American Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to The American Political Science Review. http://www.jstor.org The Transmission of Political Values from Parent to Child Author(s): M. Kent Jennings and Richard G. Niemi Source: The American Political Science Review, Vol. 62, No. 1 (Mar., 1968), pp. 169-184 Published by: American Political Science Association Stable URL: http://www.jstor.org/stable/1953332 Accessed: 21-09-2015 11:52 UTC REFERENCES Linked references are available on JSTOR for this article: http://www.jstor.org/stable/1953332?seq=1&cid=pdf-reference#references_tab_contents You may need to log in to JSTOR to access the linked references. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions THE TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD* M. KENT JENNINGS University of Michigan RICHARD G. NIEMI University of Rochester In understanding the political development of the pre-adult one of the central questions hinges on the relative and differentiated contributions of various socializing agents. The question undoubtedly proves more difficult as one traverses a range of polities from those where life and learning are almost completely wrapped up in the immediate and extended family to those which are highly complex social organisms and in which the socialization agents are extremely varied. To gain some purchase on the role of one socializing agent in our own complex society, this paper will take up the specific question of the transmission of certain values from parent to child as observed in late adolescence. After noting parent-child relationships for a variety of political values, attention will be turned to some aspects of family structure which conceivably affect the transmission flows. I. ASSESSING THE FAMILY' S IMPACT "Foremost among agencies of socialization into politics is the family." So begins Herbert Hyman's discussion of the sources of political learning.' Hyrnan explicitly recognized the importance of other agents, but he was neither the first nor the last observer to stress the preeminent position of the family. This viewpoint relies heavily on both the direct and indirect role of the family in shaping the basic orientations of offspring. Whether the child is conscious or unaware of the impact, whether the process is role-modelling or overt transmission, whether the values are political and directly usable or "nonpolitical" but transferable, and whether what is passed on lies in the cognitive or affective realm, it has been * Revised version of a paper delivered at the annual meeting of the American Political Science Association, New York, September, 1966. Financial support for the study reported here comes from The Danforth Foundation and the National Science Foundation. We wish to acknowledge the assistance of Michael Traugott in the preparation of this paper. l Herbert Hyman, Political Socialization (New York: Free Press of Glencoe, 1959), p. 69. argued that the family is of paramount importance. In part this view draws heavily from psychoanalytic theory, but it is also influenced by anthropological and national character studies, and by the great emphasis on role theory in sociological studies of socialization. In part the view stems also from findings in the area of partisan commitment and electoral behavior indicating high intergenerational agreement. Unfortunately for the general thesis, such marked correlations have been only occasionally observed in other domains of political life. Indeed, other domains of political life have been rarely explored systematically with respect to the central question of articulation in parent-child political values.2 Inferences, backward and forward extrapolations, and retrospective and projective data have carried the brunt of the argument. A recent major report about political socialization during the elementary years seriously questions the family's overriding importance. In contrast to the previously-held views that the family was perhaps preeminent or at least co-equal to other socializing agents stands the conclusion by Robert Hess and Judith Torney that "the public school is the most important and effective instrument of political socialization in the United States,"1 and that "the family transmits its own particular values in relatively few 'areas of political socialization and that, for the most part, the impact of the family is felt only as one of several socializing agents and institutions."' Hess and Torney see the primary influence of the family as the agent which promotes early attachment to country and government, and which thus "insures the stability of basic institutions."4 2 Most of these few studies, cited by Hyman, op. cit., pp. 70-71, are based on extremely limited samples and nearly all took place between 1930- 1950. 3 Robert D. Hess and Judith V. Torney, The Development of Basic Attitudes and Values Toward Government and Citizenship During the Elementary School Years, Part I (Cooperative Research Project No. 1078, U.S. Office of Education, 1965), pp. 193, 200. 4Ibid., p. 191. 169 This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 170 THE AMERICAN POLITICAL SCIENCE REVIEW Hence, "the family's primary effect is to support consensually-held attitudes rather than to inculcate idiosyncratic attitudes."' The major exception to these conclusions occurs in the area of partisanship and related matters where the family's impact is predictably high. The Hess and Torney argument thus represents a major departure from the more traditional view. They see the family's influence as age-specific and restricted in its scope. In effect, the restriction of the family's role removes its impact from much of the dynamic qualities of the political system and from individual differences in political behavior. The consensual qualities imparted or reinforced by the family, while vital for comprehending the maintenance of the system, are less useful in explaining adjustments in the system, the conflicts and accommodations made, the varied reactions to political stimuli, and the playing of diverse political roles. In short, if the family's influence is restricted to inculcating a few consensual attributes (plus partisan attachment), it means that much of the socialization which results in individual differentiation in everyday politics and which effects changes in the functioning of the political system lies outside the causal nexus of the parent-child relationship. The first and primary objective of the present article will be to assay the flow of certain political values from parent to child. Our attention will be directed toward examining the variation in the distributions of the offsprings' values as a function of the distribution of these same values among their parents. This is not to say that other attitudinal and behavioral attributes of the parents are unimportant in shaping the child's political orientations. For example, children may develop authoritarian or politically distrustful attitudes not because their parents are authoritarian or distrustful but because of other variables such as disciplinary and protection practices.6 Such transformations, while perhaps quite significant, will not be treated here. Rather, we will observe the degree to which the shape of value distributions in the child corresponds to that of his parent. Most of the values explored do not reflect the basic feelings of I Ibid., p. 192. 6 Illustrative of this argument is Frank A. Pinner's careful rendering in "Parental Overprotection and Political Distrust," The Annals, 361 (September, 1965), 58-70. See, in the same issue, Fred I. Greenstein, "Personality and Political Socialization: The Theories of Authoritarian and Democratic Character," pp. 81-95. attachment to the political system which supposedly originate in the early years,7 but much more of the secondary and tertiary values which tend to distinguish the political behavior of individuals and which contribute to the dynamics of the system. Study Design. The data to be employed come from a study conducted by the Survey Research Center of the University of Michigan in the spring of 1965. Interviews were held with a national probability sample of 1669 seniors distributed among 97 secondary schools, public and nonpublic.8 The response rate for students was 99 percent. For a random third of the students the father was designated for interviewing, for another random third the mother was designated, and for the other third both parents were assigned. In the permanent absence of the designated parent, the other parent or parent surrogate was interviewed. Interviews were actually completed with at least one parent of 94 percent of the students, and with both parents of 26 percent of the students, or 1992 parents altogether. Among parents the response rate was 93 percent.9 Two features of the student and parent samples should be underscored. First, since the sample of students was drawn from a universe of 12th graders, school drop-outs in that age cohort, estimated at around 26 percent for this time period, were automatically eliminated. Second, due mainly to the fact that more mothers than fathers constitute the head of household in single-parent families, the sample of parents is composed of 56 percent mothers.