Tedin -- The Influence of Parents on Childrens Political Attitudes.pdf

The Influence of Parents on the Political Attitudes of Adolescents Author(s): Kent L. Tedin Source: The American Political Science Review, Vol. 68, No. 4 (Dec., 1974), pp. 1579-1592 Published by: American Political Science Association Stable URL: http://www.jstor.org/stable/1959943 Accessed: 21-08-2...

The Influence of Parents on the Political Attitudes of Adolescents Author(s): Kent L. Tedin Source: The American Political Science Review, Vol. 68, No. 4 (Dec., 1974), pp. 1579-1592 Published by: American Political Science Association Stable URL: http://www.jstor.org/stable/1959943 Accessed: 21-08-2014 17:56 UTC 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 [email protected]. American Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to The American Political Science Review. http://www.jstor.org This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions The Influenceof Parentson the Political Attitudesof Adolescents KENT L. TEDIN University of Houston Introduction' A common assumption made by students of socialization is that the family is the major source of cultural transmission in society. The child is said to mirror his parents on a wide variety of social and political values. Ernest Campbell, writing in the 1969 edition of the Handbook of Socialization, claims that "we would be exceedingly surprised were we to discover, in research on any factor whatsoever, that a knowledge of the parents' position or score on a factor did not predict positively to the score of the adolescent."2 This general conception is reflected in works concerned specifically with political socialization. Roberta Sigel asserts, "The similarity of views between parents and children on such issues as foreign affairs, civil rights, economics and many other vital topics of the day is truly phenomenal."3 Herbert Hyman, in a general review of the empirical research related to the transmission of political attitudes (pre-1960), finds this research to "... establish very clearly a family correspondence in views that are relevant to matters of political orientation. Over a great many such correlations from different studies, the median value approximates.50."4 Recently, however, the effectiveness of the family, notably the parents, in the transmission of certain political values has been called into question. R. W. Connell, referring explicitly to Hyman's review, asserts that almost all investigations which report high correlation coefficients between parent-child attitudes have a "self-selec'Financial support for this project was provided by a National Science Foundation Grant for Doctoral Dissertation Research made to Donald J. McCrone (Grant #GS 34218), and by a research grant from the Graduate College of the University of Iowa. I would like to thank Professor Donald McCrone, for his sound substantive advice, and George Woodworth, Assistant Professor of Statistics, University of Iowa, for his advice on the general linear hypothesis. Thanks also to John C. Wahike for a number of helpful suggestions, and to Susan Brantley and John Tatlock for many stimulating discussions relevant to this research. 2 Ernest Campbell, "Adolescent Socialization," in Handbook of Socialization, ed. David A. Goslin (Chicago: Rand McNally, 1969), p. 827. 3 Roberta Sigel, Learning About Politics (New York: Random House, 1970), p. 104. 4 Herbert Hyman, Political Socialization (Glencoe, Illinois: The Free Press, 1959), p. 72. tion" bias in their sampling procedure. This bias, he argues, artificially inflates relationships, leading to an overestimate of the influence parents have on the political attitudes of their children.5 Post-1960 empirical evidence lends support to this contention. Based on the University of Chicago socialization study, Robert Hess and Judith Torney assert that the family has little, if any ability to transmit idiosyncratic attitudes. Rather, its influence is limited to partisanship and to reinforcing those attitudes consensually held.6 The same inference can be drawn from the family socialization project carried out by the University of Michigan's Survey Research Center. In the currently most influential intergenerational study, "The Transmission of Political Values from Parent to Child," M. Kent Jennings and Richard Niemi report relatively low correlations between attitudes held by parents and those held by their adolescent children. Over a range of issues, excluding party identification, the highest coefficient of association was found to be.34, and most coefficients were considerably lower.7 But even in this critical light, it is apparent that parents do influence the attitudes of their children. Negative correlations, in any study, are almost nonexistent. It seems probable that bivariate (uncontrolled) correlations between parent-child attitudes in any reasonably stratified sample, drawn in the early 1970s, are going to be positive and moderately low. Perhaps in a different time period (for example, the Depression, when many of the investigations reviewed by Hyman were conducted) correlations, even in the case of unbiased samples, would have been positive and moderately high. But if concern is with building cumulative knowledge, the population parameter is only of passing interest. The real interest lies in specifying conditional universal propositions. Regardless of the issue, time-frame, and population, under what conditions do parents influence the political I R. W. Connell, "Political Socialization and the American Family," Public Opinion Quarterly, 36 (Fall, 1972), 326-327. 6 Robert Hess and Judith Torney, The Development of Political Attitudes in Children (Chicago: Anchor Press, 1967), p. 113. ' M. Kent Jennings and Richard Niemi, "The Transmission of Political Values from Parent to Child," American Political Science Review, 62 (March, 1968), 172-176. 1579 This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1580 The American Political Science Review attitudes of their children? What, if any, are the general principles involved? Criteria for Studying the Transmission of Political Attitudes and Values Studies of family attitude transmission should clearly demonstrate at least three points: (1) the nature and extent of the attitude correspondence between parents and children,8 (2) the degree to which this correspondence is a function of parental transmission as distinguished from other social forces, and (3) the factors that affect the success of transmission. The Nature of Parent-Child Attitude Correspondence There are three possible types of attitude correspondence between parents and children, which we may call "relative pair," "absolute pair," and "group" correspondence.9 Relative pair correspondence, usually indicated by some type of correlational statistic, is the measure used in most studies. All correlations, whether based on intervals or ranks, determine the extent to which one distribution is a function of another. If a correlation coefficient is high, it means, for example, that the most conservative fathers will usually have the most conservative sons. It does not mean that the absolute scores of the sons on a scale of conservatism are at all close the absolute scores of their fathers. Correlation coefficients are not measures of absolute agreement. On a scale ranging from one to fifty, all the sons could be in the one-to-twenty range and all the fathers in the thirty-to-fifty range, and there could still be an extremely high correlation between their attitudes. Absolute pair correspondence is the extent to which the scale scores of the sons are exactly the same as the scale scores of their fathers, e.g., twenty-five per cent of the father-son dyads have a score of two on the scale or item in question. This correspondence is usually measured, as just indicated, by percentage agreement. Group correspondence is the extent to which the opinion distributions of each generation are similar, regardless of the pair relationships between parents and their own children. For example, fifty per cent of a student sample and fortyfive per cent of a parent sample might favor socialized medicine, a case of close group correspondence. But this says nothing about the correspondence between parents and their own children. Group correspondence is usually measured 8The discussion here will be limited to the nature of attitude correspondence. 