Transcript
Page 1: Interracial Contact in High School Extracurricular Activities

25

0042-0972/02/0300-0025/0 � 2002 Human Sciences Press, Inc.

The Urban Review, Vol. 34, No. 1, March 2002 (� 2002)

Interracial Contact in High SchoolExtracurricular Activities1

Charles T. Clotfelter

Using data from yearbooks for 193 high schools, this study examines the degree ofinterracial contact in 8,849 high school teams and other organizations. More than one-third of these groups were all-white, while only about 3% were exclusively nonwhite.Owing in large part to their overall numerical preponderance, white students rarelyfound themselves outnumbered in groups by as much as three to one; by contrast,nonwhites often were in this position. Tabulations show that the degree of interracialexposure was typically less than what would occur if all organizations in each schoolhad been racially balanced and was much less than the exposure that would haveoccurred if all organizations reflected the racial composition of the schools containingthem. Whereas the nonwhite percentage of the students enrolled in the sample highschools was 24.9%, the membership of clubs and teams was 20.7%, reflecting a lowerrate of participation by nonwhites. Furthermore, because the racial compositions ofclubs and teams were not uniform, the average white member was in an organizationthat was only 15.3% nonwhite. Although clearly less than its theoretical maximum, thisrate of contact nonetheless appears to be much higher than what would occur if friend-ships were the only vehicle for interracial contact outside the classroom. Finally, theextent of segregation associated with these organizations was the same or less in theSouth than in the rest of the country.

KEY WORDS: segregation; interracial contact; high school.

Contact among students of different racial and ethnic groups in schools re-mains an issue of profound social importance in this country. Despite the end oflegal public school segregation in 1954 and the halting but nonetheless markedincreases in school integration that occurred in the following two decades, thetopic remains a concern of policymakers and continues to receive considerableattention from scholars. As discussed in the social science literature, interracialcontact refers to the presence of members of different racial or ethnic groups inthe same small unit, such as a school, and is measured by exposure rates orsegregation indices.2 For example, an exposure rate of whites to nonwhites thatis calculated from school-level data is interpreted as the proportion nonwhite in

Address correspondence to Charles T. Clotfelter, Box 90245, Duke University, Durham, NC27708-0245; Charles.Clotfelter�Duke.edu.

Page 2: Interracial Contact in High School Extracurricular Activities

26 THE URBAN REVIEW

the average white student’s school.3 Similar measures could also be calculatedwith classrooms, or city blocks, as the unit of analysis. Measures of contactsuch as these by no means imply friendship or even meaningful interaction,although contact appears to be a prerequisite for such relationships. The pur-pose of this paper is to document one form of interracial contact heretoforevirtually unexamined: that occurring in high school clubs and sports teams.4

This paper uses data on the racial makeup of school groups gathered byexamining a recent vintage of yearbooks for 193 public and private highschools. Information on participation and interracial exposure is calculated byschool and type of organization. The paper is primarily descriptive. It does notattempt to explore the causes of observed segregation. It seeks to measure nei-ther the depth nor the durability of the friendships that may arise from suchcontact. Nor does it assess other possible consequences of interracial contact, asimportant as those issues are. By focusing only on organizational membership,the paper does not attempt to distinguish the strength of friendship ties, thedegree of reciprocity involved, or, indeed, the degree to which the patterns arevoluntary.5 Instead, the aim of the paper is to document and measure the extentof interracial contact in these organized groups, a potentially important topicabout which little empirical analysis has previously been undertaken.

The first section of the paper provides background and motivation for thepaper, highlighting some questions that invite study. The second section offers abrief summary of relevant social science research, noting the connection thathas been made between contact and racial attitudes. The next section discussesthe methods and data employed in the paper. The fourth section presents find-ings showing patterns of interracial contact based on high schools in the sam-ple, and the last section discusses criteria for evaluating those findings andgives brief conclusions.

BACKGROUND

Few areas of social policy have experienced the kind of dramatic change thathas occurred in the extent of interracial contact in the nation’s public schools.One measure of the change wrought by school desegregation is the markeddecline in the percentage of black students in the public schools who attendedschools that were all or virtually all minority in racial composition. Between1968 and 1988, the percentage of black students in the United States who at-tended public schools with minority enrollments between 90% and 100%dropped by half, from 64% to 32%. Among the nation’s regions, this percent-age declined the most in the South. In fact, by 1988, the South had the smallestpercentage of black students attending such schools, 24%, while the Northeasthad the highest, 48% (Orfield, 1983, Table 2, p. 4; Orfield & Monfort, 1992,Table 8, p. 6). Although some slippage in interracial contact has subsequently

Page 3: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 27

occurred across the board, these regional patterns appear to have remainedmore or less constant into the 1990s.6 Much of the remaining segregation can beattributed to the large disparities in racial composition that exist across districtsin many states and metropolitan areas, rather than to segregation among schoolswithin public school districts.7

