¹ A full statement of our earlier work is Keith, Bruce E. et al. , ‘The Myth of the Independent Voter’Google Scholar, delivered at the 1977 Annual Meeting of the American Political Science Association. Our principal findings had been disseminated at least eighteen months before this in Wolfinger, Raymond E., Shapiro, Martin and Greenstein, Fred I., Dynamics of American Politics (Englewood Cliffs, N.J.: Prentice-Hall, 1976), pp. 133, 151.Google Scholar

² Nie, Norman H., Verba, Sidney and Petrocik, John R., The Changing American Voter, enlarged edn (Cambridge, Mass.: Harvard University Press, 1979), p. 73.CrossRefGoogle Scholar

³ Although party identification is fairly stable, an appreciable number of voters change the strength and even the direction of their party identification in response to their evaluations of presidential candidates and the performance of the parties. See in particular Brody, Richard A., ‘Stability and Change in Party Identification: Presidential to Off-Years’Google Scholar, paper delivered at the 1977 Annual Meeting of the American Political Science Association; and Brody, Richard A. and Rothenberg, Lawrence S., ‘Dynamics of Partisanship During the 1980 Election’, paper delivered at the 1983 Annual Meeting of the American Political Science Association.Google Scholar

⁴ One well-known example is the central place of party identification in calculating the ‘normal vote’, a method used to identify groups that respond in different ways to a given campaign. See Converse, Philip E., ‘The Concept of a Normal Vote’, in Campbell, Angus, Converse, Philip E., Miller, Warren E. and Stokes, Donald E., Elections and the Political Order (New York: Wiley, 1966), pp. 9–39.Google Scholar

⁵ Kinder, Donald R. and Sears, David O., ‘Public Opinion and Political Action’, in Lindzey, Gardner and Aronson, Elliot, eds, Handbook of Social Psychology, vol. II, 3rd edn (New York: Random House, 1985), p. 684.Google Scholar

⁶ Nie, et al. , The Changing American Voter, p. 47.Google Scholar

⁷ See, for example, Campbell, Bruce A., The American Electorate (New York: Holt, Rinehart & Winston. 1979), pp. 267–9, 276.Google Scholar

⁸ Hardly anyone identifies with another party, and a handful of respondents manifest no interest in parties. Those uninterested respondents whom the NES staff classifies as ‘apolitical’ have been excluded from tabulations in this article, except those noted in fn. 15. For more on apolitical respondents, see p. 184. and Miller, Arthur H. and Wattenberg, Martin P., ‘Measuring Party Identification: Independent or No Partisan Preference?’ American Journal of Political Science, XXVII (1983), 106–21.CrossRefGoogle Scholar

⁹ The exceptions to this generalization in the Michigan trilogy are sparse enough to be enumerated in one footnote. The first book from Michigan contained several tables based on the seven-way classification, but did not acknowledge in the text that Democratic leaners were more loyal to Stevenson than Weak Democrats, or the same pattern on the Republican side. See Campbell, Angus, Gurin, Gerald and Miller, Warren E., The Voter Decides (Evanston, III.: Row, Peterson, 1954), pp. 101, 108–9.Google Scholar

This evident lack of interest in differentiating leaners and Pure Independents also appeared in Campbell, Angus, Converse, Philip E., Miller, Warren E. and Stokes, Donald E., The American Voter (New York: Wiley, 1960)Google Scholar. This book presented a great many data analyses in which all varieties of Independents were combined, and just three tables where they were not (pp. 148, 201, 390). Only the first of these bears on the controversy about how to classify leaners; it shows they were more likely than weak partisans to report that they had not always identified with the same party. (We discuss stability of party identification on pp. 166–70.)

The single seven-way table in Campbell, et al. , Elections and the Political Order (p. 218)Google Scholar, concerns one of the few variables that always has a monotonic relationship to strength of party identification – respondents' reports about the partisan consistency of their past voting choices. The same finding appears on p. 125 of The American Voter. But as we will show later (pp. 176–9). this question taps respondents' self-images, not their actual behaviour. Another table in Elections and the Political Order (p. 197) classified leaners with strong and weak partisans in a three-way distribution.