-" 7 In addition to the Hess and Torney report, evidence for this is supplied by, inter alios, Fred I. Greenstein, Children and Politics (New Haven: Yale University Press, 1965); and David Easton and Jack Dennis, "The Child's Image of Government," The Annals, 361 (September, 1965), 40- 57. 8 Of the original ninety-eight schools, drawn with a probability proportionate to their size, eighty-five (87%) agreed to participate; matched substitutes for the refusals resulted in a final total of ninety-seven out of 111 contacted altogether (87%). 9 Additional interviews were conducted with 317 of the students' most relevant social studies teachers and with the school principals. Some 21,000 paper-pencil questionnaires were administered to all members of the senior class in 85 percent of the sample schools. 10In any event, initial controls on parent (as well as student) sex suggest that parent-student agreement rates on the values examined here This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD 171 Our basic procedure will be to match the parent and student samples so that parentstudent pairs are formed. Although the actual number of students for whom we have at least one parent respondent is 1562, the base number of pairs used in the analysis is 1992. In order to make maximum usage of the interviews gathered, the paired cases in which both the mother and father were interviewed (430) are each given half of their full value.'1 A further adjustment in weighting, due to unavoidably imprecise estimates at the time the sampling frame was constructed, results in a weighted total of 1927 parent-student pairs.'2 Using 12th graders for exploring the parental transmission of political values carries some distinct characteristics. In the first place, most of these pre-adults are approaching the point at which they will leave the immediate family. Further political training from the parents will be minimal. A second feature is that the formal civic education efforts of society, as carried out in the elementary and secondary schools, are virtually completed. For whatever effect they may have on shaping the cognitive and cathectic maps of individuals, these various formal and informal modes of citizenship preparation will generally terminate, although other forms of educational preparation may lie ahead, especially for the college bound. A final consideration is that while the family and the educational system have come to some terminal point as socializing agents, the preadult has yet to be much affected by actual political practice. Neither have other potentially important experiences, such as the establishment of his own nuclear family and an occupational role, had an opportunity to exert differ little among parent-student sex combinations. This will be discussed in more detail below. 11The alternative to half-weighting these pairs is to subselect among those cases where both mother and father were interviewed. Half weighting tends to reduce the sampling variability because it utilizes more data cases. 12 It proved impossible to obtain accurate, recent figures on 12th grade enrollment throughout the country. Working with the data available and extrapolating as necessary, a sampling frame was constructed so that schools would be drawnwith a probability proportionate to the size of the senior class. After entry was obtained into the sample schools and precise figures on enrollments gathered, differential weights were applied to correct for the inequalities in selection probabilities occasioned by the original imprecise information. The average weight equals 1.2. their effects. Thus the 12th grader is at a significant juncture in his political life cycle and it will be instructive to see the symmetry of parental and student values at this juncture. AdolescentRebellion. It should be emphasized that we are not necessarily searching for patterns of political rebellion from parental values. Researchers have been hard-pressed to uncover any significant evidence of adolescent rebellion in the realm of political affairs.13 Pre-adults may differ politically from their parentsparticularly during the college years-but there is scant evidence that the rebellion pattern accounts for much of this deviance. Data from our own study lend little support to the rebellion hypotheses at the level of student recognition. For example, even of the 38 percent of the student sample reporting important disagreements with their parents less than 15 percent placed these disagreements in a broadlydefined arena of political and social phenomena. And these disagreements do not necessarily lie in the province of rebellion, as one ordinarily construes the term. There is, furthermore, some question as to whether adolescent rebellion as such occurs with anything approaching the frequency or magnitude encountered in sociological writings and the popular literature. As two scholars concluded after a major survey of the literature dealing with "normal" populations: In the large scale studies of normal populations, we do not find adolescents clamoring for freedom or for release from unjust constraint. We do not find rebellious resistance to authority as a dominant theme. For the most part, the evidence bespeaks a modal pattern considerably more peaceful than much theory and most social comment would lead us to expect. 'Rebellious youth' and 'the conflict between generations' are phrases that ring; but, so far as we can tell, it is not the 13 Hyman, op. cit., p. 72, and n. 6, p. 89. See also Robert E. Lane, "Fathers and Sons: Foundations of Political Belief," American Stciological Review, 24 (August, 1959), 502-511; Eleanor E. Maccoby, Richard E. Matthews, and Anton S. Morton, "Youth and Political Change," Public Opinion Quarterly, 18 (Spring, 1954), 23-39; Russell Middleton and Snell Putney, "Political Expression of Adolescent Rebellion," American Journal of Sociology, 68 (March, 1963), 527-535; and Robert H. Somers, "The Mainsprings of the Rebellion: A Survey of Berkeley Students in November, 1964," in Seymour Martin Lipset and Sheldon S. Wolin (eds.), The Berkeley Student Revolt (Garden City, N.Y.: Doubleday, 1965), p. 547. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 172 THE AMERICAN POLITICAL SCIENCE REVIEW ring of truth they carry so much as the beguiling but misleading tone of drama.'4 To say that rebellion directed toward the political orientations of the parents is relatively rare is not to say, however, that parent and student values are consonant. Discrepancies can occur for a variety of reasons, including the following: 1) Students may consciously opt for values, adopted from other agents, in conflict with those of their parents without falling into the rebellion syndrome. 2) Much more probable are discrepancies which are recognized neither by the parent nor the offspiing."5 The lack of cue-giving and object saliency on the part of parents sets up ambiguous or empty psychological spaces which may be filled by other agents in the student's environment. 3) Where values are unstable and have low centrality in a belief system, essentially random and time-specific responses to stimuli may result in apparent low transmission rates. 4) Another source of dissonant relationships, and potentially the most confounding one, is that life cycle effects are operative. When the pre-adult reaches the current age of his parents, his political behavior might well be similar to that of his parents even though his youthful attitudes would not suggest such congruency. This is an especially thorny empirical question and nests in the larger quandry concerning the later life effects of early socialization. II. PATTERNS OF PARENT-CHILD CORRESPONDENCES Confronted with a number of political values at hand we have struck for variety rather than any necessary hierarchy of importance. We hypothesized a range of correlations dependent in part on the play of factors assumed to alter the parent-student associations (noted above). We have purposely deleted values dealing with participative orientations and, as noted previously, those delving into sentiments of basic attachment and loyalty to the regime. The values selected include party identification, attitudinal posi- 14 Elizabeth Douvan and Martin Gold, "Modal Patterns in American Adolescence," in Lois and Martin Hoffman (eds.), Review of Child Development Research (New York: Russell Sage Foundation, 1966), Vol. II, p. 485. 