Hostname: page-component-78c5997874-xbtfd Total loading time: 0 Render date: 2024-11-14T22:41:26.618Z Has data issue: false hasContentIssue false
  • English
  • Français

The Primacy Principle: Attitude Change and Political Socialization

Published online by Cambridge University Press:  27 January 2009

Donald Searing ,
Gerald Wright  and
George Rabinowitz
Get access Rights & Permissions [Opens in a new window]

Extract

The ‘primacy principle’ comprises three assumptions about political orientations. The first is that they are learned during childhood. The second is that this childhood learning further shapes any subsequent modifications of them. The third is that the scale of any such subsequent modifications is small: fundamental political orientations tend to endure through life. We propose, using cohort analysis, to examine the extent to which three political orientations – party identification, political efficacy and political trust – do, as a matter of fact, endure through adulthood.

Type
Articles
Information
British Journal of Political Science , Volume 6 , Issue 1 , January 1976 , pp. 83 - 113
Copyright
Copyright © Cambridge University Press 1976

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 Easton, David, ‘The Theoretical Relevance of Political Socialization’, Canadian Journal of Political Science, 1 (1968), 125–46, pp. 131–3.CrossRefGoogle Scholar

2 Sources for these statements are, in order, Dawson, Richard E. and Prewitt, Kenneth, Political Socialization (Boston, Mass.: Little, Brown, 1969), p. 96Google Scholar; Greenstein, Fred I., ‘Political Socialization’, in Sills, David L., ed., International Encyclopedia of the Social Sciences (New York: Macmillan and Free Press, 1968), pp. 554–5Google Scholar; and Dennis, Jack, ‘Major Problems of Political Socialization Research’, Midwest Journal of Political Science, XII (1968), 85–114, p. 99.CrossRefGoogle Scholar Similar comments may be found in Langten, Kenneth P., Political Socialization (New York: Oxford University Press, 1969), p. 19Google Scholar; Greenstein, Fred I., Children and Politics (New Haven, Conn.: Yale University Press, 1965), pp. 74–8Google Scholar; Easton, David and Dennis, Jack, Children in the Political System (New York: McGraw-Hill, 1969), pp. 9, 75, 107Google Scholar; and Hess, Robert D. and Torney, Judith V., The Development of Political Attitudes in Children (Chicago: Aldine, 1967), p. 220.Google Scholar For general discussions of the primacy principle, see, Weissberg, Robert, Political Learning, Political Choice and Democratic Citizenship (Englewood Cliffs, N.J.: Prentice Hall, 1974)Google Scholar; Niemi, Richard G., ‘Political Socialization’, in Knutson, Jeanne N., ed., Handbook of Political Psychology (San Francisco: Jossey-Bass, 1973), pp. 117–38Google Scholar; Marsh, David, ‘Political Socialization: The Implicit Assumptions Questioned’, British Journal of Political Science, I (1971), 453–65CrossRefGoogle Scholar; Greenstein, Fred I., ‘A Note on the Ambiguity of “Political Socialization”: Definitions, Criticisms and Strategies of Inquiry’, Journal of Politics, XXXII (1970), 969–78CrossRefGoogle Scholar; and Dawson, and Prewitt, , Political Socialization, Chap. 2.Google Scholar The attitude typology described below is stated most concisely in this chapter.

3 In his presidential address, E. Lowell Kelly commented: ‘On perhaps no other major issue do widely variant psychological theories lead to such congruent predictions. Whether one is an extreme hereditarian, an environmentalist, a constitutionalist, or an orthodox psychoanalyst, he is not likely to anticipate major changes in personality after the first few years of life’. ‘Consistency of the Adult Personality’, The American Psychologist, x, (1955), 659–81.Google Scholar See the data presented by Kelly in the same article, and by Tuddenham, Read D., ‘The Constancy of Personality Ratings Over Two Decades’, Genetic Psychology Monographs, LX (1959), 3–29Google Scholar; and Kagen, Jerome and Moss, Howard A., From Birth to Maturity (New York: Wiley, 1962).Google Scholar