¹⁰ Converse, , ‘The Concept of a Normal Vote’, p. 20n.Google Scholar

¹¹ Campbell, et al. , The American Voter, pp. 122–3.Google Scholar

¹² Petrocik, John R., ‘An Analysis of Intransitivities in the Index of Party Identification’, Political Methodology, I (1974), 31–47.Google Scholar

¹³ Depicted by some writers as a challenger of the conventional wisdom about party identification, Petrocik in fact is one of its most dogged defenders. See his ‘Contextual Sources of Voting Behavior: The Changeable American Voter’, in Pierce, John C. and Sullivan, John L., eds. The Electorate Reconsidered (Beverly Hills: Sage Publications, 1980), pp. 265–6Google Scholar; and Party Coalitions (Chicago: University of Chicago Press, 1981), pp. 158–9.Google Scholar

¹⁴ Pomper, Gerald M., Voters' Choice (New York: Harper & Row. 1975). p. 40.Google Scholar

¹⁵ ‘Apolitical’ respondents accounted for about a sixth of all blacks (and a tiny fraction of whites) until 1964. Since then the proportions of ‘apolitical’ blacks and whites have been similar. ‘Apoliticals’ were included in the bases of the computations reported in the text comparing white and black party identification. White Southerners, also deleted from our 1977 paper, are included in this article. Excluding them did not alter our findings.

¹⁶ Nie, et al. , The Changing American Voter, p. 50.Google Scholar

¹⁷ Mann, Thomas E. and Wolfinger, Raymond E., ‘Candidates and Parties in Congressional Elections’, American Political Science Review, LXXIV (1980), 617–32, pp. 621–2CrossRefGoogle Scholar; and Arseneau, Robert B. and Wolfinger, Raymond E., ‘Voting Behavior in Congressional Elections’, paper delivered at the 1973 Annual Meeting of the American Political Science Association.Google Scholar

¹⁸ See particularly Petrocik, , ‘Contextual Sources of Voting Behavior’.Google Scholar

¹⁹ In the fall of 1976 the CBS/New York Times polls divided leaners and Pure Independents. Newsweek and the NBC/Associated Press surveys did this during the 1980 election season. One advantage of this practice was realized in the 1980 pre-convention period: classifying leaners as partisans made it easier for pollsters to accumulate enough cases for analysis of intraparty candidate preferences.

The Gallup Organization, on the other hand, still omitted any question about leaning (or strength of partisanship), and so continued to report that over 30 per cent of all Americans were Independents. A Gallup release in the spring of 1983 appeared under this headline: ‘Republican Affiliation Falls to 1 in 4’. (San Francisco Chronicle, 10 03 1983, p. 13.)Google Scholar

²⁰ Additional evidence about the partisan character of leaners can be found in Abramson, Paul R., Political Attitudes in America (San Francisco: W. H. Freeman, 1983), pp. 80–2, 107–10Google Scholar; Abramson, Paul R., Aldrich, John H. and Rohde, David W., Change and Continuity in the 1980 Elections, revised edn (Washington: CQ Press, 1983), pp. 169–73Google Scholar; Asher, Herbert, Presidential Elections and American Politics, revised edn (Homewood, Ill.: Dorsey Press, 1980), pp. 61–2Google Scholar; Miller, Warren E., ‘Disinterest, Disaffection, and Participation in Presidential Politics’, Political Behavior, II (1980). 20Google Scholar; Miller, Arthur H. and Miller, Warren E., ‘Partisanship and Performance: ‘Rational’ Choice in the 1976 Presidential Election’Google Scholar, paper delivered at the 1977 Annual Meeting of the American Political Science Association; and LeBlanc, Hugh L. and Merrin, Mary Beth, ‘Independents, Issue Partisanship and the Decline of Party’, American Politics Quarterly, VII (1979), 240–56.CrossRefGoogle Scholar

²¹ Both quotations are from Crotty, William, American Parties in Decline, 2nd edn (Boston: Little, Brown, 1984), pp. 42, 28–30Google Scholar. Like some other scholars with his point of view, Crotty presents tables showing differences between leaners and Pure Independents that he ignores in the text.

Other recent books that combine leaners and Pure Independents include Keefe, William et al. , American Democracy (Homewood, Ill.: Dorsey Press, 1983), p. 157Google Scholar; Ladd, Everett Carll Jr, Where Have All the Voters Gone?, 2nd edn (New York: Harper & Row, 1982), pp. 77–8Google Scholar; Phillips, Kevin P., Post-Conservative America (New York: Vintage Books, 1982), pp. 227–8Google Scholar; Prewitt, Kenneth and Verba, Sidney, An Introduction to American Government, 4th edn (New York: Harper & Row, 1983). pp. 392–3, 400, 402Google Scholar; and Sundquist, James L., Dynamics of the Party System (Washington. DC: Brookings Institution. 1983). pp. 394–408.Google Scholar

²² Crotty, , American Parties in Decline, p. 37.Google Scholar

²³ Ladd, , Where Have All the Voters Gone?, pp. 124, 77.Google Scholar

²⁴ Abramson, et al. , Change and Continuity in the 1980 Elections, p. 237.Google Scholar

²⁵ New York Times, 20 03 1980, p. A26Google Scholar. Readers will notice that the New York Times and Abramson and his collaborators have been on both sides of the dispute about Independents. Perhaps multiple authorship is the cause in both cases.

