¹ The first thoroughgoing consideration of the problems in studying political socialization, including a summary of much past research which is germane to understanding political develop ment, is Hyman's, Herbert Political Socialization (Glencoe, 1959).Google Scholar Emphasis on the socialization process as a determinant of adult behavior does not, of course, carry the implication that “only … childhood events determine adult behavior.” cf. Leites, Nathan, “Psycho-Cultural Hypotheses about Political Acts,” World Politics, Vol. 1 (1948), pp. 102–19.CrossRefGoogle Scholar

² For a full report of the larger study from which these findings are drawn, along with a more detailed discussion of its methodology, see Greenstein, Fred I., Children's Political Perspectives: A Study of the Development of Political Awareness and Preferences among Pre-adolescents, unpublished doctoral dissertation, Yale University Library.Google Scholar The major methodological point which should be noted here is that the respondents consist of the populations of three public and one private school, rather than a probability sample of New Haven children. As a rough means of compensating for this difficulty, socio-economic status was controlled in the data analysis. However, since SES differences are not relevant to the present discussion, data for all respondents are presented combined.

When data are reported by age, systematic variations between the younger and older age groups are treated as evidence of developmental changes. It is always possible, however, in a crosssectional study that some other variable is responsible for apparent age differences. Therefore, further validation of statements about development by longitudinal procedures is desirable. For a discussion of the relative merits of cross-sectional and longitudinal research, see the articles by Anderson, John B. and Mead, Margaret in Carmichael, Leonard, Manual of Child Psychology (New York 1954).Google Scholar

I am deeply indebted to Robert E. Lane for his thoughtful criticism at every stage of this project. Earlier drafts of this paper profited from comments by James David Barber, Harold D. Lasswell, Richard D. Schwartz, and Jerome H. Skolnick.

³ See “Jobs and Occupations: A Popular Evaluation,” Opinion News, Vol. 9 (September 1, 1949), pp. 3–19; Smith, Mapheus, “An Empirical Scale of Prestige Status of Occupations,” American Sociological Review, Vol. 8 (1943), pp. 185–92CrossRefGoogle Scholar; Lehman, Harvey C. and Witty, Paul A., “Further Study of the Social Status of Occupations,” The Journal of Educational Sociology, Vol. 5 (1931–1932)Google Scholar; Hartmann, George W., “The Prestige of Occupations,” The Personnel Journal, Vol. 13 (1934–1935), pp. 144–52.Google Scholar For an early attempt at cross-cultural study, see Davis, Jerome, “Testing the Social Attitudes of Children in the Government Schools in Russia,” American Journal of Sociology, Vol. 32 (1926–1927), pp. 947–52.CrossRefGoogle Scholar

⁴ The variations in Truman's first-term popularity are summarized in Berelson, Bernard et al., Voting (Chicago, 1954), p. 258.Google Scholar Public responses to Roosevelt may be found in Cantril, Hadley and Strunk, Mildred, Public Opinion 1985–1946 (Princeton, 1951).Google Scholar For Eisenhower's popularity see the regular news releases of the AIPO.

⁵ Cantril, Hadley, Gauging Public Opinion (Princeton, 1944), pp. 43–44 CrossRefGoogle Scholar; Lazarsfeld, Paul F. et al., The People's Choice (New York, 1944), p. 38 Google Scholar; Lazarsfeld, Paul F., “Public Opinion and the Classical Tradition,” The Public Opinion Quarterly, Vol. 21 (1957), p. 47.CrossRefGoogle Scholar

⁶ Cantril and Strunk, op. cit., p. 584; National Opinion Research Center, The Public Looks at Politics and Politicians, Report No. 20, March 1944.Google Scholar

⁷ Campbell, Angus et al., The Voter Decides (Evanston, 1954), pp. 187–94.Google Scholar

⁸ Mitchell, William C., in the Western Political Quarterly, Vol. 12 (1959), pp. 683–98.CrossRefGoogle Scholar

⁹ This is not to say that children never develop aversions to individual political leaders, although as is shown below, they probably are less likely than adults to do so. One item which did evoke spontaneous negative references to individual politicians asked the child to “Name a famous person you don't want to be like.” About 15 percent of the respondents named a political leader of the present or of the recent past. Interestingly, more than half of these references were to foreign leaders such as Nikita Khrushchev. Follow-up interviews suggested that some of the negative references to domestic leaders were rejections of the responsibility of being (say) president, rather than personal criticisms of the leader.

¹⁰ Moreover, they expressed the party identification which was “appropriate” for their demographic group. The children's reports of familial party preferences support the familiar thesis that membership in American political parties is “inherited.” How party identification develops and the relationship between this process and the development of other orientations which affect adult electoral choice will be discussed in a subsequent publication.