4 Miller, Arthur H., ‘Political Issues and Trust in Government: 1964–1970’, American Political Science Review, LXVIII (1974), 951–72.CrossRefGoogle Scholar This comment is not in any sense a criticism of Miller's essay. Philip E. Converse has similarly discussed aggregate changes in party identification and political efficacy: ‘Change in the American Electorate’, in Campbell, Angus and Converse, Philip E., eds., The Human Meaning of Social Change (New York: Russell Sage, 1973).Google Scholar On the relevance of adult political socialization, see Almond, Gabriel and Verba, Sydney, The Civic Culture (Princeton, N.J.: Princeton University Press, 1963), p. 373CrossRefGoogle Scholar; and Greenstein, , ‘Political Socialization’.Google Scholar

5 Dawson, and Prewitt, , Political Socialization, p. 217.Google Scholar See also Dennis, , ‘Major Problems of Political Socialization Research’, pp. 99–101.Google Scholar

6 There is, of course, no contradiction between this inference and the primacy assumption that basic orientations persist relatively unchanged. Development may take place throughout childhood and adolescence. Endurance is to be anticipated primarily during the adult years. Longitudinal change during elementary school is assessed with cross-sectional data by comparing children's responses across school grades. See Langten, , Political Socialization, p. 39.Google Scholar The same design has been applied to student-teacher comparisons by Hess, and Torney, , Development of Political Attitudes in ChildrenGoogle Scholar; to student-parent comparisons by Jennings, M. Kent and Niemi, Richard G., ‘Patterns of Political Learning’, Harvard Educational Review, XXXVIII (1968), 443–67CrossRefGoogle Scholar; and to comparisons between students and samples of the adult population by Greenstein, , Children and Politics, p. 105Google Scholar, and Easton, and Dennis, , Children in the Political System.Google Scholar For a study assessing the accuracy of cross-sectional data, see Glenn, Norval D., ‘Aging, Disengagement and Opinionation’, Public Opinion Quarterly, XXXIII (1969), 17–33.CrossRefGoogle Scholar

7 Glenn, Norval D., ‘Problems of Comparability in Trend Studies With Opinion Poll Data’, Public Opinion Quarterly, XXXIV (1970), 82–91CrossRefGoogle Scholar; Riley, Matilda W., ‘Aging and Cohort Succession: Interpretations and Misinterpretations’, Public Opinion Quarterly, XXXVII (1973), 35–49CrossRefGoogle Scholar; and Evans, William M., ‘Cohort Analysis of Survey Data: A Procedure for Studying Long-term Opinion Change’, Public Opinion Quarterly, XXIII (1959), 63–72.CrossRefGoogle Scholar Several difficulties are apparent in using cohort analysis as a quasi-longitudinal technique. First of all, fluctuations over time may be due, in part, to variation in sampling procedure. Secondly, cohort composition can change somewhat through differential birth rates, death and migration. Finally, patterns of individual change may be obscured by stability or change in the group measure. For example, when group distributions do not change much over time, individuals within the group might nonetheless be changing in different directions in such a way as to ‘cancel-out’ one another. Likewise, when the group distribution does change over time for two orientations, we are forced to assume a constant rate of ‘cancelling-out’ in order to compare them. Despite these problems, the assumptions of cohort analysis are a marked improvement over those of the alternative cross-section designs interpreted longitudinally. Until panel studies over reasonable time spans are available, cohort analysis is the most desirable technique for investigating long-term attitude change. Our analysis in the present section draws upon suggestions put forward by Cutler, Neal E., The Alternative Effects of Generations and Aging upon Political behavior (Oak Ridge, Tennessee: Oak Ridge National Laboratory, 1968).Google Scholar

8 For some purposes it is unnecessary to construct cohorts whose age spans correspond to survey intervals. See Glenn, Norval D. and Hefner, Ted, ‘Further Evidence on Aging and Party Identification’, Public Opinion Quarterly, xxxvi (1972), 31–47CrossRefGoogle Scholar; and Klecka, William R., ‘Applying Political Generations to the Study of Political Behavior: A Cohort Analysis’, Public Opinion Quarterly, XXXV (1971), 358–74.CrossRefGoogle Scholar