9Connell discusses relative pair and group correspondence, but neglects the third possibility-absolute pair correspondence. Connell, pp. 324-325. Vol. 68 by means and standard deviations, or percentages as in the example. Distinctions among the types of correspondence are not always properly made in the literature. Hyman, for example, interprets at least one case of group correspondence as relative pair correspondence.'0 While pair comparisons are most often made with correlation coefficients, terms like "transmission," "similarity," or "congruence" are often used simultaneously in ways that imply absolute pair correspondence. In the following discussion the term "congruence" will be reserved for absolute pair correspondence. Group correspondence will be used as defined above. The terms "transmission" and "influence" will indicate relative pair correspondence. Direct Parental Influence vs. Other Social Influences Direct parental influence is only one possible explanation for parent-child attitude correspondence. An alternative explanation holds that parents and children have similar attitudes because each individual family member is exposed to the same social influences. Both group and pair correspondence could be spurious. Another alternative holds that the political attitudes of children are influenced by parents indirectly." Tests for these possibilities have been extremely infrequent in past research. Spurious Relative Pair Correspondence.The problem of spurious relative pair correspondence will be illustrated by assessing the possibility that it is social class which is causing parent-child attitudes to covary. It can be argued that by belonging to the same social class, parents and children are exposed to similar media, have friends and acquaintances with like attitudes, have the same "class interests," etc. This possibility can be diagrammed as follows: Social Class Parent Attitude Child Attitude Social class is leading parents and children to have similar attitudes in an absolute sense. This may have little meaning for correlation analysis, 10 Connell, p. 325. " For example, many investigators believe that strict child-rearing patterns lead to authoritarianism in child. This authoritarianism could, in turn, "structure" policy attitudes. For a discussion of indirect socialization see Richard Dawson and Kenneth Prewitt, Political Socialization (Boston: Little, Brown, 1969), pp. 65-73; for a critique of the structuring principle see Donald D. Searing, Joel J. Schwartz and Alden Lind, "The Structuring Principle: Political Socialization and Belief Systems," American Political Science Review, 67 (June, 1973), 415-432. This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1974 Influence on the Political Attitudes of Adolescents 1581 particularly if the "third" variable is constant or has little variance. Arguments which claim that correlations are artificially high because the sample is "all of one social class," or "all from one atypical community" are almost certain to be false." In the case of the example, social class would have to be fairly heterogeneous before it would affect the parent-child correlations. The logic would be something like: the lower the family's social class, the higher both parent and child would score on some scale relative to their own group. The spurious relationship hypothesis is usually tested by partialing for the third variable." successful than if the parent is not concerned with the issue. Another condition might be the accuracy of the child's perception of the parent's attitude. If parents are directly influencing their children, an important prerequisite would seem to be some knowledge of the parent's attitude by the child. But if correlations are spurious, or influence is indirect, it should make no difference whether or not an issue is salient to a parent or whether the child knows or does not know the parent's attitude. Variance in these conditions should not affect the magnitude of correlation coefficients."4 Spurious Group Correspondence. In many circumstances group correspondence will be affected by common environmental variables. For example, religion may affect both parent and child attitudes toward legalized abortion. If the environmental factor varies, the possibility of a common effect on the two groups can be tested with a difference-in-means statistic. Often, however, there is no variance on these factors. An example would be a common community effect when the sample is drawn entirely from one community. There is no way to reject the hypothesis that group means are similar because parents and children live in the same place. But if the research concern is with parental influence, commonly affected mean values may be of little interest, as long as there is ample variance in parentchild attitudes. The research question remains: "To what extent are deviations from the child mean scores influenced by deviations from the parent mean scores?" Correlation coefficients indicating the success of attitude transmission vary greatly from issue to issue and from one study to another. However, attempts to explain these variations, i.e., to specify the conditions of influence, have been notable only for their lack of success. Jennings and Niemi, for example, did not find a single variable having any significant impact on successful transmission. Some have suggested that the content of an attitude, the attitude object, is consequential for the success of transmission." More plausible, in my view, is a failure to specify all conditions relevant to successful transmission. What might these conditions be? Some insight can be gained by analyzing those attitudes successfully transmitted on a consistent basis. Of particular interest is party identification. The influence of parents on the partisanship of their children has been well documented. This attitudinal predisposition is surrounded by a somewhat unusual set of circumstances. Party identification is subject to considerable discussion in many families at regular intervals, i.e., elections. In addition, many parents convey to their children quite early that "we" are Democrats or "we" are Republicans."' Party identification is different, then, from most other attitudes in its visibility to the child and its salience to the parents. The important point to note is that these are population parameters. They are not inherent characteristics of the attitude object.'7 The influence of these variables can be moved to the table marginals and controlled. Direct Parental Influence and the Magnitude of Correlation Coefficients. While the procedures just outlined may be used to reject the spurious alternative to parental influence, another method is better for testing the extent to which correlations between parents' and children's attitudes support the direct parental influence hypothesis. The method involves varying certain conditions which might be expected to affect the strength correlations. One condition might be issue salience to the parent. If an issue is salient to a parent, we would expect attitude transmission to be more A constant will not usually inflate correlations. One instance, however, in which it would inflate correlations is when the constant causes a large percentage of both parent and child scores to fall into one rank or interval. It is important, therefore, to check the distribution of the attitude scores for both parents and children. '" A reduction due to partialing does not, of course, necessarily mean a spurious relationship exists. A redUictionalso occurs when a developmental sequence is present. 12 Specifying Conditions of Parental Influence " This methodology is discussed by Frank F. Furstenberg, "The Transmission of Attitudes in the Family" (Ph.D. Dissertation, Department of Sociology, Columbia University, 1967), chapter 3. 1" Stephen Wieting, "Family Factors and the Re(Ph.D. Disserligious Belief Systems of Adolescents" tation, Department of Sociology, University of Minnesota, 1971), p. 87. (New Children and Politics "'Fred Greenstein, Haven: Yale University Press, 1965), pp. 71-72. 