In the middle of 1983 John Anderson was reported to believe that the ‘one-third’ of the population who ‘claim no ties’ to a party provided an opportunity for a third-party movement (Will, George F., ‘Is a Third Party Necessary?’ Newsweek, 23 05 1983, p. 88Google Scholar). A year later the Washington Post's director of polling wrote that the ‘between 35 and 40 percent’ of American adults who ‘considered themselves independents’ ‘appear to present fertile ground for the kind of effort Anderson says he is willing to undertake’ (Sussman, Barry, ‘Why Anderson's Third-Party Bid May Hit Responsive Chords’, Washington Post National Weekly Edition, 21 05 1984, p. 37).Google Scholar

²⁶ In addition to Petrocik and Abramson et al., see Beck, Paul Allen, ‘Realignment Begins? The Republican Surge in Florida’Google Scholar, paper delivered at the 1981 Annual Meeting of the American Political Science Association; Eisinger, Peter et al. , American Politics: The People and the Polity (Boston: Little, Brown, 1982). pp. 323–5, 365, 366, 374Google Scholar; and Pomper, Gerald M. and Lederman, Susan S., Elections in America, 2nd edn (New York: Longman, 1980), pp. 56, 67, 74, 99–102.Google Scholar

²⁷ See Ranney, Austin, Curing the Mischiefs of Faction (Berkeley: University of California Press, 1975), pp. 147–69.Google Scholar

²⁸ Converse, Philip E. and Markus, Gregory B., ‘Plus ça change…; The New CPS Election Study Panel’, American Political Science Review, LXXIII (1979), 32–49, p. 33.CrossRefGoogle Scholar

²⁹ Including the respondents' party identification in 1974 would increase the Sense of movement, since some respondents who were in the same partisan category in 1972 and 1976 were somewhere else in 1974.

³⁰ Richard A. Brody has shown that short-term forces from 1972 to 1974 were not kind to the Republican party in this respect. See his ‘Stability and Change in Party Identification’.

³¹ Jennings, M. Kent and Niemi, Richard G., ‘The Persistence of Political Orientations: An Over-Time Analysis of Two Generations’, British Journal of Political Science, VIII (1978), 333–63.CrossRefGoogle Scholar

³² Converse, and Markus, , ‘Plus ça change’, p. 39.Google Scholar

³³ The same finding emerges when one compares stability in two-year periods in the 1972–74–76 panel. The fact that the amount of aggregate change is not proportionate to the amount of time available for change should inhibit anyone who thinks that simple extrapolation of a two-year or four-year change will yield a useful measure of long-term party switching. Limiting the analysis to respondents who voted in both years produces slightly higher figures for directional stability but does not alter the conclusions presented here about the relative stability of leaners and outright partisans.

³⁴ Shively, W. Phillips, ‘The Nature of Party Identification: A Review of Recent Developments’Google Scholar, in Pierce, and Sullivan, , eds, The Electorate Reconsidered, pp. 232–5.Google Scholar

³⁵ Shively, , ‘The Nature of Party Identification’, p. 235.Google Scholar

³⁶ These calculations count as constant voters those people who voted for the same party's presidential candidate in the two elections, even if both votes were for the other party's candidate. That is, Strong Democrats who voted for Nixon in 1972 and Ford in 1976 are not distinguished from Strong Democrats who voted for McGovern and Carter. By the same token, Democrats in 1972 who voted for Nixon and then Carter are mixed in with Democrats who voted for McGovern and Ford. We repeated our analysis after removing those respondents whose 1972 vote did not match their 1972 party identification, defining ‘constant voters’ only as those who voted in both years for the candidate of the party with which they identified in 1972, and ‘floaters’ as those who cast a party-line vote in 1972 and switched to the other party's candidate in 1976. This increased the proportion of constant voters by eight to ten percentage points in each category of party identification. The difference in change in party identification between floaters and constant voters was about the same for leaners and weak partisans, but rose to 32 per cent for the strong partisans.