9 Campbell, Angus, Gurin, Gerald and Miller, Warren E., The Voter Decides (Evanston, Ill.: Row, Peterson, 1954), pp. 88–9Google Scholar; Campbell, Angus, Converse, Philip E., Miller, Warren E. and Stokes, Donald E., The American Voter (New York: Wiley, 1960), p. 121Google Scholar; Greenstein, , Children and Politics, pp. 73–4Google Scholar; Langten, , Political Socialization, p. 10.Google Scholar

10 Hess, and Torney, , Development of Political Attitudes in Children, p. 9Google Scholar; Jennings, and Niemi, , ‘Patterns of Political Learning’, pp. 452, 54Google Scholar; Campbell, et al. , The Voter Decides, pp. 102–3Google Scholar; Campbell, et al. , The American Voter, p. 148.Google Scholar

11 Converse, , ‘Change in the American Electorate’Google Scholar; Greenstein, , Children and Politics, pp. 71–3Google Scholar; Hess, and Torney, , Development of Political Attitudes in Children, p. 196Google Scholar; Dennis, Jack and McCrone, Donald J., ‘Preadult Development of Political Party Identification in Western Democracies’, Comparative Political Studies, in (1970), 243–63Google Scholar; and Campbell, et al. , The American Voter, pp. 135, 164.Google Scholar

12 See Hess, and Torney, , Development of Political Attitudes in Children, p. 90Google Scholar; Langton, , Political Socialization, p. 52.Google Scholar

13 Campbell, et al. , The Voter Decides, p. 99Google Scholar; and Campbell, et al. , The American Voter, pp. 98, 147.Google Scholar Obviously, there are also cases where parents do not have the same party preference. Further inquiry in such cases has shown that the mother's influence is somewhat greater than the father's: Langten, , Political Socialization, Chap. 3.Google Scholar Parent-child comparisons are reported by Jennings, M. Kent and Niemi, Richard G., ‘The Transmission of Political Values from Parent to Child’, American Political Science Review, LXII (1968), 172–4Google Scholar; and Hyman, Herbert H., Political Socialization (New York: Free Press, 1959), PP. 70, 74.Google Scholar In parent-child comparisons, party identification is the outstanding exception to a general pattern of low correlations using political attitude items: Connell, R. W., ‘Political Socialization in the American Family: The Evidence Re-examined’, Public Opinion Quarterly, xxxvi (1972), 323–33, p. 329.CrossRefGoogle Scholar

14 This is relevant to the persistence thesis because of the strong correlation between political party identification and vote for President. It seems reasonable, then, to explore persistence of party identification by examining congruence between present party identification and first vote for President. Campbell, et al. , The Voter Decides, pp. 101–2Google Scholar; Campbell, et al. , The American Voter, p. 148.Google Scholar

15 Campbell, et al. , The Voter Decides, pp. 93–4Google Scholar; Campbell, et al. , The American Voter, pp. 120–4.Google Scholar

16 Moreover, these American party identification findings are generalizable only to situations fulfilling conditions suggested by cross-national data on transmitting the orientation from parents to children, as well as by data on its continued endurance throughout life. See Greenstein, Fred I. and Tarrow, Sidney G., ‘The Study of French Political Socialization: Toward the Revocation of Paradox’, World Politics, xxii (1969–1970), 95–137, pp. 113–15CrossRefGoogle Scholar; Dennis, and McCrone, , ‘Pre-adult Development’, pp. 248–57Google Scholar; Campbell, Angus and Valen, Henry, ‘Party Identification in Norway and the United States’, in Campbell, Angus, Converse, Philip, Miller, Warren E. and Stokes, Donald E., Elections and the Political Order (New York: Wiley, 1966), pp. 245–68Google Scholar; Cameron, David R. and Summers, Laura, ‘Non-Family Agents of Political Socialization: A Reassessment of Converse and Dupeux’, Canadian Journal of Political Science, v (1972), 418–32CrossRefGoogle Scholar; Converse, Philip E., ‘Of Time and Partisan Stability’, Comparative Political Studies, II (1969), 139–72CrossRefGoogle Scholar; and Kubota, Akira and Ward, Robert E., ‘Family Influence and Political Socialization in Japan: Some Preliminary Findings in Comparative Perspective’, Comparative Political Studies, III (1970), 140–75, p. 145.CrossRefGoogle Scholar

17 In each age group the n is approximately 100. With two groups of 100 each a difference of proportions of 13 per cent is required to reject the null hypothesis at the ·05 level of confidence. If there are no real differences in the entire population, we should still expect that one in twenty values will differ by more than 13 per cent. With our data two or three values in this range are to be expected as a consequence of sampling error.