17 There are, however, differences between party identification and issues which are inherent in the attitude This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1582 The American Political Science Review Perceptual Accuracy. At some point in any study of social influence it would seem imperative to determine if the person who is believed to be influenced is aware of the attitudes and values of the person or group believed to be exerting influence. This point is often overlooked in studies of social influence. Parents may be exerting considerable influence on the attitudes of youth, but it may be of the subjective sort. That is, children are identifying with what they think or perceive are their parent's attitudes.'8 Perceptual accuracy can strongly influence successful transmission. In Furstenberg's study of mobility aspirations, the parent-child correlation for the group in which student respondents were uncertain of (didn't know or misperceived) parental attitudes was actually negative (tau-b.183). But when perception was accurate, the coefficient was a positive.480.19 There is a strong possibility that parent-child attitude correlations, for all issues, are highly dependent on the marginal distribution of perceptual accuracy for the attitude in question. Issue Salience. The finding of Jennings and Niemi that the level of political interest had little or no effect on the success of value transmission surprised many observers. This failing, and the failure of other specification variables, seems rooted in a research assumption which has seldom been questioned-the idea that some families are more successful than others in transmitting attitudes. Investigators have ignored the cognitive and emotive factors surrounding specific issues and have instead focused on all-encompassing variables like parent-child affect, parental authoritarianism and, in the same intellectual tradition, political interest. Successful transmission, rather than essentially being a family characteristic, may essentially be issue specific. That parents are "politically interested" does not mean they necessarily have interest in any particular given issue. The reverse is also possible; parents objects. For example, "independent" for party identification is not equivalent to "undecided" in the case of issue attitudes. This point will be discussed again later in the paper. 18 Richard Niemi reports that adolescents in the Michigan socialization study perceived themselves as attitudinally closer to parents by an average of.10 (tau-b) than they were in actuality. He suggests that this misperception serves an important psychological function for the youthful respondents by reducing whatever tension may arise from asymmetry in parentadolescent political attitudes. Richard G. Niemi, "Collecting Information about the Family: A Problem in Survey Methodology," in Socialization to Politics, ed. Jack Dennis (New York: John Wiley and Sons, 1973), p. 480. '9 Furstenberg, p. 172. Vol. 68 may find some political issues quite important without having any general interest in politics. To the extent that the general level of political interest and the salience of any one issue diverge, the effect of political interest on the successful transmission of the attitude toward that issue is likely to be reduced. Family Interaction. Most family interaction variables, such as parent-child affective relations, have been found to have either a slight positive effect or no effect on successful parental attitude transmission. Family interaction, not being specific to issues, may be too general to have much impact. It is conceivable that the influence of family variables is severely attenuated by a child's inaccurate perception of the parental attitude. All the family conditions necessary for parental influence could be present, and the child may in fact want to identify with the parent attitude. When questioned, he may give a response he believes similar to the attitude of the parent. But if the parental attitude is misperceived, the child's response is treated as an "error." According to theory, he should have shared the parent attitude, but did not. There is no reason to believe that there is any consequential relationship between variables like parent-child affect, parental authoritarianism, and accurate perception. It comes as no surprise, therefore, that when positive effects are reported for a family variable, the attitude object is almost invariably one that is highly visible, like party identification. The study on which the following discussion is based was designed to investigate the transmission of attitudes from parents to children in light of the above considerations. I shall proceed by first examining the extent and character of the attitude similarity between parents and children in terms of relative pair, absolute pair and groud correspondence. Next will be an analysis of the extent to which covariation between parent and child attitudes (relative pair correspondence) is due to direct parental influence as opposed to other factors. Finally, I shall attempt to elaborate, as generally as possible, the conditions requisite to successful transmission. The Sample and Measurement The Sample The sample for this analysis consists of 322 parent-child dyads. The "children" are recent high school graduates. Recent graduates were selected because at the time of the interview most were fully enfranchised members of the political community, and almost all would be eligible to vote in the 1972 presidential election. This procedure avoids the common criticism of social- This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1974 Influence on the Political Attitudes of Adolescents ization research that the subjects are too young to participate in the political process, and that by the time they come of age the attitudes and values they learned as children may have completely changed.20 Another reason for choosing recent high school graduates is that these new adults are soon to leave the family permanently, and further family socialization will be minimal. The data were collected by means of personal interview in Iowa City, Iowa (population 47,000), between June 1 and September 1, 1972.21 The sampling universe consisted of all graduating seniors from the city's largest high school (who gave Iowa City addresses) and their parents. Eighty per cent of the adolescents and at least one parent in each case were successfully interviewed.22 The sample, quite obviously, is not randomly drawn from a finite population. But this is of limited importance since the purpose here is not in describing population parameters, but in developing "conditional universal propositions." When the purpose of research is testing conditional universals, a random sample is not necessarily more useful than a nonrandom sample.23 A defensible and financially more practical procedure for building cumulative knowledge involves testing hypothesized relationships under a variety of conditions. In this paper, hypotheses will be tested using a basically white, urban, middle-class sample. Demonstration of the generalizability of the findings entails replication using, for example, a rural sample, a black sample, or possibly a non-American sample. The hypotheses can, of course, never be confirmed; they can only be disproved. Measurement Attitudes were measured by responses to closedended, Likert-type items in a personal interview 20 See David March, "Political Socialization: The Implicit Assumptions Questioned," British Journal of Political Science, 1 (October, 1971), 460-464. 21 Interviewers consisted of eight male political science and sociology graduate students between 25 and 32 years old. 22 The sample consists of 183 adolescents (54 per cent male, 46 per cent female) and 322 parents (45 per cent male, 55 per cent female). Data on both parents were obtained for 139 of the adolescents. Twenty-six adolescents came from one-parent homes. In the case of the remaining eighteen adolescents, the second parent was not home at the time of the scheduled interview or call-backs. I shall not, in this report, distinguish between mothers and fathers. Therefore, to make maximum use of the sample size each adolescent is paired with both the mother and the father (when both are present). This procedure renders tests of statistical significance based on F-ratio's inappropriate. See footnote 43 for a further discussion. 23 David Willer, Scientific Sociology (Englewood Cliffs, N.J.: Prentice-Hall, 1967), chapter 6. 