We also categorized respondents who voted in both elections according to their 1976 party identification and then repeated the original calculation. The results were very similar to those in Table 4, except that the leaners and strong partisans had about the same proportion of constant voters.

³⁷ Differences between leaners and weak partisans on thermometer ratings of ‘Democrats’ and ‘Republicans’ were tiny to nonexistent from 1972 through 1980.

³⁸ Most respondents whose comments were coded ‘±’ had both favourable and unfavourable things to say about the indicated party. This category also includes respondents who had nothing to say about one party and who answered one or both questions about the other party. Readers should note, therefore, that a small number of respondents in the ‘favourable’ categories in Table 6 said nothing about the party which we say they favour; they made only negative responses about the other party.

³⁹ Their attitudes on issues also resemble those of the outright partisans in each party, as we show in The Myth of the Independent Voter (in preparation).

⁴⁰ Campbell, et al. , The American Voter, p. 125.Google Scholar

⁴¹ Nie, et al. , The Changing American Voter, p. 54.Google Scholar

⁴² See, for example, Kritzer, Herbert M. and Eubank, Robert B., ‘Presidential Coattails Revisited: Partisanship and Incumbency Effects’, American Journal of Political Science, XXIII (1979), 615–26, pp. 617–18n.CrossRefGoogle Scholar

⁴³ Another possibility is substantial turnover from one election to the next among people claiming to be Strong Republicans. The data in Table 3 lead us to reject this alternative.

⁴⁴ Brody, , ‘Stability and Change in Party Identification’, p. 23n.Google Scholar

⁴⁵ In 1980, the question did not follow the standard set of items about party identification. Instead, it followed a new battery, also tapping party loyalty, the last item of which asks: ‘Do you ever think of yourself as a political independent, or not?’ Leaners were by far the most likely to respond affirmatively, followed by Pure Independents, as the following figures for each type of identifier demonstrate: SD, 22; WD, 31; ID, 78; I, 64; IR, 79; WR, 33; SR, 27. It is no more surprising than for earlier years, therefore, that when asked to report on their partisan consistency as a sequel to affirming their independence, all three types of Independent in 1980 overwhelmingly reported not having voted for the same party over the years (see Table 8).

⁴⁶ Campbell, et al. , The American Voter, p. 125nGoogle Scholar. The quoted passage refers to Pure Independents' answers to the question about past partisan consistency.

⁴⁷ For more evidence on this point, see Niemi, Richard G., Katz, Richard S. and Newman, David, ‘Reconstructing Past Partisanship: The Failure of the Party Identification Recall Questions’, American Journal of Political Science, XXIV (1980), 633–51.CrossRefGoogle Scholar

⁴⁸ See Katz, Richard S., ‘The Dimensionality of Party Identification: Cross-National Perspectives’, Comparative Politics, X (1979). 147–63CrossRefGoogle Scholar; Valentine, David and Wingen, John Van, ‘Partisanship, Independence, and the Partisan Identification Question.’ American Politics Quarterly, VIII (1980), 165–86CrossRefGoogle Scholar; and Weisberg, Herbert F., ‘A Multidimensional Conceptualization of Party Identification’. Political Behavior, II (1980). 33–60CrossRefGoogle Scholar. Weisberg wrote (pp. 33–4) that party identification ‘involves separate attitudes toward several distinct objects – political parties generally, the Republican party, the Democratic party, and political independence’.

⁴⁹ Valentine, and Van Wingen, , ‘Partisanship. Independence, and the Partisan Identification Question’, p. 170.Google Scholar

⁵⁰ Valentine, and Van Wingen, , ‘Partisanship, Independence, and the Partisan Identification Question’, p. 166.Google Scholar

⁵¹ Dennis, Jack, ‘On Being an Independent Partisan Supporter’, paper delivered at the 1981 Annual Meeting of the Midwest Political Science Association, pp. 5–6.Google Scholar

⁵² ‘On Being an Independent Partisan Supporter’; ‘Some Properties of Measures of Partisanship’, paper delivered at the 1981 Annual Meeting of the American Political Science Association; ‘New Measures of Partisanship in Models of Voting’, paper delivered at the 1982 Annual Meeting of the Midwest Political Science Association; and ‘Toward a Theory of Political Independence’, paper delivered at the 1983 Annual Meeting of the American Political Science Association.

⁵³ Dennis, , ‘On Being an Independent Partisan Supporter’.Google Scholar

⁵⁴ Miller, and Wattenberg, , ‘Measuring Party Identification’, p. 109.Google Scholar