We do not report rates of change of Independents for two reasons. First, the psychological membership implications of ‘Independent’ do not connote an orientation of the same order as Democratic or Republican identification. It is somewhat like declaring oneself a pagan, defined as neither Christian nor Jew. Secondly, we are interested in describing and explaining observed changes, not in what people change to, whether this is ‘Independent’ or ‘other’.

18 See Hess, and Torney, , Development of Political Attitudes in Children, pp. 69–83Google Scholar; Easton, David and Dennis, Jack, ‘The Child's Acquisition of Regime Norms: Political Efficacy’, American Political Science Review, LXI (1967), 25–38CrossRefGoogle Scholar; and Campbell, et al. , The Voter Decides, p. 187.Google Scholar

19 Easton, and Dennis, , ‘Child's Acquisition of Regime Norms’, p. 31.Google Scholar Not only is political efficacy learned early, but it is important in subsequent political behavior as well. Adults who perceive themselves as particularly efficacious are more likely to participate in elections – independently of the usual demographic correlates of political involvement. These findings buttress the view of political efficacy as an enduring psychological attribute which individuals rely upon for dealing with new political stimuli and situations. See Campbell, et al. , The Voter Decides, pp. 190–3Google Scholar; Campbell, et al. , The American Voter, pp. 104–5.Google Scholar

20 Hess, and Torney, , Development of Political Attitudes in Children, pp. 78–9.Google Scholar A deviation from this pattern has been reported for American blacks who begin with lower efficacy levels than whites, and continue to develop political efficacy during secondary school. Lyons, Schley R., ‘The Political Socialization of Ghetto Children: Efficacy and Cynicism’, Journal of Politics, XXXII (1970), 288–304, p. 294.CrossRefGoogle Scholar

21 Hess, and Torney, , Development of Political Attitudes in Children, p. 55Google Scholar, and Chap. 3; Greenstein, , Children and Politics, Chap. 3.Google Scholar This relationship implies another between personal efficacy and political efficacy. The notion is supported in the Civic Culture where, for each national sample, subjective competence (an efficacy measure for local government) is higher among those who say they participate in family, school or job decisions, with job participation being the most important factor. Using United States and Caribbean data, Langton corroborates these results, though with a different emphasis: it is politicized families which are particularly important for translating ego strength (personal efficacy) into feelings of political efficacy. In contrast to the emphasis on job participation or family politicization, Hess and Torney argue that schools serve as the critical socialization agents for political efficacy. The socialization agent with the strongest independent consequences for political efficacy's development will affect both the extent of intergenerational transmission and the degree of long-term stability for these feelings See Almond, and Verba, , Civic Culture, pp. 347–68Google Scholar; Langten, , Political Socialization, pp. 143–58Google Scholar; and Hess, and Torney, , Development of Political Attitudes in Children, pp. 126–32.Google Scholar

22 Hess, and Torney, , Development of Political Attitudes in Children, pp. 10, 78.Google Scholar

23 Langton, Kenneth P. and Jennings, M. Kent, ‘Political Socialization and the High School Civics Curriculum in the United States’, American Political Science Review, LXII (1968), 857–8.Google Scholar An exception to this pattern was the black sub-sample of high-school seniors for whom civics courses were moderately related to political efficacy scores. The author suggests that the civics courses' non-effectiveness for the white majority of the population is explained by redundancy of the material presented. There is very little in these courses which'has not already come to the middle-class youngster by way of previous instruction, or through the family and the media. Jamaican data are presented in Langton, Political Socialization, pp. 40–1.Google Scholar