15For an analysis of students' and parents' knowledge of each other's political attitudes and behavior, see Richard G. Niemi, "A Methodological Study of Political Socialization in the Family" (unpublished Ph.D. thesis, University of Michigan, 1967). tions on four specific issues, evaluations of socio-political groupings, and political cynicism. For comparative purposes we shall glance briefly at parent-student congruences in the religious sphere. To measure agreement between parents and students we rely primarily on correlations, either of the product-moment or rank-order variety. While the obvious alternative of percentage agreement may have an intuitive appeal, it has several drawbacks. Percentage agreement is not based on the total configuration of a square matrix but only on the "main diagonal." Thus two tables which are similar in percentage agreement may represent widely differing amounts of agreement if deviations from perfect agreement are considered. Moreover, percentage agreement depends heavily on the number of categories used, so that the degree of parent-student similarity might vary for totally artificial reasons. Correlations are more resistant to changes in the definition of categories. Finally, correlations are based on relative rankings (and intervals in the case of product-moment correlations) rather than on absolute agreement as percentage agreement usually is. That is, if student scores tend to be higher (or lower) than parent scores on a particular variable, but the students are ranked similarly to their parents, a high correlation may be obtained with very little perfect agreement. Party Identification. Previous research has established party identification as a value dimension of considerable importance in the study of political behavior as well as a political value readily transmitted from parents to children. Studies of parent-youth samples as well as adult populations indicate that throughout the life cycle there is a relatively high degree of correspondence between respondents' party loyalities and their parents'. Our findings are generally consistent with those of these earlier studies. The substantial agreement between parent and student party affiliations is indicated by a tau-b (also called tau-beta) correlation of .47, a statistic nearly unaffected by the use of three, five, or all seven categories of the party identification spectrum generated by the question sequence." The magnitude of this 18 This figureis based on parent-student pairs in which both respondents have a party identification; eliminated are the 2 percent of the pairs in which one or both respondents are apolitical or undecided. The product-moment correlation for these data is .59. The standard SRC party identification questions were used: see Angus CampThis content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD 173 statistic reflects the twin facts of the presence of a large amount of exact agreement and the absence of many wide differences between students and parents. When the full 7 X 7 matrix of parent-student party loyalties is arrayed (Table 1), the cells in which parents and students are in unison account for a third of the cases. The cells representing maximum disagreement are very nearly empty. Despite our earlier contention, collapsing categories and considering percentage agreement in the resulting table does make good substantive sense with regard to party identification. In this instance the collapsed categories have a meaning beyond just broader segments of a continuum, and are associated with a general orientation toward one party or the other or bell, Philip E. Converse, Warren E. Miller, and Donald E. Stokes, The American Voter (New York: Wiley, 1960), Ch. 6. toward a neutral position between them. Thus arrayed, 59 percent of the students fall into the same broad category as their parents, and only seven percent cross the sharp divide between Republicans and Democrats. The observed similarity between parents and students suggests that transmission of party preferences from one generation to the next is carried out rather successfully in the American context. However, there are also indications that other factors (temporarily at least) have weakened the party affiliations of the younger generation. This is most obvious if we compare the marginal totals for parents and students (Table 1). The student sample contains 'almost 12 percent more Independents than the parent sample, drawing almost equally on the Republican and Democratic proportions of the sample. Similarly, among party identifiers a somewhat larger segment of the students is but weakly inclined toward the chosen party. TABLE 1. STUDENT-PARENT PARTY IDENTIFICATION Students Parents Strong Weak Ind. md. Ind. Weak Strong Total Dem. Dem. Dem. d Rep. Rep. Rep. Party Identification Strong Dem. 9.7% 8.0 3.4 1.8 .5 .9 .5 24.7% Weak Dem. 5.8 9.0 4.2 2.6 .7 1.6 .7 24.7 (32.6)a (13.2) (3.6) (49.4) Ind. Dem. 1.6 2.1 2.1 1.7 .8 .7 .2 9.3 Ind. 1.1 1.6 1.6 2.7 1.2 .9 .5 9.7 Ind. Rep. .1 .5 .8 .9 .9 1.3 .5 4.9 (7.0) (12.7) (4.1) (23.9) Weak Rep. .3 2.1 1.6 2.3 1.9 5.0 1.9 15.0 Strong Rep. .2 .9 .8 .8 2.4 3.3 3.5 11.7 (3.4) (9.7) (13.6) (26.7) Total 18.8% 24.2 14.5 12.8 8.4 13.6 7.7 100.0% (43.0) (35.7) (21.3) tau-b= .47 N = 1852 a The full 7X7 table is provided because of the considerable interest in party identification. For some purposes, reading ease among them, the 3 X3 table is useful. It is given by the figures in parentheses; these figures are (within rounding error) the sum of the numbers just above them. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 174 THE AMERICAN POLITICAL SCIENCE REVIEW Nor are these configurations simply an artifact of the restricted nature of the parent sample, since the distribution of party identification among the parents resembles closely that of the entire adult electorate as observed in November, 1964 (SRC 1964 election study). A number of factors might account for the lesser partisanship of the students, and we have only begun to explore some of them. On the one hand, the students simply lack their parents' long experience in the active electorate, and as a consequence have failed as yet to develop a similar depth of feeling about the parties.17 On the other hand, there are no doubt specific forces pushing students toward Independence. The experience of an ever-widening environment and the gradual withdrawal of parental power may encourage some students to adopt an Independent outlook. The efforts of schools and of teachers in particular are probably weighted in the same direction. If these forces are at work, high school students may be gradually withdrawing from an earlier position of more overt partisanship. But, whatever the exact nature of the causes, they clearly draw off from the partisan camp a small but significant portion of the population as it approaches full citizenship. Opinions on Specific Issues. One way in which political values are expressed is through opinions on specific issues. However, as Converse has shown, many opinions or idea elements not only tend to be bounded by systems of low constraint but are also quite unstable over relatively short periods of time among mass publics.'8 Hence in comparing student responses with parent responses the problem of measurement may be compounded by attitude instability among both samples. Rather than 17 This is suggested by an analysis of different age groups among the active electorate: see Ibid., pp. 161ff. For evidence that the depth of adult attachment to party is not necessarily uniform across electoral systems see M. Kent Jennings and Richard G. Niemi, "Party Identification at Multiple Levels of Government," American Journal of Sociology, 72 (July, 1966), 86-101. 18 Philip E. Converse, "The Nature of Belief Systems in Mass Publics," in David E. Apter (ed.), Ideology and Discontent (New York: Free Press of Glencoe, 1964), pp. 206-261. The following section borrows from Converse's discussion. Robert E. Agger takes a somewhat different view of instabilities in "Panel Studies of Comparative Community Political Decision-Making," in M. Kent Jennings and L. Harmon Zeigler (eds.), The Electoral Process (Englewood Cliffs: Prentice-Hall 1966), pp. 265-289. being a handicap instabilities actually sharpen the test of whether significant parent-tochild flows occur. One would not expect unstable sentiments to be the object of any considerable political learning in the family. It seems unlikely that many cues would be given off over matters about which the parents were unsure or held a fluctuating opinion. Even in the event of numerous cues in unstable situations, the ambivalent or ambiguous nature of such cues would presumably yield instability in the child. In either case the articulation between parent and child beliefs would be tempered. We have selected four specific issues for examination. Two involve public schools; given the populations being studied, schools are particularly relevant attitude objects. Furthermore these two issues envelope topics of dramatic interest to much of the public-integration in the schools and the use of prayers in schools. After an initial screening question weeded out those without any interest at all on the issues, the respondents were asked if they thought the government in Washington should "see to it that white and Negro children go to the same schools" or