¹¹ The report of these findings in Table III shows not only that ratings of the three incumbent political executives were exceptionally favorable, but also that few children failed to express an opinion. In fact, two-thirds of the sample evaluated all three leaders, although, as can be seen from Table II, many fewer were able to describe the duties of each of these individuals. This raises the possibility that responses to the evaluation item may be “invalid”, as a result, for example, of perseveration. I do not believe that this is so for the following reasons:

(1) Some variations in ratings of the three leaders emerged, within the context of a general willingness to (a) evaluate and (b) make a highly positive assessment. For example, “don't know” was checked with respect to the lesser known governor more often than with respect to the mayor and president.

(2) Seventh and eighth grade children evaluated a non-incumbent political leader, Adlai Stevenson, as well as the three incumbents. Stevenson's ratings were much less favorable than Eisenhower's (with a modal response of “fairly good”).

(3) Within the general tendency to produce favorable ratings, differences between subgroups in the sample were evident. Children with Democratic party identifications were significantly more likely to give high ratings to the Democratic Mayor and Governor, and Republicans were significantly more favorable toward Eisenhower.

(4) The possibility of checking “don't know” was emphasized in the administration instructions, and this alternative appeared before the other possibilities in the questionnaire. In addition, each evaluation item was separated from the next by a series of intermediary items. The anonymity of questionnaire responses was stressed.

Rather than being an artifact of the questionnaire, I suspect that the responses reported in Table III are further evidence of children's early propensity to make political judgments in the absence of much information. cf. Babchuck, Nicholas and Gordon, C. Wayne, “The Child as a Prototype of the Naive Informant in the Interview Situation,” American Sociological Review, Vol. 23 (1958), pp. 196–98.CrossRefGoogle Scholar

¹² At least three points of non-comparability should be noted: the differences in population; the differences in question wording; and the differences in response alternatives. With these major reservations in mind, the following additional AIPO findings on President Eisenhower, for the months surrounding this survey, may be noted. (AIPO releases November 1957, January and March 1958).

The November AIPO report, which controls for region, suggests that Southern responses may have inflated the “disapproval” column. In that month, Eastern states were reported as 65% approve, 21% disapprove, and 14% no-opinion. This still is considerably more than the 5% of the New Haven children who checked the “bad” and “not very good” alternatives.

¹³ For example, in the Survey Research Center's 1952 election study sample, only a fifth of the respondents claimed that “it would make a great deal of difference to the country whether the Democrats or the Republicans win the elections.” Campbell, Angus et al., The Voter Decides (Evanston, 1954), p. 38.Google Scholar The same reservations about comparability of samples, question wording, and response alternatives expressed in note 12, above, are also relevant here.

The Survey Research Center reports that 43 percent of the respondents replied “some difference, minor differences.” Compare a 1946 crosssection of the national population which was asked: “Do you think it makes much difference or only a little difference which party wins the elections for Congress this fall?” About half said it made “much difference,” about 30 percent “little,” and the remaining 20 percent was divided evenly between “no difference” and “no opinion.” Cantril and Strunk, op. cit., p. 582.

¹⁴ Another class of political imagery is worth noting. A small percentage of the respondents (but some in each of the four schools which make up the sample) coped with the problem of organizing their fragmentary political information by using hierarchical concepts. They saw politics, even in the case of individuals and institutions which are formally co-ordinate, in terms of a chain of command. For example:

The Mayor gets ordere from the President … The President Rives orders to the Governor. (Fifth grade girl)

[The Mayor] handles the minor problems and if it is too big he goes to the Governor. (Seventh grade boy)

[The Mayorl takes the problems of the town. He [the Governor] takes the hard problems. (Fifth grade boy)

One notably ingenious misconception combined hierarchy with benevolence. The fifth grade boy whose statement that “The Mayor pays working people like banks” was cited above went on to say, “The Governor pays Mayors…. The President pays the governor.” Again the appropriate instrument might reveal that this way of perceiving politics is reasonably common in childhood and, perhaps, that it is related to certain types of nonpolitical—e.g., primary group—experiences. See footnote 21, below.

¹⁵ For a useful summary of this literature, see Child, Irvin L., “Socialization,” in Lindzey, Gardner, Handbook of Social Psychology (Cambridge, 1954), pp. 655–92.Google Scholar Also Alex Inkeles and Daniel J. Levinson, “National Character: the Study of Modal Personality and Sociocultural Systems,” ibid., especially pp. 997–98.

¹⁶ cf. Hyman, op. cit., ch. 4.

¹⁷ cf. Greenstein, op. cit., pp. 55–74.

¹⁸ Contemporary children's literature, which presents a painfully benevolent portrayal of the wider environment, would tend to reinforce such a frame of reference. Books such as Our Friend the Farmer and How the Policeman Helps Us are couched in language which closely resembles some of the pre-adolescent descriptions of political leaders reported above.