24 Campbell, et al. , The Voter Decides, p. 191.Google Scholar

23 Lane, Robert E., Political Life (New York: Free Press, 1959), p. 151.Google Scholar

26 See, for instance, Lane's extended discussion of political efficacy and ego strength in his Political Life, pp. 147–55.Google Scholar Campbell et al., refer to political efficacy as one of a set of ‘broader and more enduring’ orientations (The Voter Decides, p. 187Google Scholar), or ‘as lying at a relatively “deep” level’, and as ‘highly generalized orientations toward the world of polities' (The American Voter, p. 516Google Scholar). If sense of political efficacy does lie at a relatively deep psychological level, then it seems reasonable to expect durability. Many of the discussions we have seen do indeed imply that political efficacy is embedded within the individual's psychic makeup. This is in contrast to political trust which is more often spoken of as a manifestation of personal trust. The distinction is important in that orientations integrated with the ‘political self should not be overly affected by political events. If, by contrast, the orientation is merely a derivative of stable psychological dimensions, it can maintain these ties and still be moved considerably by Zeitgeist factors. A report by Philip E. Converse would seem to support this view by demonstrating efficacy changes for the population as a whole during the ‘sixties (‘Change in the American Electorate’).

27 Dennis et al. found that responses to these items developed rather differently from one another at successive grade levels in four Western democracies. Combination of the items would conceal considerable complexity in the patterns the authors discovered (‘Political Socialization to Democratic Orientations in Four Western Systems’). On the same point, see also Converse, , ‘Change in the American Electorate’.Google Scholar

28 The six-year difference between the 1958 and the 1964 surveys raised the problem of constructing cohorts whose age boundaries (four years) would remain equal to the time elapsed between surveys. Our solution was to group the 1958 respondents such that in 1960 they would fall within our established age categories. For instance, age group one bounds the 21–24-year-olds. Respondents who, in 1958, were aged 19–22 were put into this group. Hence, they become 21–24 in 1960, 25–28 in 1964 and so on. We have in effect constructed the groups so that what ideally would have been the 1960 measurement was taken two years early. This does not cause serious problems for the cohort analysis – cohort boundaries are accurate. Some error is, however, introduced in the aging tests since 1958 respondents are actually two years younger than the age group into which they are classified. Nevertheless, this procedure seems preferable to distorting the cohorts.

29 Greenstein, , Children and Politics, p. 32Google Scholar; Hess, and Torney, , Development of Political Attitudes in Children, p. 129Google Scholar; Easton, and Dennis, , Children in the Political System, p. 137Google Scholar; Jennings, and Niemi, , ‘The Transmission of Political Values from Parent to Child’, p. 178Google Scholar; Jennings, and Niemi, , ‘Patterns of Political Learning’, p. 463.Google Scholar Robert Weissberg has suggested that this may be changing somewhat as a new generation of adolescents shows considerably less political trust than did its predecessor: ‘Adolescents' Perceptions of Political Authorities: Another Look at Political Virtue and Power’, Midwest Journal of Political Science, XVI (1972), 147–68.Google Scholar

30 Agger, Robert E., Goldstein, Marshall N. and Pearl, Stanley A., ‘Political Cynicism: Measurement and Meaning’, Journal of Politics, XXIII (1961), 477–506CrossRefGoogle Scholar; Litt, Edgar, ‘Political Cynicism and Political Futility’, Journal of Politics, XXV (1963), 312–23CrossRefGoogle Scholar; and Aberbach, Joel, ‘Alienation and Political behavior’, American Political Science Review, LXIII (1969), p. 92n.Google Scholar

31 Only two conditions are necessary for high correlations between a very stable and less stable attitude. First, each measurement must be restricted to a single cross-sectional measurement. Secondly, the source of instability must be systematically related to this and affect all age groups in a similar manner.

32 Easton, and Dennis, , Children in the Political System, pp. 298–307Google Scholar; Jennings, and Niemi, , ‘Patterns of Political Learning’, p. 463Google Scholar; and Agger, et al. , ‘Political Cynicism: Measurement and Meaning’Google Scholar.

33 Dawson, and Prewitt, , Political Socialization, pp. 22–4.Google Scholar

34 Though a decided improvement over cross-sectional data, cohort analysis is prone to underestimate actual attitude change. Since it does not treat the same individuals across time, it is blind to changes hidden by ‘cancelling out’ effects. In 1956–60, the Survey Research Center did a panel study which included the same party identification and efficacy items as used in our cohort analysis. Comparing the cohort and panel results for this period shows that change in party identification is underestimated by 11 per cent, and political efficacy by 12 per cent.