1583 situation. The questionnaires were designed to secure information about the young adult, his parents, his best friend, and his perception of certain attitudes of his parents and best friend. This information consists basically of party identification, attitudes toward three issues, evaluation of eight public figures, and a series of questions relating to adolescent/parent/best friend interactions. This paper will be limited to an analysis of parents and adolescents with respect to party identification, the three issues, and certain interaction variables. Selection of the issues and their indicators was made only after numerous pretests, first on college students, then on college students and their parents, and finally on a small sample of high school seniors and their parents. The issues finally chosen were: (1) racial integration of the public schools, (2) United States policy toward Communist China, and (3) laws concerning the use of marijuana (these questions are reproduced in the appendix). The attitude toward each issue was measured by a scale consisting of four opinion items, which were scattered throughout the measuring instrument to minimize any tendency to respond to the items in a blanket fashion. Control of response set was effected by wording some of the items negatively and some positively. The reliability and validity of the scales derived from these questions is, of course, crucial. Fortunately, these estimates can be made with some confidence. Since each scale is used three times (once for the adolescent, once for the adolescent's perception of the parent's attitude, and once for the actual parent attitude), there are three opportunities to check on scale properties. Reliability (defined as internal consistency) was estimated using Cronbach's alpha.24 Nine coefficients were computed. All were above.70, which is satisfactory. As for validity, a good argument can be made for the content validity of the scales used here. They are composed of specific opinion items measuring attitudes toward concrete relatively well-defined concepts (issues). More important is the assessment of construct validity-the extent to which relationships between items are isomorphic with the relationships predicted by the nominal definition of the concept. This form of validity can be estimated by factor analysis, which will indicate those items meas 'ring the same concept, and how successful they are in this measurements 24 George W. Bohrnstedt, "Reliability and Validity Assessment in Attitude Measurement," in Attitude Measurement, ed. Gene Summers (Chicago: RandMcNally, 1970), pp. 87-88. 25 Fred N. Kerlinger, Foundations of Behavioral Research (New York: Holt, Rinehart and Winston, 1964), pp. 453-455. This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions The American Political Science Review 1584 Table 1. Product-Moment Correlations between Parents and Adolescents on Three Issues and Party Identification Means Attitude Objects Corre-lation Racial Integration China Policy Marijuana Laws Party Identification.32.27.40.48 Parents Adolescents 8.8 6.4 9.2 4.4 7.4 5.4 6.9 3.9 For purposes of issue description, a scale ranging from 1-17 was computed by the additive method. The higher the scale value, the more conservative the respondent. For party identification (measured by the standard SRC seven-point scale), the higher scale values indicate Republican preference. All twelve opinion items were factored together, once for the adolescents, once for the adolescent's perception of the parent's attitudes, and once for the actual parent attitudes. It was hypothesized that three distinct factors would emerge, with all the racial integration items loading on one factor, all the China policy items on another factor, and all the marijuana items on a third factor. The results supported the hypothesis. The analysis revealed an extremely clean factor structure, with only one of the predicted thirty-six loadings (four items, three factors, three administrations) falling below.50.26 The actual scale values for each individual respondent were computed with factor scores.27 Bivariate and Within-Category Relationships Bivariate Relationships The product-moment correlations between parent and adolescent attitudes, presented in Table 1, are as high as or higher than one might expect on the basis of recent literature. The correlations for party identification and racial integration are comparable to those found by Jennings and Niemi (.59 and.34, respectively).28 These authors also report a number of very lowcorrelations, but for attitude objects which lacked immediacy at the time of the interview (1965). When immediacy is comparable, the rela28 The common factor model was used with communalities estimated by squared multiple correlation coefficients. The rotation was oblique. The item falling below.50 was the first racial integration item (see appendix) which loaded at.38. 2"The factor scores were estimated by multiple regression following the procedures outlined in Harry Harmon, Modern Factor Analysis (Chicago: University of Chicago Press, 1968), pp. 350-354. 28 Jennings and Niemi use the product-moment coefficient for party identification, and tau-b for the issues. Vol. 68 tionships reported here and those reported by Jennings and Niemi are similar. In terms of group correspondence, the adolescents are in all cases more liberal than the parents but the differences indicate no great generation gap. The standard deviations (data not presented) on each issue for the two groups are almost identical. Absolute congruence between parents and adolescents can best be described as moderately high. This information was computed by subtracting the adolescent's scale score on each issue (computed by the additive method) from the complementary parent's score. These difference scores had a possible range of ? 16, with zero indicating perfect agreement. The means of these difference scores for the race issue, China policy, and the marijuana issue were 1.3, 1.0 and 2.2, respectively.29 Given this information, the correlation coefficients can be interpreted as indicating a functional relationship between attitudes that are reasonably similar in an absolute sense. There exists no large gap between parent and adolescent attitudes in either group or congruence terms. Conditions for Parental Influence (Within-Category Relationships) Having established the magnitude of the parentadolescent attitude correlations, I shall now attempt to specify the conditions which are conducive to successful transmission. There are two reasons for this analysis. First, it is consequential for determining whether parent-adolescent attitude correlations are due to parental influence or due to other social forces.30 If attitude correlations between parents and adolescents are spurious, it should make no difference for the magnitude of coefficients if certain specification variables are manipulated. Beyond demonstrating or failing to demonstrate direct parental influence, the analysis is designed to uncover conditional universal propositions concerning parent-child attitude transmission. 29 Positive scores mean adolescents are more liberal than parents, negative scores mean the reverse. The distribution of these scores approximates a normal curve. The standard deviations for the race issue, China policy and marijuana laws are 4.1 3.7, and 4.3, respectively. " The most plausible hypothesis, alternative to direct parental influence, is that parents and children have similar attitudes because each individual family member is exposed to the same social influences. The example used earlier was social class. Variables of this type may determine the position of the parent and child within their respective groups thus producing a spurious correlation. One test of the spurious relationship hypothesis is controlling for a third variable. Controls were introduced for education, social class, and religion; at no point, for any issue or party identification, was the bivariate correlation reduced by more than.02. This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1974 Influence on the Political Attitudes of Adolescents Perceptual Accuracy. Accurate perception of the parental attitude by the child has been shown by Furstenberg to have a very significant effect on the success of attitude transmission-at least in the case of mobility aspirations. The same should be true of political attitudes, with those parental attitudes most accurately perceived by adolescents being those which are the most successfully transmitted. Hypothesis1: The more accurate the adolescent's perception of the parent's attitude, the more influence the parent will have on the adolescent's attitude. 