if the government should "stay out of this area as it is none of its business." On the prayers in school question the respondents were asked if they believed "schools should be allowed to start each day with a prayer" or that "religion does not belong in the schools.""9 Taken in the aggregate the high school seniors proved less likely to sanction prayers in school than did the parents (although a majority of both answered in the affirmative) and more willing to see the federal government enforce segregation than were the adults (with both yielding majorities in favor). These differences are moderate; no more than 14 percentage points separate like-paired marginals on the prayer issue and no more than 10 points on the integration issue. The crosstabulation of parent and student responses produces moderately strong coefficients, as shown in the first two entries of Table 2. Combining as they do some very visible population groupings along with topics of more than usual prominence in the mass 19 Sizeable proportions of both parents and students elected to state a middle or "depends" response, particularly on the first question, Such responses occupy a middle position in our calculation of the rank order correlations. On the first issue 10 percent of the pairs were dropped because either the parent or child opted out on the initial screen; the corresponding figure for the second issue is 19 percent. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD 175 TABLE 2. CORRELATIONS BETWEEN PARENTSTUDENT ATTITUDES ON FOUR ISSUES Federal government's role in integrating the schools .34a Whether schools should be allowed to use prayers .29 Legally elected Communist should be allowed to take office .13 Speakers against churches and religion should be allowed .05 a Each of the correlations (tau-b) in this table is based on at least 1560 cases. media and local communities, it would be surprising if there were not at least a moderate amount of parent-student overlap. The wonder is not that the correlations are this high, but rather that they are not higher. If correlations no higher than this are produced on issues which touch both generations in a manner which many issues assuredly do not, then one would speculate that more remote and abstract issues would generate even less powerful associations. This hypothesizing is borne out by the introduction of two other issues. Both parents and students were asked to agree or disagree with these two statements: "If a Communist were legally elected to some public office around here, the people should allow him to take office"; and "If a person wanted to make a speech in this community against churches and religion, he should be allowed to speak." In general, the pre-adults took a slightly more libertarian stance on the two issues than did the parents, but the differences in any of the like-paired marginals do not exceed 14 percent. These similarities mask extremely tenuous positive correlations, however, as the second pair of items in Table 2 reveals. These two issues carry neither the immediacy nor the concreteness which may be said to characterize the two issues dealing with integration and prayers in the schools. Indeed, one might question whether the two statements represent issues at all, as the public normally conceives of issues. At any rate it is improbable that the students are reflecting much in the way of cues emitted from their parents, simply because these topics or related ones are hardly prime candidates for dinnertable conversation or inadvertent cue-giving. Nor do they tap some rather basic sentiments and attitude objects which permeate the integration and prayers issues. Such sentiments are more likely to be embedded in the expressive value structure of the parents than are those having to do with some of the more abstract "fundamental" tenets of democracy as exemplified in the free speech and right to take office issues. That adults themselves have low levels of constraint involving propositions about such fundamental tenets has been demonstrated by McClosky, and Prothro and Grigg.20 Given this environment, the lower correlation for the two more abstract propositions is predictable. Although the issues we have examined by no means exhaust the variety of policy questions one might pose, they probably exemplify the range of parent-student correspondences to be found in the populace. On all but consensual topics-which would perforce assume similar distributions among virtually all population strata anyway-the parent-student correlations obtained for the integration and prayer issues probably approach the apex. In part this may be due to unstable opinions and in part to the effects of agents other than the family. It is also possible that the children will exhibit greater correspondences to their parents later in the life cycle. But for this particular point in time, the articulation of political opinions is only moderately strong on salient, concrete issues and virtually nil on more abstract issues. Evaluattionsof Socio-Political Groupings. Collectivities of people which are distinguished by certain physical, locational, social, religious, and membership characteristics (the list is obviously not exhaustive) often come to serve as significant political reference groups for individuals. While distinguishable groups may carry affective neutrality, it seems to be in the nature of mass behavior that these groups most often come to be viewed with greater or lesser esteem. The intersection of group evaluations and the political process comes when claims or demands are made by or upon significant portions of such groupings. The civil rights movement of the past decade is perhaps the most striking contemporary example. As Converse has suggested, social groupings are likely to have greater centrality for mass publics than abstract idea elements per se.2' Thus when particular issues and public policies become imbued with group-related properties, 20 Herbert McClosky, "Consensus and Ideology in American Politics," this REVIEw, 58 (June, 1964), 361-382; and James WV.Prothro and Charles W. Grigg, "Fundamental Principles of Democracy: Bases of Agreement and Disagreement," Journal of Politics, 22 (May, 1960), 276- 294. 81 Converse, op. cit. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 176 THE AMERICAN POLITICAL SCIENCE REVIEW the issues acquire considerably more structure and concreteness for the mass public than would be the normal case. To what extent is the family crucial in shaping the evaluations of social groupings and thus-at a further remove-the interpretation of questions of public policy? Some insight into this may be gained by comparing the ratings applied by the parents and students to eight socio-political groupings. While the groups all carry rather easily recognized labels, they do differ in terms of their relative visibility and their inclusive-exclusive properties. They include Protestants, Catholics, Jews, Negroes, Whites, Labor Unions, Big Business, and Southerners. To measure the attitudes toward these groups, an instrument dubbed the "feeling thermometer" was used. The technique was designed to register respondents' feelings toward a group on a scale ranging from a cold 0 to a warm 100. In the analysis we will treat this scale as interval level measurement. We have also examined the data using contingency tables and ordinal statistics; our conclusions are the same regardless of the method used. Turning first to the mean ratings, given in Table 3, we find a striking similarity in student and parent aggregate scores. The largest difference is five points and the average difference is only 2.2 points. Additionally, the standard TABLE 3. CORRELATIONS BETWEEN PARENTSTUDENT GROUP EVALUATIONS Group Parent-Student Mean Ratings Evaluated Correlations Parent Student Catholics .36a 72 70 Southerners .30 66 62 Labor Unions .28 60 60 Negroes .26 67 69 Jews .22 67 63 Whites .19 84 83 Protestants .15 84 79 Big Business .12 64 63 a Each of the product-moment correlations in this table is based on at least 1880 cases. The corresponding tau-b's are (top to bottom) .28, .22, .22, .20, .18, .19, .13, .08. deviations for the two samples (not shown) are extremely similar across all groupings. Nor were there significant tendencies for one sample to employ more than the other the option of "unawareness") or "no feelings" (a reading of 50 on the thermometer) about the groupings. Moreover, the aggregate differences which do occur are not immediately explicable. For example, students rate Southerners slightly lower than parents, as we expected, but the difference in ratings of Negroes is negligible, which was unanticipated. Students rate Whites and Protestants somewhat lower than parents. This is not matched, however, by higher evaluations of the minority groups-Jews, for example. Given these extraordinarily congruent patterns it is rather startling to see that they are patently not due to uniform scores of parentchild pairs. As shown in Table 3, the highest correlation