1 Campbell, A., Converse, P., Miller, W., Stokes, D., The American Voter (New York: Wiley, 1964), 88Google Scholar; Hyman, H. H., Political Socialization (Glencoe: The Free Press, 1959), 109, 123Google Scholar; Jennings, M. and Niemi, R. G., The Political Character of Adolescence (Princeton: Princeton University Press, 1974), 17Google Scholar; Berelson, B., Lazarsfeld, P., and McPhee, W., Voting (Chicago: University of Chicago Press, 1954), 121Google Scholar; McCloskey, H. and Dahlgren, H., “Primary Group Influence on Party Loyalty,” American Political Science Review 53 (1959), 757–76CrossRefGoogle Scholar; Newcomb, T. M., Personality and Social Change: Attitude Formation in a Student Community (New York: Dryden Press, 1943)Google Scholar; Newcomb, T. M., Koenig, K., Flacks, R., Warwick, O., Persistence and Change: Bennington College and Its Students after 25 Years (New York: Wiley, 1967)Google Scholar.

2 Blalock, H. M. Jr, “Status Inconsistency and Interaction: Some Alternative Models,” American Journal of Sociology 73 (1967), 305–15CrossRefGoogle Scholar, Blalock, H. M. Jr, “The Identification Problem and Theory Building: The Case of Status Inconsistency,” American Sociological Review 31 (1978), 52–61CrossRefGoogle Scholar.

3 Lazarfeld, P., Berelson, B., and Gaudet, H., The People's Choice (New York: Columbia University Press, 1948)Google Scholar.

4 This is fundamentally a cognitive dissonance approach (Festinger, L., A Theory of Cognitive Dissonance [Evanston: Row, Peterson, 1957]Google Scholar) though developed and put forward at an earlier point in time.

5 After reviewing a large body of the available evidence, Hopkins failed to find a single piece of statistically significant evidence to support the view that American men upwardly mobile from the working to the middle class “over-conformed” by becoming more conservative than their new peers (Hopkins, A., “Political Overconformity by Upwardly Mobile American Men,” American Sociological Review 38 (1973), 143–48)CrossRefGoogle Scholar.

6 Among sociologists, the work of Homans dealing with various aspects of social exchange is widely known, and along with Skinner's work (on which Homans' own efforts are quite explicitly based), has generated increasing interest across many sectors of the social sciences (Homans, G. C., Social Behavior: Its Elementary Forms [New York: Harcourt, Brace & World, 1961]Google Scholar and The Human Group [New York: Harcourt, Brace, 1950]Google Scholar). This interest is now manifesting itself in a considerable variety of journal publications including several which have an explicitly political focus: Anderson, C. W. and Nesvold, S. A., “A Skinnerian Analysis of Conflict Behavior: Walden II Goes Cross-National,” American Behavioral Scientist 15 (1972), 883–909CrossRefGoogle Scholar; Cook, T. J. and Scioli, R. P. Jr, “A Critique of the Learning Concept in Political Socialization Research,” Social Science Quarterly 52 (1972), 949–62Google Scholar; Freedman, A. E., “The Planned Society: An Analysis of Skinner's Proposals,” Behaviordelia (1972), 50 pp.Google Scholar; Stillman, P. G., “The Limits of Behaviorism: A Review Essay on B. F. Skinner's Social and Political Thought,” American Political Science Review 69 (1975), 202–13CrossRefGoogle Scholar; Watts, M. W., “B. F. Skinner and the Technological Control of Social Behavior.” American Political Science Review 69 (1975), 214–27CrossRefGoogle Scholar.

7 Staats, A. and Staats, C., Complex Human Behavior (New York: Holt, Rinehart, & Winston, 1963), 36Google Scholar.

8 Research has demonstrated that nearly all the autonomic responses an organism can produce, including many of the accepted physiological indices of emotion (such as gastro-intestinal secretions or vaso-motor reactions), can be conditioned through the respondent learning procedure, and to almost any stimulus object. See for example, Bem, D. J., Beliefs, Attitudes, and Human Affairs (Belmont: Brooks/Cole Publishing, 1970), 42Google Scholar; Bandura, A., Social Learning Theory (Morristown: General Learning Press, 1971), 12Google Scholar; Lovello, W. R., “The Conditioning of Autonomic Responses,” Biological Psychology Bulletin 1 (1971), 18–23Google Scholar.