1585 ceptual accuracy. Children are simply more likely, for reasons alreadydiscussed,to perceive parental party identification more accurately than they perceive attitudes toward current issues. Table 2 shows that while 72 per cent of the adolescents have a correct perception of their parent's party identification,not more than 12 per cent have a completely accurate perception of their parent'sattitude toward any one issue. Evenwhen the two categoriesindicatingthe most accurate perception are collapsed together, not more than 36 per cent could be classedas having an accurateperceptionof any one parentalattitude. Perceptual accuracy for party identification was operationalized by asking each adolescent to report the party identification of his parents. This Theseobservationslead to the secondhypothesis: 2: Parentswill haveequalinfluenceon Hypothesis report was collapsed into three categories (Demothe partyidentification and issueattitudesof adocrat, Independent and Republican) and compared accuracyis heldconstant. lescentswhenperceptual with the actual parent party identification (also collapsed into three categories). If the two were The correlationsnecessaryto test hypothesis1 congruent, perception was classed as accurate. and hypothesis2 are presentedin Table 3. If they were not congruent or the adolescent reThe effectsof perceptualaccuracyon parental sponded "I don't know," perception was classed influence are very strong and offer substantial as uncertain. The situation for the issues is a bit supportfor hypothesis1.32 The data, in addition, more complicated. These attitudes were measured supportthe directparentalinfluenceexplanation with four opinion indicators. Adolescents were for parent-adolescentattitude correlations. For asked to report parent opinions for all four items all four attitudeobjects there is a linear increase using a five point agree-disagree scale. A sixth in parentalinfluenceas perceptionbecomesmore option was an "I don't know" response. These accurate.As predicted,adolescentsuncertainof items and the corresponding parent items were parental attitudes are subject to almost no pacollapsed into three categories-agree, undecided, rentalinfluence. disagree- and a comparison was made to deterHypothesis 2 also receives support from the mine the accuracy of perception. It is possible for data presentedin Table 3. Within categoriesof adolescent respondents to manifest uncertainty perceptualaccuracy,the coefficientsfor the issues on all four indicators at one extreme, or to per- compare quite favorably with those for party ceive the parent's opinions correctly on all four identification.This correspondencesuggests the items at the other. The result is a scale ranging operation of a general process, with the prinfrom zero to four, the higher score indicating ciples involved in the transmissionof attitudes the greater accuracy. toward issues being the same as those involved Some analysts assume that party identification in the transmissionof party identification.Note is an attitude intrinsically different from those also that the attitudetowardmarijuanalaws does toward issues or public figures, believing it is this not deviate from the overall pattern. If there difference which accounts for its successful trans- exists, as claimed by some sociologists, a youth mission. Connell, for example, sees a differ- culture exerting countervailinginfluenceto that ence in social patterning-the objects of opinion of parents,it should at this point be manifested. being institutions, and the choice among them being a relatively simply dichotomy. Jennings and Issue and PartisanSalience. Jenningsand Niemi Niemi, in discussing the successful transmission found that political interesthad no effect on the of party identification, note the low correlations success of attitude transmission.The same patobserved in other domains of political life and tern, using the same operational indicator, is assert that there is no ground for any "general 32 Perceptual accuracy was collapsed into three catethesis" explaining attitude transmission." It was gories to insure that each category had an adequate observed earlier that there are inherent differ- sample size. The linear patterns was, however, even ences between party identification and other more pronounced in the five-category case, although attitudes. But there are also noninherent differ- the N's were small. In the five-category breakdown, ences which are controllable. One of these is per- the correlations in the category exhibiting the greatest for the race issue, China policy, 31 Connell, pp. 329-330, Jennings and Niemi, p. 169. perceptual accuracy and the marijuana issue were.66,.72, and.61, respectively (based on 30-35 cases). This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1586 The American Political Science Review Vol. 68 Table 2. Distribution of Perceptual Accuracy for the Issues and Party Identification Perceptual Accuracy Attitude Objects Racial Integration China Policy Marijuana Laws Party Identificationb Zero One Two Three Four 11%0 24%0 15%o 28%0 23%o 19%o 23%o 72%0 30%0 20%0 28%o 24%0 24%0 24%0 11%0 12%0 09%o - Totala 100% (N= 318) 100%0(N=318) 100% (N= 317) 100%o(N= 319) a May not total to 100%,0because of rounding. Perceptual accuracy for party identification is a dichotomy. The adolescent either knows (one) or does not know (zero) the parent's attitude. b found in the data employed here.33The parentadolescent correlations within categories of political interest appear to be randomly distributed (data not presented). The next step is to substitute issue and partisan salience for political interest. Salience of party identification to the parent was operationalized by breaking parental partisanship into three categories-independents, weak partisans, and strong partisans. A scale of parental issue salience was constructed using three items: (1) how important is the issue to the parent, (2) how often do the parent and adolescent talk about the issue, and (3) would it make any difference to the parent if the adolescent adopted an attitude on the issue different from that of the parent. The scale ranged from one to seventeen with the median scores for the race, China, and marijuana issues being 7.8, 6.8, and 11.4, respectively. It is obvious that the marijuana issue is considerably more salient to parents than are the other two issues. Since the distribution of perceptual accuracy did not differ on the three issues, we immediately suspect that the reason for the higher parent-adolescent correlation on the marijuana issue, as opposed to the other two issues, is the greater issue salience to parents. Hypothesis3: The greater the issue and partisan salienceto the parent,the more influencethe parent will have on the adolescentattitude. "The question is: "How often do you discuss politics with your son or daughter?" (1) never, (2) a few times a year, (3) several times a month, (4) several time, a week. 34 It is possible that adolescents are influencing parents. But if this were the case, correlations should not be so strongly affected by adolescent perceptual uncertainty, unless there is a functional relationship be- Table 4 indicates clear support for the hypothesis, although the relationship in the "strong partisan" category for party identification is somewhat greater than the "high salience" category for the issues. It must be noted, however, that salience for the issues and salience for party identification are not exactly comparable. The term "Independent" for party identification does not have the same connotations as the term "low" for issue salience. It is possible for independents to be highly concerned and interested in partisan politics, while "low" issue salience necessarily entails low "interest" in that issue. Finally, Table 4 like Table 3, offers strong support for the "direct parental influence" explanation for parentadolescent attitude covariation. While we cannot rule out all explanations alternative to parental influence, the direct influence hypothesis seems at this point the most plausible.34 Table 3. Parent-Adolescent Product/Moment Correlations Within-Categories of Perceptual Accuracy Perceptual Accuracy" Attitude Objects Racial Integration China Policy Marijuana Laws Party Identificationb Uncertain -.15 -.08.12.01 (107) (139) (122) (90) Medium Accurate.30 (96).21 (65).25 (90) Accurate.56.51.60.57 (115) (116) (106) (229) Based on Table 2 (for the issues), the scores of zero and one for perceptual accuracy were combined for the uncertain category; the score of two equals the medium accurate category, and the scores of three and four were combined for the accurate category. b Perceptual accuracy for party identification is a dichotomy. This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1974 Influence on the Political Attitudes of Adolescents 1587 Table 4. Parent-Adolescent Issue Attitude and Partisan Correlations Within-Categories of Issue and Partisan Salience to the Parent Parental Issue Salience Attitude Object Racial Integration China Policy Marijuana Laws Low Medium High.00 (64).03 (73) -.21 (35).27 (127).25 (142).35 (71).48 (125).44 (100).48 (211) Parental Partisan Salience Independent Weak Partisan Strong Partisan.31 (73).68 (83).28 (163) party Identification InclUdes both "pure" Independents and Independents who "lean" toward either the Republicans or the Democrats. Family Interaction. Family interaction variables were criticized earlier as being too general to have much impact on successful transmission. Hypothesis 4: Conditions of family interactionwill not affect the success of attitudetransmission. I have chosen for analysis two of the most frequently used family variables-family authoritarian ideology and parent-child affective relations. Many students of the family believe that authoritarian as opposed to permissive child-rearing will affect the success of attitude transmission. Empirically, Jennings and Niemi, as well as Middleton and Putney, report a curvilinear relationship between parental authoritarianism and successful transmission.35 Other investigators, however, tween adolescent uncertainty over parental attitudes and parental uncertainty over adolescent attitudes. Any conclusive test would require longitudinal data. Another nonrejectable alternative is the possibility that parents and peers have attitudes which are strongly related. If this were the case, it could be argued that the adolescent's attitude is due to peer influence. 