¹⁹ This thesis is advanced in Lasswell, Harold D., Power and Personality (New York, 1948), pp. 156–57.Google Scholar Also see Wahl, C. W., “The Relation Between Primary and Secondary Identifications,” in Burdick, Eugene and Brodbeck, Arthur J., American Voting Behavior (Glencoe, 1959), pp. 262–80.Google Scholar

²⁰ DeGrazia, Sebastian, “A Note on the Psychological Position of the Chief Executive,” Psychiatry, Vol. 8 (1945), pp. 267–72CrossRefGoogle Scholar; Fairbairn, W. R. D., An Object Relations Theory of Personality (New York, 1954), pp. 223–29Google Scholar; Franz Alexander, “Emotional Factors in Voting Behavior,” in Burdick and Brodbeck, op. cit., pp. 300–307; Richard E. Renneker, “Some Psychodynamic Aspects of Voting Behavior,” ibid., pp. 399–413.

²¹ The hierarchical imagery reported in footnote 14 provides an illustration of the types of linkage which may take place. In each instance the child clearly has meager factual information about the roles he is attempting to describe. His allusion to a chain of command seems to result from an at tempt to organize what fragmentary information he has. This may be seen in further detail in a follow-up interview with a sixth grade boy who had responded hierarchically on his question naire:

Interviewer: What sorts of things does Mayor Lee do?

Larry: Well, he keeps the city together and tells them what to do.

I: Tells who what to do?

L: Well, let's see. He probably tells some of the most important people and then they tell the ones that are less important and they keep on going.

I: What kinds of people are these? Do you mean any people in the city?

L: No, like the police chief and the head of the schools or something.

……‥

I: What do you think the Governor does?

L: Well, he probably tells the mayors what to do and he probably tells other people what to do.

I: When you say “probably,” does that mean you're sort of guessing about what he does?

L: Yes … He has to tell somebody what to do.

……‥

I: President Eisenhower, what does he do?

L: He probably tells the governors what to do.

I: Can you think of any other things that he might do?

L: He probably tells the people that are going to foreign countries to present a thing or something he wants that if they might be interested in. (sic)

I: What kind of a thing do you mean?

L: Well, like a peace treaty or something.

That Larry turns “automatically” to hierarchy as an organizing principle, suggests that his pre-political experiences have taught him to view the world in terms of “telling people what to do.” One might hypothesize, for example, that children responding in this way are more likely to come from families in which parental authority follows similar patterns.

²² Child, op. cit., p. 678.

²³ Angus Campbell and Donald E. Stokes “Partisan Attitudes and the Presidential Vote,” Burdick and Brodbeck, op. cit., p. 357.

²⁴ The sequence of development of certain classes of political response in the New Haven sample is summarized in Greenstein, op. cit., ch. 11.

²⁵ Projective devices of the Thematic Apperception variety are promising techniques. For a novel projective procedure, which has received preliminary application, see Kennedy, John L. and Lasswell, Harold D., “A Cross-Cultural Test of Self-Image,” Human Organization, Vol. 17 (1958), pp. 41–43.CrossRefGoogle Scholar

²⁶ A bibliography of portions of this study which have been published is contained in Anderson, Harold H. et al., “Image of the Teacher by Adolescent Children in Four Countries: Germany, England, Mexico, United States,” Journal of Social Psychology, Vol. 50 (1959), pp. 47–55.CrossRefGoogle Scholar The Andersons have kindly made available to me unpublished portions of their data.

²⁷ For a guide to ongoing research, see the regular reports of the Children's Bureau in the bulletin Research Relating to Children. A bibliogra phy of longitudinal studies has been assembled by Stone, A. A. and Onque, Gloria C., Longitudinal Studies of Child Personality: Abstracts with Index (Cambridge, 1959).CrossRefGoogle Scholar

²⁸ A logical first step would be to extend cross-sectional sampling of various age groups beyond the age-range of the present study. For an inter esting attempt to use survey data collected over a period of several years to simulate developmental analysis, see Evan, William M., “Cohort Analysis of Survey Data,” The Public Opinion Quarterly, Vol. 23 (1959), pp. 63–72.CrossRefGoogle Scholar

Lasswell has proposed an ingenious procedure for combining the convenience of cross-sectional research with the merits of studies following the same individuals over time. Lasswell, Harold D., “The Method of Interlapping Observations in the Study of Personality and Culture,” Journal of Abnormal and Social Psychology, Vol. 32 (1937), pp. 240–45.CrossRefGoogle Scholar In addition, the systematic examination of adult recollections, expecially under clinical conditions (see note 21) has barely begun.

²⁹ Criteria for the Life History (New Haven, 1935), p. 27.