9 Singh, R., “Reinforcement and Attraction: Specifying the Effects of Affective States,” Journal of Research in Personality 8 (1974), 294–305CrossRefGoogle Scholar; Byrne, D. and Rhamey, R., “Magnitude of Positive and Negative Reinforcements as a Determinant of Attractions,” Journal of Personality and Social Psychology 2 (1965), 884–89CrossRefGoogle Scholar; Lamberth, J., Gay, R., Dyck, D., “Differential Reward Magnitude and Human Conditioning with Social Reinforcers,” Psychonomic Science 28 (1972), 213–33CrossRefGoogle Scholar.

10 Bandura, Social Learning Theory, refers to this process as symbolic conditioning to emphasize its cognitive aspects.

11 Lazarsfeld, Berelson, and Gaudet, The People's Choice; Verba, S., Small Groups and Political Behavior (Princeton: Princeton University Press, 1961)Google Scholar; Campbell, et al., the American Voter; Butler, D. and Stokes, D., Political Change in Britain (New York: St. Martin's Press, 1971)CrossRefGoogle Scholar; McCloskey and Dahlgren, “Primary Group Influence on Party Loyalty”; and Jennings and Niemi, The Political Character of Adolescence are a few of the better known examples.

12 McCloskey and Dahlgren, “Primary Group Influence on Party Loyalty,” Jennings and Niemi, The Political Character of Adolescence; and Fitton, M., “Neighborhood and Voting: A Sociometric Examination,” British Journal of Political Science 3 (1973), 445–72CrossRefGoogle Scholar, have done some excellent research on this area.

13 Jennings and Niemi, The Political Character of Adolescence.

14 Ibid.

15 Braungart, R., “Parental Identification and Student Politics,” Sociology of Education 44 (1971), 462–73CrossRefGoogle Scholar; Winch, R., Identification: Its Familiar Determinants (New York: Bobbs-Merrill, 1962)Google Scholar; Mortimer, J., “Occupational Value Socialization in Business and Professional Families,” Sociology of Work and Occupation 2 (1975), 29–53CrossRefGoogle Scholar; Jennings, M. and Langton, N., “Mothers vs. Fathers: The Formation of Political Orientations Among Young Americans,” Journal of Politics 31 (1969), 329–57CrossRefGoogle Scholar.

16 Jennings and Niemi, The Political Character of Adolescence.

17 The limiting of the model to primary group contacts should not be construed to imply that the mass media and other more impersonal societal inputs are unimportant considerations. Rather, the exclusion of these factors reflects a fairly simple cost-benefit analysis. As has been mentioned, the weight of the evidence strongly suggests that personal interaction is simultaneously the most effective and the most pervasive source of influence on many, if not most, political orientations, particularly partisanship. Further, compared to the vague and rather tenuous nature of societal media impact, the interpersonal lines of influence are relatively easy to identify and assess. Trying to incorporate additional causal elements beyond face-to-face contacts with significant others would not involve any new principles and the additional complexity and methodological problems engendered would be difficult if not impossible to deal with in any adequate way given the constraints of time, funding, and the present “state of the art.” Then, too, in indexing the attitudinal impact of significant others, we are indirectly taking at least some of the less personal milieu factors into account through their effect on those persons' viewpoints and their subsequent cue-giving to the respondents.

18 An additional complication is raised in considering the kind of cue-giving which takes place. There is a difference between a significant other saying “I'm a Democrat,” or in some other way being identified as such by himself or someone else, and asserting “I'm for the Democrats because—they're honest, brave, trustworthy, loyal, chaste, etc.” In the former situation the pairing is fairly direct and straightforward along the lines which we have been describing. In the latter case, the situation is more complex. Words will, of course, function as UCS in their own right and any conditioning process involving a significant other and partisan identification will be accelerated or perhaps contravened to some extent by the nature of the descriptive adjectives (that is, their own affective loading) used to refer to the political party; constituting, in effect, additional respondent learning trials every time they are employed. Still further complicating things is the fact that individuals can respondently and operantly condition themselves. See Bandura, Social Learning Theory.