35Jennings and Niemi, p. 181, and Russell Middleton and Snell Putney, "Political Expression of Adolescent Rebellion," American Journal of Sociology, 68 (March, 1963), 532-533. have reported no relationship.36 The measure of family authoritarian ideology used here is a fiveitem scale taken from Ernhart and Loevinger.17 The data support hypothesis 4. The within-category correlations display no patterned distribution, either linear or curvilinear.38 It is also commonly thought that warm parentadolescent interpersonal relations should lead to a desire by the adolescent to adopt the values and attitudes of the parent. Jennings and Niemi found parent-adolescent affective ties had no influence on the success of transmission. Middleton and Putney, on the other hand, report a positive effect.39 The parental "attractiveness" scale employed here is operationalized by items in36Philip Nogee and Martin Levin, "Some Determinants of Political Attitudes among College Voters," Public Opinion Quarterly, 22 (Winter, 1958-1959), 460. 3 Claire B. Ernhart and Jane Loevinger, Authoritarian Family Ideology: A Measure, Its Correlates, and Its Robustness (Austin, Texas: Society for Multivariate Experimental Psychology, 1969). 38A four- as opposed to three-category breakdown is reported because any curvilinear pattern is more likely to be evident in the former. The results with three categories reflect the same pattern. Jennings and Niemi, p. 181, and Middleton and Putney, pp. 532-533. Table 5. Parent-Adolescent Product/Moment Correlations Within-Categories of Authoritarianism for the Issues and Party Identification Parental Authoritarianism Attitude Objects High Med-High Med-Low Low Racial Integration China Policy Marijuana Laws Party Identification.32 (62).16.55.31.31 (85).23.32.58.24 (93).31.32.40.38 (66).35.36.46 This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions The American Political Science Review 1588 Table 6. Parent-Adolescent Correlations WithinCate-ories of Parental Attractiveness for the Issues and Party Identification Parental Attractiveness Attitude Object Low Racial Integration China Policy Marijuansi Laws Party Identification.17 (76).14.46.24 Medium.36 (167).26.37.58 High.42 (67).41.46.52 quiring into how "close" the adolescent feels to the parent, and how much the adolescent "admires" the parent. It is apparent that, contrary to hypothesis 4, attractiveness does in some instances have an effect on successful transmission. But the pattern is unstable. The transmission of attitudes toward marijuana laws seems impervious to changes in attraction. On the other hand, there is a moderately strong effect on the other two issues and party identification (although there is a mild reversal in the calf' of the latter). Multivariate Models Statistical Methodology Both issue salience40 and perceptual accuracy, taken by themselves, have effects on successful transmission as strong or stronger than any variable reported in past research. In this section it shall be determined how these variables, along with "attractiveness," operate together. This will require a multivariate statistical model. The model employed is a test of the general linear hypothesis for multiple covariance. The specific form of the model used is somewhat complex going a step beyond the analysis of covariance. Conceptually, the dependent variable is parental influence. Operationally, this is not a single variable, but a relationship between two variables. Its indicator is the regression of the adolescent attitude on the parent attitude. Analysis of covariance would test the effect of a categorical variable on the regression intercept. First order statistical interaction, or nonadditivity, exists if the slope of the relationship between parent and adolescent attitudes is dependent on a second (or nth) independent variable. Clearly, the hypotheses predict this interaction. That is, within categories of a second independent variable-for example, issue salience- the slope of the parent-adolescent attitude relationship should increase as issues become more salient to parents. Consequently, the firstorder interaction will be built into the statistical 4" From this point forward, the term "issue salience" will be used to refer to partisan salience as well as issue salience. Vol. 68 model. This model is the summary analog of comparing relationships within categories of a specification vaiiiable or variables-the method used to test hypotheses in the previous section.41 The procedure might be thought of as a two-way analysis of variance, but instead of means, the cell entries are regression coefficients.42As in all tests of the general linear hypothesis, a maximally predictive equation is fitted to the data, and against it are tested less complex equations. If the F-ratio resulting from a comparison of these models is statistically significant, it means the additional terms in the more complex model make a difference.40The procedure, in addition, yields multiple correlation and partial regression coefficients. Issue Salience and Perceptual Accuracy The effect of issue salience and perceptual accuracy on parental influence are combined in a number of models, presented in Figure 1, which are consistent with my conceptual assumptions. I am assuming parental influence to be the degeneral form of model is: a. + a, + b0Xj + b1Xj + e, where a0 represents the grand mean, a, represents the deviation of the grand mean from the kth category, b0Xj is the regression of Y on the interval independent variable which is common to all categories, bjXj represents the deviation of the optimum regression function in the kth category from the common slope. The question of interest is whether or not the common regression is parallel to the within-category regression. If they are parallel, it means that categorizing the parent attitude (by, for example, issue salience) has no effect. If the, common regression slope and the slope computed within categories are not parallel, it means the regression within categories provides the line of best fit. The sums of squares due to fitting the line within the categories will thus be greater than the suxns of squares due to the common regression. The explained variance and the correlation coefficient will increase because of this nonparallelism. See John E. Overall, "Multiple Covariance Analysis by the General Least Squares Regression Method," Behavioral Science, 17 (May, 1972), 313-320. 42Most readers are no doubt aware that two-way analysis of variance computed by standard methods requires equal cell frequencies. Unbiased estimates are usually obtained, when cell frequencies are not equal, by using the weighted squares of means approximation. The multiple covariance linear model is the mathematical equivalent of this approximation. James Fenngessey,"The: General Linear Model: A New Perspective on Some Familiar Topics," American Journal of Sociology, 74 (July, 1968), 4. 43Tests of statistical significance based on F-ratio's are not appropriate for these data because the least squares assumption of "independence" in independent random samples is not met (183 adolescents are paired with 322 parents). This means that the degrees of freedom for the residual sums of squares cannot be estimated in an unbiased manner. Consequently, the mean square residual (the denominator in the F-ratio) cannot be computed. The sums of squares, the crossproducts, and the degrees of freedom for the nurierator are, however, unaffected. 