1 But see Fass, Paula S., Outside In: Minorities and the Transformation of American Education (New York, 1989).Google Scholar

2 Ueda, Reed, Avenues to Adulthood: The Origins of the High School and Social Mobility in an American Suburb (New York, 1987), 221. For a broadly pitched narrative on the nineteenth-century development of American high schools, see Reese, William J., The Origins of the American High School (New Haven, Conn., 1995).Google Scholar

3 Graebner, William, Coming of Age in Buffalo: Youth and Authority in the Postwar Era (Philadelphia, 1990). For an analysis of similar mechanisms as they are currently operative, see Eder, Donna, School Talk: Gender and Adolescent Culture (New Brunswick, N.J., 1995).Google Scholar

4 Labaree, David F., The Making of an American High School: The Credentials Market and the Central High School of Philadelphia, 1838–1939 (New Haven, Conn., 1988).Google Scholar

5 Gutowski, Thomas W., “Student Initiative and the Origins of the High School Extracurriculum: Chicago, 1880–1915,” History of Education Quarterly 28 (spring 1988): 72; Mirel, Jeffrey, “From Student Control to Institutional Control of High School Athletics: Three Michigan Cities, 1883–1905,” Journal of Social History 16 (1982): 83–100. Tyack, David and Hansot, Elisabeth, Learning Together: A History of Coeducation in American Schools (New Haven, Conn., 1990), 193.Google Scholar

6 Fass, , Outside In, 108. Stephen Lassonde's dissertation on New Haven's Italian American children during the same period reflects a similar sense of deeply differentiated ways of being in schools' formal and informal social groupings, although he additionally emphasizes how parents' ideas of what school was all about were often held in dialectical tension with that of their children. See Lassonde, Stephen A., “Learning to Forget: Schooling and Family Life in New Haven's Working Class, 1870–1940 (Connecticut)” (Ph.D. diss., Yale University, 1994). Graebner, William, “Outlawing Teenage Populism: The Campaign against Secret Societies in the American High School, 1900–1960,” Journal of American History 74 (Sept. 1987): 412.Google Scholar

7 Endacott, John Earl, “An Analysis of Friendship Selections and Other Social Relations and Activities of 103 Seniors in a Kansas High School” (master's thesis, University of Kansas, 1942). Our findings are of course subject to such systematic and random error that Endacott's work entailed. The meticulousness of Endacott's procedures is apparent from the thesis itself (including the exceptionally zealous reproduction of the raw data) and is consistent with the practice of his mentor, Mapheus Smith, a well-published sociologist, who exemplified the “dry-as-dust Midwest empiricism” common in the sociology of the day. The most obvious shortcoming of the data for social description is its restriction of nominated friendships to other Abilene High School seniors, a common problem of sociometric technique (which necessarily relies upon an a priori bounded set of eligibles). The patterns we discuss undoubtedly hold, but the parameters are no doubt occasionally mismeasured.Google Scholar

8 Endacott, quite appropriately, presented his data on each student with numbers in place of the names of individual students. But since his data were coterminous with the senior class and since the entries were, with a very few interruptions, in strictly alphabetical order, we could easily match them to the senior listings in the 1939 Abilene High yearbook (using gender and organizational participation to discern and identify nonsequential entries), and from there link to other nominal records. Our ability to research at a distance owes much to the timely assistance of Pat Aker (Abilene Public Library), Margaret McCaffrey (Carnegie Mellon University), William Neuenswander (Abilene Unified School District), Brian Pawley (Abilene High School), and William Tuttle (University of Kansas).Google Scholar

9 In fact, Mapheus Smith published his advisee's findings about friendship patterns in the journal Sociometry in 1944—without extending junior-author status to Endacott—although he cited the thesis. Smith, , “Some Factors in Friendship Selections of High School Students,” Sociometry 7 (Aug. 1944): 303–10.Google Scholar