between the parent and student ratings is .36 and the coefficients range as low as .12. Even the highest correlation is well below that found for party identification (where the product-moment coefficient was .59 for the seven-fold classification), and for several groupings the relationships between parent and student scores are very feeble. If the child's view of socio-political groupings grows out of cue-giving in the home, the magnitude of the associations should exceed those observed here. It is beyond the task of this paper to unravel thoroughly these findings. The range of correlations does provide a clue as to the conditions under which parent-student correspondences will be heightened. In the first place the three categories producing the lowest correlations appear to have little socio-political relevancy in the group sense. Whites and Protestants are extremely inclusive categories and, among large sectors of the public, may simply not be cognized or treated in everyday life as groupings highly differentiated from society in general. They are, in a sense, too enveloping to be taken as differentiated attitude objects. If they do not serve as significant attitude objects, the likelihood of parent to child transmission would be dampened. In the third case-Big Business-it seems likely that its visibility is too low to be cognized as a group qua group. As the parent-student correlations increase we notice that the groupings come to have not only highly distinguishable properties but that they also have high visibility in contemporary American society. Adding to the socio-political saliency thereby induced is the fact that group membership may act to increase the parentstudent correlations. One would hypothesize that parent-student pairs falling into a distinguishable, visible grouping would exhibit higher correlations in rating that same grouping than would nonmembers. Taking the four groupings for whom the highest correlations This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD 177 were obtained, we divided the pairs into those where both the parent and the child-except in the case of labor unions-were enveloped by the groupings versus those outside the groupings. Although none of the hypothesized relationships was contravened, only the coefficients for evaluations of Southerners provided a distinct demarcation between members and nonmembers (tau-b =.25 for Southern pairs, .14 for non-Southerners). It is quite possible that measures capturing membership identification and intensities would improve upon these relationships. As with opinions on specific issues, intrapair correlations on group evaluations are at best moderately positive, and they vary appreciably as a result of socio-political visibility and, to a small degree, group membership characteristics. What we begin to discern, then, is a pattern of congruences which peak only over relatively concrete, salient values susceptible to repeated reinforcement in the family (and elsewhere, perhaps), as in party identification and in certain issues and group evaluations. It is conceivable that these results will not prevail if we advance from fairly narrow measures like the ones previously employed to more global value structures. We now turn to an illustrative example. It so happens that it also provides an instance of marked aggregate differences between the two genera- tions. Political Cynicism. Political cynicism and its mirror image, trust, offer an interesting contrast to other variables we are considering. Rather than referring to specific political issues or actors, cynicism is a basic orientation toward political actors and activity. Found empirically to be negatively related to political participation, political cynicism has also been found to be positively correlated with measures of a generally distrustful outlook (personal cynicism).22 Political cynicism appears to be a manifestation of a deep-seated suspicion of others' motives and actions. Thus this attitude comes closer than the rest of our values to tapping a basic psycho-political predisposition. Previous research with young children suggests that sweeping judgments, such as the essential goodness of human nature, are formed early in life, often before cognitive development and information acquisition make the 22 Robert E. Agger, Marshall N. Goldstein, and Stanley A. Pearl, "Political Cynicism: Measurement and Meaning," Journal of Politics, 23 (August, 1961), p. 490; and Edgar Litt, "Political Cynicism and Political Futility," Journal of Politics, 25 (May, 1963), 312-323. evaluated objects intelligible. Greenstein, and Hess and Easton, have reported this phenomenon with regard to feelings about authority figures; Hess and Torney suggest similar conclusions about loyalty and attachment to government and country.23 Evaluative judgments and affective ties have been found among the youngest samples for which question and answer techniques are feasible. This leads to the conclusions that the school, mass media, and peer groups have had little time to influence these attitudes. It seems to follow that the family is the repository from which these feelings are initially drawn. Either directly by their words and deeds or indirectly through unconscious means, parents transmit to their children basic postures toward life which the children carry with them at least until the development of their own critical faculties. Although our 12th graders have been exposed to a number of influences which could mitigate the initial implanting, one should expect, according to the model, a rather strong correspondence between parent and student degrees of political cynicism. To assess the cynicism of parents and students, a Guttman scale was constructed from five questions asked of both samples. All questions dealt with the conduct of the national government.24 In each sample the items formed 23 Greenstein, op. cit., Ch. 3; Robert D. Hess and David Easton, "The Child's Changing Image of the President," Public Opinion Quarterly, 24 (Winter, 1960), 632-644; and Hess and Torney, op. cit., pp. 73ff. 24 The items are as follows: 1) Do you think that quite a few of the people running the government are a little crooked, not very many are, or do you think hardly any of them are? 2) Do you think that people in the government waste a lot of the money we pay in taxes, waste some of it, or don't waste very much of it? 3) How much of the time do you think you can trust the government in Washington to do what is right-just about always, most of the time, or only some of the time? 4) Do you feel that almost all of the people running the government are smart people who usually know what they are doing, or do you think that quite a few of them don't seem to know what they are doing? 5) Would you say that the government is pretty much run by a few big interests looking out for themselves or that it is run for the benefit of all the people? This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 178 THE AMERICAN POLITICAL SCIENCE REVIEW a scale, with coefficients of reproducibility of .93 and .92 for parents and students, respectively. The aggregate scores reflect a remarkably lesser amount of cynicism among students than among parents. This is apparent in the marginal distributions in Table 4, which show the weight of the parent distribution falling much much more on the cynical end of the scale. Similarly, while a fifth of the students were more cynical than their parents, three times this number of parents were more cynical than their children. The students may be retreating from an even more trusting attitude held earlier, but compared to their parents they still see little to be cynical about in national political activity. Here is a case where the impact of other socialization agents-notably the schoollooms large. The thrust of school experience is undoubtedly on the side of developing trust in the political system in general. Civic training in school abounds in rituals of system support in the formal curriculum. These rituals and curricula are not matched by a critical examination of the nation's shortcomings or the possible virtues of other political forms. Coupled with a moralistic, legalistic, prescriptive orientation to the study of government is the avoidance of conflict dimensions and controversial issues.25A direct encounter with the realities of political life is thus averted or at least postponed. It would not be surprising, 25 These are old charges but apparently still true. After a survey of the literature on the subject and on the basis of a subjective analysis of leading government textbooks in high schools, Byron G. Massialas reaches similar conclusions: see his "American Government: 'We are the Greatest'," in C. Benjamin Cox and Byron G. then, to find a rather sharp rise in the level of cynicism as high school seniors move ahead in a few years into the adult world. Students on the whole are less cynical than parents; relative to other students, though, those with distrustful, hostile