4'The Y This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1974 Influence on the Political Attitudes of Adolescents pendent variable, and that issue salience is antecedent to perceptual accuracy. These assumptions are not unreasonable. It is difficult to imagine parental influence as a determinant of the other two variables.44 Also, conditions id which perceptual accuracy on the part of the adolescent might lead to issue salience for the parent, are difficult to fathom. One intuitively suspects that perceptual accuracy is largely a function of issue salience. If an issue is important to the parent, it will be discussed or commented on within the confines of the family, and other family members will become aware of that particular parental attitude. +' If, however, parental influence were indirect (for example, overprotected adolescents being conservative on China policy) it is possible that this indirect influence could lead to adolescent perceptual 'accuracy. Adolescents, because they are influenced, could be motivated to learn the parent's attitude. But if parental influence is direct, as is argued for the attitudes studied here, it is very unlikely that parental influence could be antecedent to adolescent perceptual accuracy. 1 AMdorel I - Model2 P I Model 3 P Z I Z FZ Model 5 Model4 ,- P P liP 1589 Table 7. Product-Moment Correlations Between Parental Issue Salience and Adolescent Perceptual Accuracy for the Issues and Party Identification Attitude Object CorrelationBetween Issue Salienceand PerceptualAccuracy Racial Integration China Policy MarijuanaLaws Party Identification.25.22.25.23 Hypothesis5: The greater the issue salience to the parent,the more accuratethe adolescentperception of the parentalattitude. The data only marginally support hypothesis 5. There is unquestionably a relationship, but it is obvious that factors other than issue salience are responsible for a considerable amount of the variance in the perceptual accuracy of the adolescent. Nevertheless, while the correlations between issue salience and perceptual accuracy can best be described as modest, they are consistent across all three issues and party identification, and are strong enough to warrant rejection of Model Two. The multiple covariance model will now be used to analyze the independent and combined eAfects of issue salience (I) and perceptual accuracy (P) on parental influence (Z), where Zzt=ao+boX (the regression of the adolescent attitude on the parent attitude). 3 3 9 Hypothesis 6: Y=ao+ai+ap+aip+boX Z Figure 1. Models of 10-terrelationships Amll Salience, Perceptual Accurcy and Parental lafluence Issue I- Parental Issue Salience P -=Adolescent Perceptual Accuracy =-Parental Influence on the Ad lescent Attd Model 1: This model posits*^Fwith.a dirt -effecton Z. and a spuriousFelations~i~p between-P andZ The two latter covary only because of the com'ori in. fluence of I. Model 2: Model two posits no relationship between I and P. but with each having an independent",additive effect on Z. I leads -to P Model3: T~hedevelopmental-seqjuenck.~ which in turn leads to Z.X Model 4: The developmental sexfenc'e; buvwth wit. having an independenteffect-von i~nadditfoiton its effect through P. Model 5: I and P act on Z in a msriadditive manner. 3 3 9 +(bi +bp+bip)X+e.45 With the additive effects of issue salience and perceptual accuracyheld constant, parentalinfluenceis affected by a nonadditive combination of issue salience and perceptualaccuracy. 3 3 3 3 hypothesis 7: Y=ao+ai +ap+boX+(bi+bp)X+e. Witb issue salience held constant, parental influence is affected by perceptual accuracy. 3 3 3 3 fltpothesis8: Y=ao+ap+ai+boX+(bp+bi)X+e. With perceptual accuracy held constant, parental influence is affected by issue salience. 41In these equations the a's are the intercepts and the b's are the regressions. The a's are necessary for unbiased estimates of the within-category slopes (their effects on Y are partialed out). The numerals above the aig and b's are the number of categories used Rt c'onstricting the dummies and the first-order interactiOn terms (although perceptual accuracy for party identification has two categories rather than three). This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions The American Political Science Review 1590 Table 8. Effects of Issue Salience and Perceptual Accuracy on Successful Parent-Adolescent Attitude Transmission Explained Variance (Explained S.S./d.f.) Multiple Correlation. Racial Attitude Parent Attitude 228.0.32 Test of Hypothesis 7: Effect of Perceptual Accuracy, Control Issue Salience Issue Salience 37.5.38 Perceptual Accuracy, Control Issue Salience 99.5.48 Test of Hypothesis 8: Effect of Issue Salience, Control Perceptual Accuracy Perceptual Accuracy Issue Salience, Control Perceptual Accuracy Interaction 113.0.44 24.0.48 15.0.51 China Attitude Parent Attitude 97.0.27 Test of Hypothesis 7: Effect of Perceptual Accuracy, Control Issue Salience Issue Salience 21.5.32 Perceptual Accuracy, Control Issue Salience 67.5.45 Test of Hypothesis 8: Effect of Issue Salience, Control Perceptual Accuracy Perceptual Accuracy Issue Salience, Control Perceptual Accuracy Interaction 73.5.41 15.5.45 7.0.46 Marijuana Attitude Parent Attitude 411.0.40 Test of Hypothesis 7: Effect of Perceptual Accuracy, Control Issue Salience Issue Salience 48.0.44 Perceptual Accuracy, Control Issue Salience 47.0.48 Test of Hypothesis 8: Effect of Issue Salience, Control Perceptual Accuracy Perceptual Accuracy Issue Salience, Control Perceptual Accuracy Interaction 62.0.45 33.0.48 8.0.49 Party Identification Parent Attitude 165.0.48 Test of Hypothesis 7: Effect of Perceptual Accuracy, Control Issue Salience Vol. 68 9.5.50 Issue Salience PerceptualAccuracy,Con27.0.54 trol Issue Salience Test of Hypothesis8: Effect of Issue Salience, Control PerceptualAccuracy PerceptualAccuracy Issue Salience Interaction 32.0 7.0.52.54 0.5.54 The information necessary to test these hypotheses and reject the remaining models is presented in Table 8.46 Hypothesis 6 and the accompanying Model Five can first be rejected. This model posits interaction, but it is apparent from Table 8 that the effects of I and P are strongly additive. When these effects are controlled, the increase in the multiple correlation coefficient due to the secondorder interaction is quite small. Examining the variance explained by issue salience and perceptual accuracy, it is obvious that the latter has the far greater impact on parental influence. Model One would be consistent with the data only if a control for issue salience very sharply reduced the effect of perceptual accuracy. One can see this is definitely not the case. With issue salience controlled, perceptual accuracy retains a strong impact.47The arrow between P and Z cannot be removed, and Model One must be rejected (hypothesis 7 is supported). Remaining is Model Three, a developmental sequence, and Model Four, the same but with I having an independent effect in addition to being part of a developmental sequence. Both issue salience and perceptual accuracy, taken by themselves, have a strong influence on successful transmission. Controlling for perceptual accuracy reduces by 26-36 per cent the variance explained by issue salience. This is evidence for the operation of a developmental sequence, as depicted in Model Three. Finally, it can be asked whether issue salience has an independent as well as a developmental 46The information presented in Table 8 will be used to test hypotheses 6, 7, and 8, and to partition the explained variance between issue salience and perceptual accuracy. Total explained variance for the partitioning analysis is that variance explained by issue salience and perceptual accuracy. It is important to note that by the terms "issue salience" and "perceptual accuracy" is meant the parental attitude categorized by these two variables, or the effect of the two variables on the parent-adolescent regression (parental influence). 