10 Warner and his associates focused their research on several different towns in the United States. The first of these works cited below studied “Yankee City,” a large north-eastern town. The middle four seem to have studied the same small Midwestern city (though each gave it a different pseudonym). Throughout this paper we will refer to this midwestern town as “Jonesville.” The final of the works cited below studied “Deep South,” a southern town. In this paper we are careful to distinguish between these different sites in our discussions of Warner's work. Warner, W. Lloyd and Lunt, Paul S., The Social Life of a Modern Community (New Haven, Conn., 1941); Warner, W. Lloyd, Democracy in Jonesville: A Study of Quality and Inequality (New York, 1949); Neugarten, Bernice L., “Social Class and Friendship among School Children,” American Journal of Sociology 51 (Jan. 1946): 305–13; Junker, Buford H. and Loeb, Martin B., “The School and Social Structure in a Midwestern Community,” School Review 50 (Dec. 1942): 686–95; Hollingshead, August B., Elmtown's Youth: The Impact of Social Classes on Adolescents (New York, 1949); Gardner, Burleigh B., Gardner, Mary R., and Loeb, Martin B., “Social Status and Education in a Southern Community,” School Review 50 (Mar. 1942): 179–91.Google Scholar

11 Social historians have perhaps overly disdained the Warner school since Stephan Thernstrom's solid attack on its ahistoricism. Thernstrom, Stephan, Poverty and Progress: Social Mobility in a Nineteenth-Century City (Cambridge, Mass., 1964). Warner's distinctively non-Marxian social class is well (and critically) discussed in Centers, Richard, “Toward an Articulation of Two Approaches to Social Class Phenomena: I. Similarities and Differences in Theoretical Formulation, Method, and Results,” International Journal of Opinion and Attitude Research 4 (1950–51): 499–514.Google Scholar

12 Coleman, James S., The Adolescent Society: The Social Life of the Teenager and Its Impact on Education (New York, 1961).Google Scholar

13 Federal Writers' Project of the Works Progress Administration for the State of Kansas, Kansas: A Guide to the Sunflower State (New York, 1949), 354.Google Scholar

14 Osdol, Marvin Van, “A Brief History of the City of Abilene, Kansas, and a Survey of the Abilene Public School System” (master's thesis, Kansas State Teachers College of Emporia, 1941), 30–31. Bureau of the Census, Sixteenth Census of the United States, 1940: Population, vol. iv, Characteristics by Age, part 1 (Washington, D.C., 1943), 42, 54, 66; Bureau of the Census, Sixteenth Census of the United States, 1940: Characteristics of the Population, Kansas (Washington, D.C., 1943), table 11; Bureau of the Census, Historical Statistics of the United States: Colonial Times to 1970, part 1 (Washington, D.C., 1976), 379. For a detailed discussion of enrollment trends in Kansas high schools around 1940, see Hanson, Dale, “School Attendance of Kansas Youth Sixteen and Seventeen Years of Age” (master's thesis, Kansas State Teachers College of Emporia, 1948).Google Scholar

15 Osdol, Van, “Brief History,” 60–62; Berner, Loren Richard, “Curriculum Trends in One Hundred Kansas High Schools Between 1925 and 1935” (master's thesis, Kansas State College of Agriculture and Applied Science, 1937), 4–7. See also, Swanson, Frank Arvid, “The Organization and Content of the Curricula of Kansas High Schools” (master's thesis, Kansas State Agricultural College, 1928). For an excellent analysis of the rise of the women's vocational education movement, see Powers, Jane Bernard, The “Girl Question” in Education: Vocational Education for Young Women in the Progressive Era (London, 1992).Google Scholar

16 Rather than burden our paper with large numbers of tables, we present most of our findings discursively, seeking to maintain the flow of exposition and argument, and presuming that readers will recognize that we base our account of Abilene High, except where we indicate otherwise, upon our augmentation of Endacott's data set as described above. Skeptical or curious readers may request tabulations from the senior author. We omit tests of statistical significance because our data represent a complete census rather than a sample of the population to which we refer. Warner, , Democracy in Jonesville, 207.Google Scholar

17 Kefauver, Grayson N., Noll, Victor H., and Drake, C. Elwood, Secondary-School Population (Washington, D.C., 1933), 26.Google Scholar

