec” as both “desirable” and “workable.” The problem of validity we raise would also explain why so many “independentists” in the Carlos et al. 1973 study (ibid., 229) voted for other parties than the P.Q. (17 per cent in Montreal, 37 per cent elsewhere) while our own figures are much lower (3 and 12 per cent, respectively). Conversely, about the same proportions of non-independentists voted for the P.Q. in both studies (15 per cent in Montreal and 9 per cent elsewhere in Carlos et al.; 13 and 11 per cent, respectively, in our study). Who are all the “independentists” in Carlos et al. who vote for other parties, in particular for the federalist Liberals? The stronger association between our measure of separatism and a vote for the P.Q. indicates that our measure has greater external validity than theirs. Their data also revealed that no less than 31 per cent of the respondents favourable to independence were also favourable to federalism (recomputed from ibid., 226, Table 5). These data suggest that our question is a more accurate measure of people's options regarding the independence of Quebec. Other data also support this contention, as indicated in the next footnote.

58 A lot has been said in Quebec about the use of one term or another to assess people's options on independence. In this study, a first effort was made to assess empirically the meaning of the terms and the effect of using one rather than another. Apart from the usual question on “separation,” each respondent was asked whether or not the expression “the separation of Quebec” meant the same thing as the expressions “the independence of Quebec”; if s/he answered “no,” two questions followed, asking (1) what was the difference between the two terms; and (2) whether s/he was “for or against Quebec becoming an independent country.” (Notice that the latter question did not ask: for or against the independence of Quebec.) Fifty-seven per cent answered that the two terms meant the same thing, while 30 per cent said different things and 13 per cent said they did not know. Among those who answered “different things,” the answers to the open-ended question asking to state what the difference was revealed a very great degree of confusion, vagueness, or ignorance, and these answers were difficult to code in great detail. The code used, however, was aimed at finding whether one of the terms implied a more radical or complete break away from Canada than the other; the following results were obtained. First no less than 46 per cent answered “don't know” or gave answers which were so vague or unclear as to be impossible to code. For an additional 15 per cent, only one of the terms was dealt with (separation described as complete division, 7 per cent, or not, 1 per cent; independence as a complete division, 4 per cent, or not, 3 per cent). For the rest, the great majority (31 per cent) gave answers implying that independence was less radical than separation, while only 7 per cent gave the opposite answer. The ways in which independence was seen as a less complete division was again not made entirely clear. For some, independence seemed to imply only provincial autonomy; for others, some form of continuing association with Canada or even explicitly sovereignty-association. (Here are examples of these responses: “Indépendant—pas séparé des autres provinces; séparation (signifie qu'on se) séparerait d'Ottawa et des autres provinces.” “Séparation du Québec—pays différent; indépendance—pas nécessairement un pays différent.” “L'indépendance, on reste affilié avec le Canada; séparation veut faire un pays.” “La séparation représente le divorce complet du fédéral; l'indépendance représente l'autonomie du Québec”) At a minimum, therefore, it appears that for some respondents, to be in favour of the “independence of Quebec” means less than the official P.Q. option. At any rate, if for those who said the two terms were different, we take their answer to the last question (for or against “Quebec becoming an independent country” rather than for or against “separation”), the net gain of people favourable to independence is only 2 per cent, raising the proportion of independentists to 19 per cent, while 63 per cent would be opposed and 18 per cent undecided (N = 1,006). These results are consistent with those of Table 8 and seem to indicate that except for those who are ideologically very sensitive, terms like “separation,” “independent country” (but not “independence” used alone) are rather equivalent stimuli, valid to measure the P.Q. option.

59 The proportion of separatists the P.Q. loses is about equal to the proportion of non-separatists it gains, but the absolute number of non-separatists being larger, the P.Q. makes net gains.

60 It should be added, incidentally, that those in favour of separation are not only more likely to vote, but are also more likely to reveal their preference to an interviewer. In our data, only 13 per cent of those in favour of separation and who said they voted did not reveal their choice, as opposed to 38 per cent among those opposed to separation. Thus when considering only those who voted and revealed their preference, the data indicate that 25 per cent are in favour of separatism. This contributes to the slight overestimation of the P.Q. vote in our poll (see footnote 2). (The data reported in the last paragraph and in this note include all respondents, not only French Canadians.)

61 The two could also of course be the result of third factors.

62 N = 74. The 18 per cent is a minimum since some of those who said they had been in favour of separation “between 2 and 5 years” could have become so after the 1970 election.

63 In a 1974 Quebec poll, 53 per cent of the respondents said that the P.Q. chances of coming to power would increase “if the P.Q. were to decide to remove the idea of Quebec's independence from its platform,” while only 14 per cent said they would decrease, the rest saying they would not change (22 per cent) or D. K. (11 per cent). La Presse, November 14, 1974, A13.

64 But recall the fact that independentists are more likely to reveal their vote (see footnote 59 above), so that the non-separatist vote of the P.Q. is possibly underestimated here. The few non-French P.Q. voters, as we have seen, were all in favour of independence.

65 For the undecided these patterns generally hold, except for education, where the proportion of undecided voting P.Q. increases with education.

66 Based on Sarah Samuels’ unpublished analysis of data from the 1970 study of the independentist movement.

67 With unemployment, the patterns are the same, except that there is no reversal for the P.Q., but only a substantial decrease in the size of the percentage differences.

68 See Hamilton and Pinard, “The Bases of Parti Québécois Support,” 21ff.

69 It should be noted, however, that if social conservatives were more likely to vote for these third parties, this was not so among economic conservatives. For instance, there is little relationship between support for third parties and answers to a question asking whether our governments should do more to reduce the differences of incomes between the rich and poor.

70 The Changing American Voter, 157ff.

71 It should be noted that no effort was made in this paper to relate the growth of the Parti Québécois to general conditions for the rise of third parties; this was done elsewhere (see Postscript to The Rise of a Third Party, esp. 278ff.). Similarly, while the independence issue largely explains the P.Q. vote, there remains the important and complex task of explaining support for independence itself; unfortunately, this could not be attempted here. For a recent effort in this direction and a review of previous literature, see Cuneo, Carl J. and Curtis, James E., “Quebec Separatism: An Analysis of Determinants within Social-Class Levels,” The Canadian Review of Sociology and Anthropology 11 (1974), 1–29.CrossRefGoogle Scholar