4 Using the race issue as an example, entering perceptual accuracy immediately after the parent attitude increases the correlation coefficient from.32 to.44. Entering perceptual accuracy after the parent attitude and issue salience (issue salience controlled) increases the correlation from.38 to.48. This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions Influence on the Political Attitudes of Adolescents 1974 effect. If controlling for P substantially wipes out the variance explained by I, we know there can be no arrow between I and Z. The data show that with P controlled, I accounts for at least seventeen and as much as thirty-five per cent of the explained variance. While not high, it seems to me that this impact, being present for all three issues and party identification, is large enough to attribute an independent effect to issue salience (hypothesis 8 is supported).48It is concluded, therefore, that Model Four best fits the data. In this model, issue salience, which is conceptualized as being antecedent to perceptual accuracy, accounts for approximately twenty-five per cent of the explained variance (through both independent and developmental effects) and perceptual accuracy accounts for the remaining approximately seventy-five per cent (although the split for the marijuana issue is about fifty-fifty).4') With all the terms in the maximally predictive equation, the multiple correlation coefficients are quite similar for all attitudes. Direct Effects of the Parental Attitude. A comparison can now be made of the individual, direct effect that parental party identification and parental attitudes on race, China, and marijuana have on the same adolescent attitudes when issue salience and perceptual accuracy are held constant. This requires computation of the partial regression coefficient. Reported in Table 9 are the unstandardized coefficients (b's) which, according to Blalock, are generalizable across populations.50 Since the issues and party identification are measured with different scales, the standardized 48 In the case of the race issue, entering issue saliencc immediately after the parent attitude increases the correlation from.32 to.38. Entering issue salience after the parent attitude and perceptual accuracy (perceptual accuracy controlled) raises the correlation from.44 to.48. 48 The variables are entered in the order called for by the equation for hypothesis seven. ' Hubert M. Blalock, Jr., "Causal Inferences, Closed Populations, and Measures of Association," American Political Science Review, 60 (March, 1967), 130136. Table 9. B's and Beta Weights for the Effect of the Parent Attitude on the Adolescent Attitude with Issue Salience and Perceptual Accuracy Controlleda Beta B Std. Error.52.46.55.54.52.45.58.45. 10.09.09.04 Racial Integration China Policy Marijuana Laws Party Identification a Additive effects only. 1591 Table 10. Multiple Correlation with Attractiveness Added to the Equation for Hypothesis Seven Attitude Object Multiple Correlation Racial Integration China Policy Marijuana Laws.49 (+.01)Y.47 (+.02) (Not computed due to lack of linear progression within categories).57 (+.03) Party Identification a Increase in the multiple correlation coefficient when attractiveness is added to the equation for hypothesis 7. coefficients (beta weights) are also reported. These latter coefficients, which will be interpreted, represent the change that would occur in the adolescent attitude if the parent attitude were to change by one standard unit, with issue salience and perceptual accuracy controlled. The beta weights (see table 9) for the effect of the parent attitude on the adolescent attitude, with issue salience and perceptual accuracy held constant, are similar enough for all issues and party identification to warrant the assertion of a rough general explanation for the successful transmission of political attitudes from parent to child. Parental Attractiveness Finally, parental attractiveness can be added to the equation for hypothesis seven." Operating by itself, attractiveness was found to have moderately strong effects on two issues and party identification. Because of its inconsistent impact and because its proves to be only weakly related to either perceptual accuracy or issue salience,52 attractiveness was excluded from the model until the combined effects of issue salience and perceptual accuracy had been completed. Nevertheless, when added to the equation, it may explain variance to warrant subenough additional stantive interpretation. 3 3 3 Hypothesis 9: Y=ao+ai+ap+a+bcX 3 3 3 +(bi +bp+ba)X+e. With I and P held constant, parental influence is affected by attractiveness. The data reveal that, with a reservation, hypothe51 If the variable "attractiveness" is added to the equations for either hypothesis 7 or hypothesis 8, the results will be statistically identical. The correlations between attractiveness and perceptual accuracy for the attitudes toward racial integration, China policy, marijuana laws and party identification are.08,.09, -.05 and.05, respectively; for attractiveness and issue salience they are.01,.00,.12, and _.05, respectively. 52 This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions 1592 The American Political Science Review sis nine can be rejected. Attractiveness has no impact on the successful transmission of the attitude toward marijuana laws. The impact it has on attitudes toward the race issue, China policy and party identification`-while small-prevents dismissing this variable or others like it (nonissue specific) as being completely inconsequential.54 Conclusion In recent years, students of political socialization have downplayed the family as a source of attitudes toward public policy issues. This conclusion is not unwarranted given the data. But it would seem not to be an inherent characteristic of parent-child attitude transmission. Rather, quite the opposite appears to be the case. Parents have an inherent potential for successful transmission. Whether this potential is realized depends basically on the distribution within the population of (1) issue salience to the parent, and (2) adolescent perceptual accuracy of the parent attitude. These variables are specific to the individual attitude objects, as opposed to variables of the family interaction variety or "political interest" which are general to all attitude objects. Selected for analysis were attitude objects different enough to allow a fairly stringent test of any general thesis. One was party identification, to which some investigators have attributed unique transmissible characteristics. The others were substantive issues. One of the issues (marijuana laws) was deliberately chosen with the idea that it would be susceptible, to countervailing peer-group influence. But when issue salience and perceptual accuracy are held constant in a multiple covariance equa' It should be noted from Table 6 that there is a reversal for the within-category progression for party identification. The statistical model will make no adjustment for this fact. If we assume conceptually that the progression must be linear, then the model overestimates the impact of attractiveness on the transmission of party identification. "IIt should be kept in mind that what is being discussed is the independent effect of attractiveness on successful transmission al ter the effects of issue salience and perceptual accuracy have been partialed out. Vol. 68 tion, the beta weights indicating the success of transmission are very similar for all the attitudes that were studied. It may be tentatively concluded, therefore, that the influence of parents on the party identification and public policy attitudes of their children can. be explained by a general equation; and that the level of parental influence, at any one point in time, will be highly dependent on the distribution of issue salience and perceptual accuracy for the particular attitude object in question. Appendix OpinionItems MeasuringAttitudesTowardIssues Racial Integrationof the Public Schools 1. It is desirableto have raciallyintegratedschools in the United States. 2. The Governmentshould stay out of the questionof whether blacks and whites go to the same school. 3. In places where both blacks and whites live, the Governmentshould make sure that there are both black and white childrenin all the public schools. 4. Busing should be used to integrate the public schools. UnitedStates Policy towardCommunistChina 1. The United States is getting too friendly with CommunistChina. 2. The United States Governmentshould have more firmly opposed letting Communist China into the United Nations. 3. The United States should send ambassadors to CommunistChina and give the Chinesediplomatic recognition. 4. The United States should engage in greatly expanded trade with CommunistChina. Laws Regulatingthe Use of Marijuana 1. The Governmentis too easy on those caught with possession of marijuana. 2. The laws dealingwith those caught with possession of the marijuanaare too tough. 3. Anyonecaughtwith possessionof marijuanashould spend some time in jail. 4. The use of marijuanashould be made legal. This content downloaded from 128.146.120.57 on Thu, 21 Aug 2014 17:56:52 UTC All use subject to JSTOR Terms and Conditions