35 Change in religious identification seems an indicator of basic attachments whether measured by percent Catholic, Jewish or Protestant. While we have chosen to use percent Catholic, Jewish identification would also be theoretically appropriate, as would the broad class ‘Protestant’ – but not specific Protestant denominations.

With issue beliefs, item wording changes across more than two surveys are often minor, and might be disregarded for some purposes. But they are unacceptable in the present context. This is so because the potential error would artificially inflate issue-belief change rates, making it easier for political orientations to appear comparatively stable. The alternative we have chosen is to limit the time span to 1964–68, using identical items. Issue beliefs likely undergo their maximum and minimum changes at different times. Hence, deriving a mean for nine such items from one survey interval seems to provide a not unreasonable comparison with religious identification and political orientations taken over longer time spans. Issue items used in the present analysis are: government aid housing and utilities; government guarantee work, government aid school building; foreign aid; government promote integration (jobs and housing); military aid to anti-Communist nations; government prevent unions influencing government; government promote school integration; and government aid health care.

36 Comparing the 1956–60 cohort and panel data we find that cohort scores underestimate real change by the following magnitudes: religion, 0 per cent (gamma ·94); political party identification, 11 per cent (gamma ·79); political efficacy, 12 per cent (gamma ·69); and issue beliefs, 32 per cent (gamma ·31). Plainly, issue-belief items mask much more actual change than do the political orientations, while religious affiliation masks considerably less. In other words, if the figure were redrawn using panel data, the spread between the three belief types would likely be widened impressively. See the data presented by Converse, Philip E., ‘The Nature of Belief Systems in Mass Publics’, in Apter, David E., ed., Ideology and Discontent (New York: Free Press, 1964), p. 240.Google Scholar It might be argued that the period chosen for our issue-belief measure (1964–68) is a poor one, that it is atypical because these were years of unusual political change for issue beliefs. Yet they were years of considerable orientation change as well. And the orientations changed less than the issue items.

37 Hyman, , Political Socialization, p. 124.Google Scholar

38 It is often difficult to specify the age groups involved. Many analysts have singled out generations entering the occupational structure or the electorate. Yet there are no convincing theoretical reasons for always focusing on these groups rather than others. Impressionable cohorts are likely unique to the character of the orientation and historical event involved in each case. For instance, the Second World War may have been as influential for young children who were without their fathers as for men and women who served in the military.

39 Campbell, et al. , The American Voter, pp. 153–6.Google Scholar

40 This does not necessarily invalidate the argument that this age group was extraordinarily strong and stable in its Democratic identification from the 1930s through 1960. These voters were, in fact, the strongest Democratic identifiers in 1952 and 1956. And data presented by Glenn and Hefner show that they were the strongest Democratic identifiers in 1945 as well: ‘Further Evidence on Aging and Party Identification’, pp. 36–7.Google Scholar

41 Lipset, Seymour Martin, Political Man (Garden City, N.Y.: Doubleday, 1960), 279–86Google Scholar; and Hyman, , Political Socialization, Chap. 6.Google Scholar

42 Easton, and Dennis, , Children in the Political System, pp. 306–10.Google Scholar

43 Hyman, , Political Socialization, pp. 123–55.Google Scholar Age-related roles, of course, can differ in different societies. But within one society, this concept presumes some role continuity over time. It has been argued that aging is an unsatisfactory concept because it poses as a surrogate variable, disguising a host of unnamed socialization experiences: Prewitt, Kenneth, von der Muhll, George and Court, David, ‘School Experiences and Political Socialization: A Study of Tanzanian Secondary School Students’, Comparative Political Studies, III (1970), 203–25, pp. 205–6.CrossRefGoogle Scholar Still, the term has utility so long as we are aware that its domain is restricted to physiological developments and other age-related variables. It serves to distinguish such variables from diffuse cultural sources of attitude change. In this sense we concur with Hans Toch that aging phenomena are bona fide psychological variables: ‘Attitudes of the “Fifty Plus” Age Group: Preliminary Considerations Toward a Longitudinal Survey’, Public Opinion Quarterly, XVII (1953–1954), 391–4, P. 392.Google Scholar