18 Junker, and Loeb, , “School and Social Structure in a Midwestern Community.” We are precluded from making straightforward quantitative comparisons with Abilene by the lack there of the reputational criteria for social class so conscientiously employed by Warner and his followers. For the present paper, we employ an index of family socioeconomic status (SES) including as equally weighted components: occupational level of household head, credit rating of household head, and (a whiff of Warner) family membership in a small set of relatively prestigious local organizations. This segmentation pattern has repeated itself widely, not just in the United States but almost everywhere in the world where schooling at a given level has become universalized. Success in achieving graduation at that level temporarily becomes a function of socioeconomic background, but once graduation becomes effectively universal, then entry to the next level of schooling becomes the socioeconomically discriminated gateway to adult placement. See Mare, Robert D., “Change and Stability in Educational Stratification,” American Sociological Review 46 (Feb. 1981): 72–87; Shavit, Yossi and Blossfeld, Hans-Peter, Persistent Inequality: Changing Educational Attainment in Thirteen Countries (Boulder, Colo., 1993).Google Scholar

19 This pattern was effectively identical to that found by Warner and Lunt in Yankee City despite curricular choices that differed considerably in detail. See Warner, and Lunt, , The Social Life of a Modern Community, 364. For evidence from a few years earlier, see a survey covering a number of high schools scattered through the country: Kefauver, Grayson N., Noll, Victor H., and Drake, C. Elwood, Horizontal Organization of Secondary Education (Washington, D.C., 1934), ch. 6.Google Scholar

20 This fits with Hansot and Tyack's discussion of coeducation in this period. They suggest that while there was much “policy talk” about separating the sexes to adequately prepare them for different lives, participation in the segregated curricula was minimal, and even those who opted-in usually took many courses with those in other curricula as well. See Hansot, and Tyack, , Learning Together, chs. 7–8. See Berner, , “Curriculum Trends in One Hundred Kansas High Schools between 1925 and 1935,” 4–10.Google Scholar

21 Booster, 12 May 1938. See especially Sloan, Clair M., “A Survey of the Practices in the Out-of-Class Activity Program of Thirty High Schools in Kansas With Enrollments of Varying Sizes” (master's thesis, Kansas State Teachers College of Emporia, 1940). See also Caldwell, Bessie Ellis, “The Socialization of Extracurricular Activities in Lincoln High School, Salina, Kansas” (master's thesis, Kansas State Teachers College, 1949); Nation, Marion Allen, “The Objectives and Values of Clubs in the Junior High Schools of Southeast Kansas” (master's thesis, Kansas State Teachers College, 1932); Jennings, Alice Marie, “The Out-of-School Interests and Activities of High School Girls in Rural and Small Town Communities in Kansas” (master's thesis, Kansas State College of Agriculture and Applied Science, 1936).Google Scholar

22 Joiners at school were most likely also to be joiners outside the school; in fact, not even one student who belonged to an extraschool organization belonged to none within the school. It should come as no surprise that participation in school activities was more closely correlated with student popularity than was participation in extraschool activities.Google Scholar

23 We created a measure of students' attractiveness by asking eight people of differing age (none of high-school age, however) and gender (but uniformly middle-class) to sort the senior pictures of the Abilene High class of 1939 (which we had mounted one-by-one with no identification whatever) into five piles, ranked from most attractive to least attractive. We then determined “looks” by taking an average of the z-scores (each, that is, standardized by the rater's mean for all students of that sex). Average rater-to-rater correlations (Pearsonian) were 0.53 for boys' looks and 0.45 for girls' looks. Changing standards of beauty, different class and cultural perceptions of beauty, and the relative homogeneity of our judges surely make this a very crude variable. Nevertheless, , “looks” does not simply measure other attributes indirectly: for boys, correlations of “looks” with age was only 0.12 (younger seniors judged handsomer) and with family SES was but 0.08; for girls, the figures were 0.01 and 0.18. Overall, the variable's salience at every stage in this analysis convinced us that it captures a meaningful dimension of life in Abilene High in 1939.Google Scholar

24 Factor analysis is a common, bluntly empirical procedure that examines the covariance of a set of variables to discern possible “latent dimensions,” called factors, among them. A given factor includes maximally intercorrelated variables and is itself largely uncorrelated with other factors. It is the analyst's task to name the latent dimension in view of the variables that it incorporates. Endacott's questionnaire and the 1939 Abilene High Yearbook gave us separate listings of organizational membership of the students. For the most part, if one listed a given student as a member of a given organization, so did the other, but there were exceptions of both sorts. In our analysis, we treated each listing as independent, reasoning that those listed in only one source were perhaps more tenuously affiliated than those listed in both, but were nevertheless more affiliated than those listed in neither source. (Trial tabulations, including a dichotomous ever-mentioned/never-mentioned distinction, yielded virtually identical correlational results.) Most of the factors are self-explanatory. Somewhat less obvious are boys' “intellectual” organizations, which included the high school annual, the junior and senior play, the school newspaper, and debate; and girls' “school spirit” organizations, which included band, orchestra, mixed chorus, and cheerleaders.Google Scholar

