1 Przeworski, A. and Teune, H., The Logic of Comparative Social Inquiry (New York: Wiley, 1971), 20, 22, 25Google Scholar.

2 “Social phenomena do not have a property of ‘being comparable’ or ‘not comparable.’ ‘Comparability’ depends upon the generality of the language that is applied to express observations” (ibid., 10).

3 “Finally, a large part of the dispute about the usefulness of party identification stems from unnecessary disagreements about the theoretical meaning of the concept” (“Party Identification,” 433; emphasis added).

4 See the author's “Party Strategy and Party Identification: Some Patterns of Partisan Allegiance,” this Journal 9 (1976), 27–32, for a discussion of the different place of the concept in two theories of voting; p. 36 describes differences in some characteristics of party identification which might be predicted by one of the conceptualizationsGoogle Scholar.

5 Elkins, “Party Identification,” 421.

6 More than one-quarter of very strong identifiers who changed did so at the time of the last election. As well, of those who report unstable identifications, 25 per cent classify themselves as strong identifiers and 50 per cent as fairly strong. The relationship between intensity of identification and instability is weak, as is the relationship between time of change and intensity (H. Clarke, J. Jenson, L. LeDuc. and J. Pammett, Political Choice in Canada [Toronto: McGraw-Hill Ryerson, forthcoming]. Tables 5.8 and 5.9).

7 “Party Identification,” 433–35, which describes the process of vote change and change in identification as equivalent. The equivalency is also indicated by his use of the concept “fidelity.”

8 Campbell, A., Converse, P., Miller, W., and Stokes, D., The American Voter (New York: Wiley, 1960), 121Google Scholar.

9 Brody, R. A. and Page, B. I., “Comment: An Assessment of Policy Voting,” American Political Science Review 66 (1972), 451 (emphasis added)CrossRefGoogle Scholar.

10 “We are convinced that the relationships in our data reflect primarily the role of enduring partisan commitments in shaping attitudes toward political objects. Our conviction on this point is rooted in what we know of the relative stability and priority in time of party identification and the attitudes it may affect” (Campbell et al.. The American Voter, 135 and also 120).

11 Converse provides an elaborate justification of his argument that where there is agreement of partisan direction between party identification and issue preferences that a party to issue causal flow probably exists. The crux of the justification is the “fantastically stable” identifications that have been observed by means of several measurement techniques (Converse, P., “Public Opinion and Voting Behavior,” in Greenstein, F. and Polsby, N. [eds.], Nongovernmental Politics, Vol. 4 of The Handbook of Political Science [Reading: Addison-Wesley, 1975], 127ff., especially 130–31)Google Scholar.

12 Converse concludes that normal vote analysis is the proper technique (ibid., 132–33). Brody and Page, “Comment,” provide similar justifications for their discussion of the technique.

13 Campbell et al., The American Voter, 135.

14 This is clearly not a black or white case of perfect clarity in causal influence in the theory. However, the authors of The American Voter make it clear that the over-whelming expectation is for changes in proximal forces rather than identifications in cases of potential conflict. See ibid., 128ff.

15 Elkins, “Party Identification,” 434.

16 It is interesting to note that Sniderman, et al, in “Party Loyalty and Electoral Volatility: A Study of the Canadian Party System,” this Journal 8 (1974), seem to suggest the opposite when they point to the lack of innovation in the party system (288)Google Scholar.

17 The American Voter, 136–37.

18 Ibid., 139.

19 See Clarke et al., Political Choice in Canada, chap. 5, for an attempt at such a measure.

20 See ibid, for a discussion of the short-term nature of some party orientations, particularly among partisans with a history of partisan instability.

21 Brody and Page make the contrast clear when they distinguish between “our leading models” and the work of A. S. Goldberg (“Comment,” 451, note 8).

22 Jenson, “Party Strategy and Party Identification.”

23 Elkins, “Party Identification.” 432.

24 This table displays Taus rather than Lambda, which is what Elkins reports because Lambda tends to zero values in certain cases of imbalance in marginals. This occurs in this case for the Canadian parties except for the Liberals. The equivalency of the two measures and their characteristics are discussed in Blalock, H., Social Statistics (2nd ed.; New York: McGraw Hill, 1972), 300ffGoogle Scholar. It was not clear to me what data Elkins had used to calculate his statistics so I returned to the original sources cited in my article for the American data. This is Campbell, A., Gurin, G., and Miller, W., The Voter Decides (Evanston: Row Peterson, 1954)Google Scholar.

25 “Party Loyalty and Electoral Volatility,” 285.

26 Whether one would want to consider this degree of agreement to provide evidence that voters are “lashed” to the party of their parents is something of another matter. It may be more momentary agreement between quite changeable pairs. The available data can not allow any conclusions one way or the other.

27 Sniderman et al., “Party Loyalty and Electoral Volatility,” 288.