44 Among such specific variables political crises and even changes in government administration or in level of education are regarded as especially important. For a discussion of Zeitgeist-induced change see Hyman, , Political Socialization, Chap. 6.Google Scholar

45 Exploring aging and Zeitgeist effects through cohort analysis involves many methodological difficulties which have often been overlooked: Mason, Karen O., Mason, William M., Winsborough, H. H. and Poole, W. Kenneth, ‘Some Methodological Issues in Cohort Analysis of Archival Data’, American Sociological Review, XXXVIII (1973), 242–58.CrossRefGoogle Scholar After much experimentation, it seems to us that operationalizing the relative influence of aging and Zeitgeist in terms of explanatory power is the only technically feasible option. Had we been able to ‘divine’ the correct bAi and bZj coefficients, we might have chosen to compare these instead. But we are unable to do this. In any event, for most empirical instances, use of either explanatory power or the relative magnitude of the correct b coefficients will produce consistent results. When one set of the correct coefficients, either Zeitgeist or aging, has a mean squared value twice that of the other, results between the two approaches will always be consistent. Since it seems impossible to analyze even the simple linear model, we obviously cannot analyze a more complex model including interaction terms. From this perspective, it should be noted that we are not interested in analyzing variance in change, but rather in analyzing the change itself. ‘Percent change accounted for’ is conceptually equivalent to percent variance accounted for in traditional analyses. It is defined as 100 times the following:

where is the predicted change and CAiZj is the observed change.

46 Campbell, et al. , The American Voter, pp. 160–6Google Scholar; Campbell, and Valen, , ‘Party Identification in Norway and the United States’, pp. 250–1.Google Scholar

47 Easton, and Dennis, , Children in the Political System, p. 139Google Scholar; Lane, , Political Life, p. 151.Google Scholar One recent study, which places greater weight on the potential impact of political events, provides an exception to this general picture: Converse, ‘Change in the American Electorate’. Moreover, it has been argued that evidence for political efficacy's endurance ignores the very probable effects of differential drop-out rates. But if differential drop-out is a major cause of observed change, it should have a predictable pattern. Presumably, lower SES and lower efficacy voters die at a faster rate than do their counterparts with higher scores on these variables. We should therefore expect aging to yield an increase in political efficacy: most observed change should be related to aging rather than Zeitgeist variables. Yet, despite the probable effects of drop-out, it will be found below that Zeitgeist rather than aging factors have the greater explanatory strength.

48 In presenting these distributions graphically, four-year cohorts have been aggregated into these ‘enlarged’ cohorts, which are simple unweighted means of the scores of the cohorts which form them. This was done to simplify the presentation (thirteen lines on a single graph obscure as much as they show) without losing too much information contained in the basic data tables.

49 Converse, , ‘Change in the American Electorate’.Google Scholar

50 Searing, Donald, Schwartz, Joel and Lind, Alden, ‘The Structuring Principle: Political Socialization and Political Belief Systems’, American Political Science Review, LXVII (1973), 415–32.CrossRefGoogle Scholar

51 McGuire, William J., ‘The Nature of Attitudes and Attitude Change’, in Lindzey, Gardner and Aronson, Elliot, eds., The Handbook of Social Psychology, Vol. III, 2nd edn. (Reading, Mass.: Addison-Wesley, 1969), p. 248.Google Scholar

52 Zigler, Edward and Child, Irvin L., ‘Socialization’Google Scholar, in Lindzey, and Aronson, , Handbook of Social Psychology, p. 503.Google Scholar

53 Zigler, and Child, , ‘Socialization’Google Scholar See also Brim, Orville G. Jr., ‘Adult Socialization’Google Scholar, in Sills, , ed., International Encyclopedia of the Social Sciences, Vol. 14, p. 558.Google Scholar

54 Easton, , ‘The Theoretical Relevance of Political Socialization’, p. 141.Google Scholar

55 See Hess, and Torney, , Development of Political Attitudes in Children, p. 9Google Scholar; and Easton, David and Hess, Robert D., ‘The Child's Political World’, Midwest Journal of Political Science, VI (1962), 229–46, p. 236.CrossRefGoogle Scholar