25 Residential proximity helps give us a sense of the mechanisms at work. Even in a small city like Abilene, and despite having for six years attended schools that served the whole city, senior boys and girls both chose their friends from those who lived relatively close to themselves. The average high school student living in the city dwelled just about ten blocks from his or her average classmate, but two blocks closer to his or her average friend. That on average friends lived closer reflected the joint operation of the maintenance of neighboring friends from early childhood, the tendency to choose friends from like socioeconomic level (often indexed by neighborhood), and simple convenience. On the other hand, seniors—especially girls—who chose many friends chose them from further away. This pattern was truest for best friends, indicating that popular, confident children chose their best friends especially selectively, with least regard thus to residential proximity. (Fifty-one students lived in the city and gave addresses. We employed a six-by-six grid on the city and used the Pythagorean theorem to determine how far every student's home was from every other student's home.) Google Scholar

26 Ten boys out of forty-nine were chosen as a top friend by one or more girls, most of these by only one, but one by three girls. Twelve of fifty-four girls were chosen as a top friend by at least one boy, only one of these by more than one boy. Overall, although the pattern is not very pronounced (for instance, the boy chosen by three girls was chosen by none of the boys), children popular among their own gender are relatively popular with the other gender, too. In high schools in upstate New York at about the same time, however, girls were less often named as friends in a sociometric inquiry than were boys. Kuhlen, Raymond G. and Lee, Beatrice J., “Personality Characteristics and Social Acceptability in Adolescence,” Journal of Educational Psychology 34 (Sept. 1943): 321–40.Google Scholar

27 Intriguingly, membership in school organizations represented, overall, only roughly about as strong a basis for friendship choice as did parental organizational participation. The procedure follows that of Neugarten, , “Social Class and Friendship among School Children.” See also Warner, , Democracy in Jonesville, ch. 5. The rare exceptions were mainly best friendships, not lesser friendships, a finding that reminds us of the lessened significance of residential proximity among best friends.Google Scholar

28 The observation here holds, whether one is tabulating just first friend choices in Abilene, or all choices. The question in Jonesville asked high schoolers to name “the boy or girl I would choose … if I could have anybody I wanted for a friend.” Neugarten, , “Social Class and Friendship among School Children,” 306. See also Warner, , Democracy in Jonesville, ch. 5. Hollingshead, , Elmtown's Youth, 216.Google Scholar

29 There was a minute tendency for the girls in the secretarial track to choose one another as friends, but no tendency whatever for students in the general curriculum to do so. This perhaps reflects the frequent movement permitted between the non-professional curricula: students without a long-term curricular identification would not be expected to favor as friends those in their own curriculum.Google Scholar

30 We owe deep thanks to Professor Kathleen Carley of the Carnegie Mellon University Department of Social and Decision Sciences, an especially thoughtful and especially generous sociological student of social networks, for her instruction, computer assistance, and enthusiastic discussion of our ongoing work on this paper.Google Scholar

31 Warner, , Democracy in Jonesville, 208–10. Warner also notes a parallel tendency in Deep South for high school grades to be “highly influenced by the relative social status of the pupils' families.” Warner, W. Lloyd, “Social Anthropology and the Modern Community,” American Journal of Sociology 46 (May 1941): 790, citing Warner, W. Lloyd and Davis, W. Allison, “A Comparative Study of American Caste,” in Race Relations and the Race Problem: A Definition and an Analysis , ed. Thompson, Edgar T. (Durham. N.C., 1939), 219–40.Google Scholar

32 In fact, of course, most boys would have these somewhat determinate career trajectories re-opened by military service. See Modell, John, “Did the Good War Make Good Workers?” in The Home-Front War: World War II and American Society, ed. O'Brien, Kenneth Paul and Parsons, Lynn Hudson (Westport, Conn., 1995), 139–55.Google Scholar