parents should themselves be more suspicious of the government, while those with trusting parents should find less ground for cynicism. Against the backdrop of our discussion, it is remarkable how low the correspondence is among parentstudent pairs. Aside from faint markings at the extremities, students' scores are very nearly independent of their parents' attitudes (Table 4). The cynicism of distrustful parents is infrequently implanted in their children, while a smaller group of students develops a cynical outlook despite their parents' views. Political cynicism as measured here is not a value often passed from parent to child. Regardless of parental feelings, children develop a moderately to highly positive view of the trustworthiness of the national government and its officials. These findings do not mean that parents fail to express negative evaluations in family interaction nor that children fail to adopt some of the less favorable attitudes of their parents. What is apparently not transmitted is a generalized cynicism about politics. Thus while warmth or hostility toward specific political objects with high visibility may be motivated by parental attitudes, a more pervasive type of belief system labelled cynicism is apparently subject to heavy, undercutting influences outMassialas (eds.), Social Studies in the United States: A Critical Appraisal (New York: Harcourt, Brace, & World, Inc., 1967), pp. 167-195. TABLE 4. RELATIONSHIP BETWEEN PARENT-STUDENT SCORES ON THE CYNICISM SCALE Students Row Marginal Least Cynical Most Cynical Rowa Marginalse Parents 1 2 3 4 5 6 Least Cynical-i 25% 27 33 13 1 2 101% 8% 2 19 28 38 9 1 5 100 12 3 18 28 37 10 3 4 100 33 4 16 23 41 13 3 4 100 17 5 15 19 35 19 3 9 100 9 Most Cynical-6 12 22 36 18 4 8 100 21 Marginal Totalsa 17% 25 37 13 3 5 100% tau-b= .12 N = 1869 a Marginal totals show the aggregate scaler patterns for each sample. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD 179 side the family nexus. These influences are still operative as the adolescent approaches adult status. Working with another encompassing set of values we encountered much the same patterns as with cynicism. After obtaining their rank orderings of interest in international, national, state, and local political matters the respondents were allocated along a 7-point scale of cosmopolitanism-localism through an adaptation of Coombs' unfolding technique.26 On the whole the students are considerably more cosmopolitan than the parents, and the paired correlation is a modest .17. Both life cycle and generational effects are undoubtedly at work here,27 but the central point is that the students' orientations only mildly echo those of their parents. What results from juxtaposing parents and their children on these two measures of cynicism and cosmopolitanism-localism is the suspicion that more global orientations to political life do not yield parent-student correspondences of greater magnitude than on more specific matters. If anything, the opposite is true-at least with respect to certain specifics. It may be that the child acquires a minimal set of basic commitments to the system and a way of handling authority situations as a result of early experiences in the family circle. But it appears also that this is a foundation from which arise widely diverse value structures, and that parental values are an extremely variable and often feeble guide as to what the pre-adult's values will be. Religious Beliefs. Up to this point we have traversed a range of political and quasipolitical values, and have witnessed varying, but generally modest degrees of parent-student correspondences. To what extent does this pattern also characterize other domains of social values? For comparative purposes we can inject a consideration of religious beliefs. Like party preference, church affiliation among pre-adults is believed to be largely the same as parental affiliation. Such proves to be the case among our respondents. Of all parent-student pairs 74 percent expressed the same denomina- 26 A description of this operation and some results are given in M. Kent Jennings, "Pre-Adult Orientations to Multiple Systems of Government," Midwest Journal of Political Science, XI (August, 1967), 291-317. The underlying theory and technique are found in Clyde Coombs, A Theory of Data (New York: Wiley, 1964), esp. Ch. 5. 27 This is discussed in more detail in Jennings, op. cit. tional preference. That this percentage is higher than the agreement on the three-fold classification of party identification (Democrat, Republican, Independent) by some 15 percent suggests that by the time the preadult is preparing to leave the family circle he has internalized the church preference of his parents to a moderately greater extent than their party preference. There are some perfectly valid reasons for this margin. To a much greater extent than party preference, church preference is likely to be reinforced in a number of ways. Assuming attendance, the child will usually go to the same church throughout childhood; the behavior is repeated at frequent intervals; it is a practice engaged in by greater or lesser portions of the entire family and thus carries multiple role-models; formal membership is often involved; conflicting claims from other sources in the environment for a change of preference are minimal except, perhaps, as a result of dating patterns. Religious affiliation is also often imbued with a fervid commitment. In contrast, party preference is something which the child himself cannot transform into behavior except in rather superficial ways; reinforcement tends to be episodic and varies according to the election calendar; while the party preference of parents may vary only marginally over the pre-adult years, voting behavior fluctuates more and thus sets up ambiguous signals for the child; other sources in the environment-most noticeably the mass media-may make direct and indirect appeals for the child's loyalty which conflict with the parental attachments. Given the factors facilitating intrafamilial similarities in church preference, and the absence of at least some of these factors in the party dimension, it is perhaps remarkable that congruity of party identification approaches the zone of churchpreference congruity. We found that when we skipped from party identification to other sorts of political values the parent-student correlations decreased perceptibly. May we expect to encounter similar behavior in the realm of religious values? One piece of evidence indicates that this is indeed the case. Respondents were confronted with a series of four statements having to do with the literal and divine nature of the Bible, ranging from a description of the Bible as "God's word and all it says is true" to a statement denying the contemporary utility of the book. Both students and parents tended to view the Bible with awe, the parents slightly more so than the students. But the correlation (tau-beta) among parent-student pairs is only This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 180 THE AMERICAN POLITICAL SCIENCE REVIEW a moderately strong .30. As with political values, once the subject matter moves out from central basic identification patterns the transmission of parental values fades.28 And, as with political values, this may be a function of instability-although this seems less likely for the rendering of the Bible-the impingement of other agents-particularly likely in this case-or the relative absence of cue-giving on the part of the parents. The more generalizable proposition emerging from a comparison of political and religious orientations is that the correlations obtained diminish when the less concrete value orientations are studied. III. FAMILY CHARACTERISTICS AND TRANSMISSION PATTERNS We have found that the transmission of political values from parent to child varies remarkably according to the nature of the value. Although the central tendencies lie on the low side, we may encounter systematic variations in the degree to which values are successfully transmitted according to certain properties of family structure. That is, whether the transmittal be conscious and deliberate or unpurposive and indirect, are there some characteristics of the family unit which abet or inhibit the child's acquisition of parental values? We shall restrict ourselves to a limited set of variables having theoretical interest. In order to dissect the parent-student relationships by controlling for a variety of independent variables, we shall retain the full parent-student matrices and then observe correlations within categories of the control variables.28The political values to be examined include party identification, political cynicism, political cosmopolitanism, four specific politi- 28 To compare directly the amount of correspondence on interpretation of the Bible with church membership information, which is nominal-level data, we used the contingency coefficient. Grouping parent and student church affiliations into nine