19 Along with corroborative evidence from related areas of psychology (Hull, C., Principles of Behavior [New York: Appleton-Century, 1943]Google Scholar; Fishbein, M., “A Behavior Theory Approach to the Relations Between Beliefs about the Object and the Attitude Toward the Object,” Readings in Attitude Theory and Measurement [New York: John Wiley and Sons, 1967], 389–400Google Scholar; Newcomb, et al., Persistence and Change: Bennington College and Its Students after 25 Years, 54) and political behavior (McCloskey and Dahlgren, “Primary Group Influence on Party Loyalty”).

20 See Niemi, R. G., How Family Members Perceive Each Other (New Haven: Yale University Press, 1974), 60–61Google Scholar, for a supporting view on the relationship between partisan intensity and cue-giving. Some sort of combined index of the amount of political discussion a respondent had with significant others plus the latter's general degree of political activity would probably have been a better indicator. This information is not available in the data base.

21 Some persons feel that external observer perceptions of significant other attitudes and behaviour are subject to serious self-directed bias. The work of Niemi (How Family Members Perceive Each Other), Langton, K. P. (Political Socialization [New York: Oxford University Press, 1969])Google Scholar, and this author (“Self-Report and Observer Perception: Some Additional Considerations on the Validity Question,” Political Methodology, forthcoming) suggests that this is not an important problem with respect to partisan identification.

22 The data base was not a random probability sample, nor does it need to be. Making inferences from the respondent population to a wider universe concerning some substantive preference or trait was not the objective, rather, it was to test the applicability of basic principles among a diverse population, something the sample had in abundance.

23 Maccoby, E., Matthews, R., and Morton, A., “Youth and Political Change,” Public Opinion Quarterly 18 (1954), 23–29CrossRefGoogle Scholar; Langton, Political Socialization; Jennings and Niemi, The Political Character of Adolescence.

24 Since the cell sizes for affective evaluations < 0 were very small, producing wildly fluctuating (though overwhelmingly negative) correlation figures, they have been omitted from this and all succeeding tables.

25 It should be pointed out that this measure is tapping a learning history at some distance in time from the actual events. This can be expected to reduce the strength and regularity of the observed relationship since the magnitude and direction of the affective response elicited from the respondent by the significant other might well have been different when those learning trials were occurring than when assessed in this study.

26 The product-moment correlation between significant-other partisan intensity and the “how sure” scale was .51.

27 McCloskey and Dahlgren, “Primary Group Influence on Party Loyalty”.

28 Powell, G. B. Jr, “Political Cleavage Structure, Cross-Pressure Processes, and Partisanship: An Empirical Test of the Theory,” American Journal of Political Science 20 (1976), 1–23CrossRefGoogle Scholar.

29 There were 1,410 cases (about 56 per cent of the intact family triples) where both parents were liked a +5 or 6, and in 63 per cent of those, the affective ratings were exactly the same.

30 What percentage of these cases might also be response errors in filling out the questionnaire is unknown.

31 It should be born in mind that judging the efficacy of the model by the closeness of the means is not appropriate since the range of the predicted value is more or less arbitrary and varies between ±8.7 versus ±3.75 for the actual score.

32 For the entire sample there were 208 cases (7.5%) where the respondents claimed a partisanship contrary to that predicted by the model. In an additional 157 cases (5.7%) a partisan preference was predicted and the respondent indicated he had none or vice versa. Some slippage was allowed. If the respondent claimed no preference whatsoever (0), and the predicted intensity was less than or equal to .10 in either partisan direction (6 per cent of the mean, 8 per cent of the S.D., and 1 per cent of the range), it was not counted as an error. By trichotomized grade groupings the error percents were:

33 This interpretation is also consistent with the common finding that the longer individuals have occupied a new class position, the more they tend to resemble its dominant sentiments in their political views. With time, the number of conditioning trials they are exposed to from their more recently acquired associates will be more and more likely to offset the earlier pre-mobility learning, supplanting it with the new views.