general categories, the coefficient is .88, compared to .34 for the Bible ques- tion. 29 A more parsimonious method is to develop agreement indexes and to relate the control variables to these indexes. This results in a single statistic and contingency table for each control variable rather than one for each category of the control variable. Experience with both methods indicates that similar conclusions emerge, but retaining the full matrices preserves somewhat better the effects of each category of the control variable. cal issues, and the ratings assigned to three minority population groupings-Catholics, Negroes, and Jews. This makes ten variables altogether, but for some purposes the issues and the group ratings are combined into composite figures. Parent and Student Sex Combinations. Various studies of adolescents have illustrated the discriminations which controls for sex of parent and sex of child may produce in studying the family unit.30 Typically these studies have dealt with self-development, adjustment problems, motivational patterns, and the like. The question remains whether these discriminations are also found in the transmission of political values. Part of the common lore of American political behavior is that the male is more dominant in political matters than the female, in his role both of husband and of father. And among preadults, males are usually found to be more politicized than females. While our findings do not necessarily challenge these statements, they do indicate the meager utility of sex roles in explaining parent-student agreement. The correlations between parent-student values show some variation among the four combinations of parent and student sex, but the differences are usually small and inconsistent across the several values. Of the sixty possible comparisons for the ten political variables (i.e., (2) =6 pairs of correlations for each variable), only eight produce differences in the correlations greater than .10, and thirty-three fall within a difference of less than .05. The average parent-student correlations for these variables are: Mother-Son, .22; Mother-Daughter, .24; Father-Son, .20; Father-Daughter, .22. Thus the values of the father are not more likely to be internalized than those of the mother; nor do sons register consistently different rates of agreement than daughters. Finally, the particular sex mix of parent and child makes little difference. We also found that the use of sex combinations as controls on other bivariate s0 See, e.g., Charles E. Bowerman and Glen H. Elder, "Adolescent Perception of Family Power Structure," American Sociological Review, 29 (August, 1964), 551-567; E. C. Devereux, Urie Bronfenbrenner, and G. J. Suci, "Patterns of Parent Behavior in the United States of America and the Federal Republic of Germany: A CrossNational Comparison," International Social Science Journal, 14 (UNESCO, 1963), 1-20; and Morris Rosenberg, Societyand theAdolescentSelfImage (Princeton: Princeton University Press, 1965), Ch. 3. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD 181 relationships usually resulted in minor and fluctuating differences. Whatever family characteristics affect differential rates of value transmission, they are only marginally represented by sex roles in the family. Affectivity and Control Relationship. Another set of family characteristics employed with considerable success in studies of the family and child development has to do with the dimension of power or control on the one hand, and the dimension of attachment or affectivity on the other.' One salient conclusion has been that children are more apt to use their parents as role models where the authority structure is neither extremely permissive nor extremely autocratic and where strong (but not overprotective) supportive functions and positive affects are present. Although these dimensions have been employed in various ways in assessing the socialization of the child, they have rarely been utilized in looking at value transmission per se. In the nearest approach to this in political socialization studies, college students' reports suggested that; perceived ideological differences between parent and child were higher when there was emotional estrangement, when the parental discipline was perceived as either too high or too low, and when the parent was believed to be interested in politics.32 Somewhat related findings support the idea that affective and power relationships between parent and child may affect the transferral of political orientations.33 Affectivity and control relationships between pre-adults and their parents were operationalized in a number of ways, too numerous to give in detail. Suffice it to say that both parent and offspring were queried as to how close they felt to each other, whether and over what they 31 A discussion of these dimensions is found in Murray Straus, "Power and Support Structure of the Family in Relation to Socialization," Journal of Marriage and the Family, 26 (August, 1964), 318-326. See also Wesley C. Becker, "Consequences of Different Kinds of Parental Discipline," in Martin and Lois Hoffman (eds.), Review of Child Development (New York: Russell Sage Foundation, 1964), Vol. 1, pp. 169-208; William H. Sewell, "Some Recent Developments in Socialization Theory and Research," The Annals, 349 (September, 1963), 163-181; Glen H. Elder, Jr., "Parental IPower Legitimation and Its Effects on the Adolescent," Sociometry, 26 (March, 1963), 50-65; and Douvan and Gold, op. cit. 32 Middleton and Putney, op. cit. 3 Lane, op. cit.; and Maccoby et al., op. cit. disagreed, the path of compatibilities over the past few years, punishment agents, perceived level of parental control, parent and student satisfaction with controls, the nature and frequency of grievance processing, and rule-making procedures. In accordance with the drift of previous research we hypothesized that the closer the student felt to his parent the more susceptible he would be to adopting, either through formal or informal learning, the political values of the parent. This turned out to be untrue. The closeness of parents and children, taking either the parent's report or the child's report, accounts for little variation in the parent-student correlations. This is true whether closeness to mother or father is considered and regardless of the student's sex. Similarly, other measures of affective relationships give little evidence that this dimension prompted much variation in the correlations among pairs. Turning to the power relationships between parent and child we hypothesized two types of relationships: 1) the more "democratic" and permissive these relationships were the greater congruency there would be; and 2) the more satisfied the child was with the power relationships the greater would be the congruency. Where patterning appears it tends to support the first hypothesis. For example, those students avowing they have an "average" amount of autonomy agree slightly more often with their parents than do those left primarily to their own resources and those heavily monitored by their parents. More generally, however, the power configuration-either in terms of its structure or its appraised satisfactoriness -generated few significant and consistent differences. This proved true whether we relied on the parent's account or the student's. As with sex roles, the affective and control dimensions possess weak explanatory power when laid against parent-to-student transmission patterns. In neither case does this mean that these characteristics are unimportant for the political socialization of the young. It does mean that they are of little help in trying to account for the differential patterns of parent-student congruences. Levels of Politicization. Another set of family characteristics concerns the saliency and cue-giving structure of political matters within the family. One would expect parents for whom politics is more salient to emit more cues, both direct and indirect. Other things being equal, the transmission of political values would vary with the saliency and overt manifestations of political matters. Cue-giving This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 182 THE AMERICAN POLITICAL SCIENCE REVIEW TABLE 6. FAMILY POLITICIZATION AND PARENT-STUDENT AGREEMENT ON A RANGE OF POLITICAL VALUESa .. Cosmopoli- Prayer and Party Political tnsm - Group . Freedom Frequencyof: . < . . Localism Rains Integration Identification Cynicism sRatngSb Issues' Localism Issuese Husband-Wife Political Conversations Very often .54 .19 .22 .20 .36 .13 Pretty often .49 .15 .11 .20 .30 .10 Not very often .45 .11 .14 .24 .28 .06 Don't talk .32 .08 .22 .23. .32 .08 Student-Parent Political Conversations Several times/week .49 .16 .17 .22 .32 .08 Few times/month .45 .12 .16 .21 .35 .14 Few times/year .41 .10 .18 .30 .18 -.05 Don't talk .47 .02 .12 .20 .26 .06 a Each tau-b correlation in this table is based on at least 82 cases. b Average ratings of Negroes, Catholics, and Jews on the "feeling thermometer." e See p. 175 for a description of these issues. would structure the political orientations of the child and, in the absence of rebellion, bolster parent-student correspondences. The absence of cue-giving would probably inject considerable instability and ambiguity in the child's value structure. At the same time this absence would invite the injection of other socializing agents whose content and direction might vary with parental values. In either event parental-offspring value correspondences should be reduced in the case of lower political saliency and cue-giving. Turning to the data, it is evident that while the hypothesis receives some support for party identification and political cynicism, it does not hold generally. Illustratively, Table 6 provides the parent-student correlations for party indentification, cynicism, cosmopolitanism, averaged group evaluations, and two pairs of issues. The two politicization measures capture different elements of family politicization-the extent of husband-wife conversations about politics (reported by parents) and the frequency of student-parent conversations related to political affairs (reported by students). The correspondence between parent and student cynicism is mildly related to both of these measures, while party identification is clearly affected by parental conversations, but not by student-parent political discussions. The other opinions and values show no consistent relationships with either measure of politicization. Similar results were obtained when politicization was measured by the general political interest among parents and students, parentstudent disagreements regarding political and social matters, and parents' participation in political campaigns. That the level of family politicization affects somewhat the flow of party identification and cynicism but is unrelated to the transmission of other variables should not be ignored. The extremely salient character of party loyalties, which results in the higher overall parentstudent correlation, and the summary nature of the cynicism variable suggest characteristics that may determine the relevancy of family politicization for the transmission of political values. The essential point, though, is that the level of politicization does not uniformly affect the degree of parent-student correspondence. Students with highly politicized backgrounds do not necessarily resemble their parents more closely than students from unpoliticized families, Whether it is measured in terms of student or parent responses, taps spectator fascination with or active engagement in politics, or denotes individual-level or family-level properties, varying amounts of politicization do not uniformly or heavily alter the level of correThis content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions TRANSMISSION OF POLITICAL VALUES FROM PARENT TO CHILD 183 spondence between parent and offspring values.34 Since our findings are mostly on the null side, it is important to consider the possibility that interaction effects confound the relationship between family characteristics and transmission patterns. Previous work suggests that affectivity and power relations in the family will be related to parent-child transmission primarily among highly politicized families. Only if politics is important to the parents will acceptance or rejection of parental values be affected by the parent-child relationship. In order to test this hypothesis, student-parent agreement was observed, controlling for family politicization and affectivity or power relations simultaneously. No strong interaction effects emerge from this analysis. The affectivity and power dimensions sometimes affect only the highly politicized, sometimes the most unpoliticized, and at other times their effect is not at all dependent on the level of politicization.35 The lack of impressive bivariate relationships between family characteristics and the transmission rate of political values is not due to the confounding influence of multiple effects within the family. With hindsight, reasons for the failure of the hypothesized relationships bearing on family structure can be suggested. But to give a clear and thorough explanation and test alternative hypotheses will be difficult and time-consuming. One exploratory avenue, for example, brings in student perceptions of parental attitudes as an intervening variable. Another is concerned with the relative homogeneity of the environment for children of highly politicized backgrounds versus youngsters from unpoliticized familes. A third possibility is the exisence of differential patterns of political learning and, in particular, a differential impact of the various socializing agents on children from politically rich versus those from politically barren backgrounds.6 It is also possible that 8' Nor was the intensity of parental feelings related in any consistent fashion to the amount of parent-student correspondence. 35 There is a moderate tendency for those children feeling most detached from their parents to exhibit greater fluctuation in agreement with their parents-at various levels of politicizationthan is true of those feeling most attached to their parents. 36 At another level, the explanation may be in the lack of validity of students' and parents' reports of family structure. See Niemi, op. cit. Ch. II and pp. 184-185. knowledge about later political development of the students would help explicate these perplexing configurations. IV. A CONCLUDING NOTE In our opening remarks we noted the conflicting views regarding the importance of the family as an agent of political learning for the child. This paper has been primarily concerned with a fairly narrow aspect of this question. We sought evidence indicating that a variety of political values held by pre-adults were induced by the values of their parents. Thus our test has been rather stringent. It has not examined the relative impact of the family visa-vis other socializing agents, the interaction effects of the family and other agents, nor the other ways in which the family may shape political orientations. Having said this, it is nevertheless clear that any model of socialization which rests on assumptions of pervasive currents of parentto-child value transmissions of the types examined here is in serious need of modification. Attitude objects in the concrete, salient, reinforced terrain of party identification lend support to the model. But this is a prime exception. The data suggest that with respect to a range of other attitude objects the correspondences vary from, at most, moderate support to virtually no support. We have suggested that life cycle effects, the role of other socializing agents, and attitude instabilities help account for the very noticeable departures from the model positing high transmission. Building these forces into a model of political learning will further expose the family's role in the development of political values. A derivative implication of our findings is that there is considerable slack in the valueacquisition process. If the eighteen-year old is no simple carbon copy of his parents-as the results clearly indicate-then it seems most likely that other socializing agents have ample opportunity to exert their impact. This happens, we believe, both during and after childhood. These opportunities are enhanced by the rapid socio-technical changes occurring in modern societies. Not the least of these are the transformations in the content and form of the mass media and communication channels, phenomena over which the family and the school have relatively little control. It is perhaps the intrusion of other and different stimuli lying outside the nexus of the family and school which has led to the seemingly This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions 184 THE AMERICAN POLITICAL SCIENCE REVIEW different Weltanschauung of the post-WorldWar-II generation compared withit s immediate predecessor. The place of change factors or agents thus becomes crucial in understanding the dynamics at work within the political system. Such factors may be largely exogenous and unplanned in nature, as in the case of civil disturbances and unanticipated consequences of technical innovations. Or they may be much more premeditated, as with radical changes in school organization and curriculum and in enforced social and racial interaction. Or, finally, they may be exceedingly diffuse factors which-result in numerous individual studentparent differences with no shift in the overall outlook of the two generations. Our point is that the absence of impressive parent-to-child transmission of political values heightens the likelihood that change factors can work their will on the rising generation. Shifting demands on the political system and shifting types of system support are natural outgrowths of these processes. This content downloaded from 147.231.52.62 on Mon, 21 Sep 2015 11:52:08 UTC All use subject to JSTOR Terms and Conditions