Despite these impressive reductions in segregation between schools, severalfactors may tend to limit the actual interracial contact within schools. One ofthese is academic tracking, the policy of dividing the students in some aca-demic subjects into separate classes offering instruction at different levels. Ifstudents of different racial groups are disproportionately assigned to differenttracks, interracial contact will necessarily be less than it would have been withrandom assignments. The racial disparities that result from such assignmentshave been one argument against tracking.8 A second force that would tend toreduce interracial contact in high schools would be a tendency toward self-segregation in students’ choices of friends. If students, like most adults, tend togravitate toward those similar to themselves, school racial compositions wouldoverstate actual exposure rates. A third factor influencing actual interracial con-tact in high schools is a largely overlooked vehicle for interaction: the teams,clubs, and other groups associated with extracurricular activities. Extracurricu-lar activities play a significant role in the high school experience, as illustratedby the fact that over half of all high school students participate in athletic teamsalone (McNeal, 1998, p. 187).

The purpose of this paper is to examine the last of these three mechanisms,to document the extent of actual contact among students of different racial andethnic groups on sports teams and in other student organizations. Casual obser-vation of such features as segregated cafeterias and nearly all-white soccer orgolf teams would suggest that there is less interracial contact within schoolsthan schoolwide enrollment figures would suggest, but there is little formalresearch to document those impressions. Among the questions that might beasked about interracial contact in high school extracurricular activities, proba-bly the most significant one is simply how much contact exists. To answer thisquestion, the paper presents measures of racial composition and interracial ex-posure that have been commonly used in studies of racial segregation. Relatedquestions of interest include how interracial contact differs across regions, be-tween public and private schools, and among types of organizations; how thiscontact compares to schools’ overall racial composition; whether students ofdifferent racial and ethnic groups participate in extracurricular activities at thesame rates; how many students of another race a typical student associates within organizations; and whether there exist racial thresholds beyond which stu-dents are reluctant to participate in an organization.

Page 4: Interracial Contact in High School Extracurricular Activities

28 THE URBAN REVIEW

PREVIOUS RESEARCH

Two strands of social science research provide relevant background to thisinvestigation of extracurricular activities. First, research on interracial contactwithin schools offers a benchmark for assessing the relative importance of mea-sured differences in contact. Second, research on the effects of contact on atti-tudes and friendships suggests the social significance of the current investiga-tion.

The Extent of Interracial Contact Within Schools

In contrast to the extensive research on segregation between schools, rela-tively few studies have measured the interracial contact within schools. Perhapsthe most comprehensive attempt to measure within-school contact is Morganand McPartland’s (1981) examination of classroom assignments in 43,738 pub-lic schools in the fall of 1976. They found a small degree of intraschool seg-regation in elementary and middle school grades and a more pronounced degreein high schools. They found such segregation to be more extreme in the Souththan elsewhere, and more in schools with approximately equal numbers ofblacks and whites than in those that were predominantly white or black. Theirfindings suggest that in-school segregation had the effect of reducing the inter-racial exposure rate by about 11% on average below what it would have beenhad classrooms been racially balanced.9 More recently, Oakes (1994) and Mick-elson (2001) have analyzed the effects of academic tracking on the racial com-position of individual classes. In particular, Mickelson’s study of high schoolsreveals a marked degree of segregation in some courses. This finding impliesagain that within-school exposure rates are lower than corresponding schoolracial compositions, but her calculations do not make it possible to say exactlyhow much the overall exposure rate is reduced.

Turning from classroom assignments to friendship choices gives another per-spective on interracial contact within schools. Hallinan and Williams (1989)used students’ reports from the High School and Beyond sample to determinewho their best friends were in school. They found that same-race friendshipswere much more common than those between blacks and whites. Among soph-omores, same-race friendships were six times as common as cross-race ones;among seniors the ratio was 5.6 (p. 74). Joyner and Kao (2000, pp. 818–819)examined similar questions put to students in Grades 7–12 in the NationalLongitudinal Study of Adolescent Health, focusing on each respondent’s first-named friend. Of those reporting a best friend of the same sex, they foundwhites, whose average school was 73% white, were much more likely to namea white as their best friend; males were 7.3 times as likely to name a same-racefriend, and females were 10.1 times. Blacks, who attended schools that were on

Page 5: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 29

average 44% black, had comparable ratios of 3.8 and 5.7. At least on the basisof this criterion for friendship, therefore, students appear to give themselveslittle opportunity, outside class and other organized school activities, to associ-ate with members of other racial groups.

Effects of Interracial Contact on Attitudes

A question of long-standing interest has been the extent to which interracialcontact may lead to more tolerant racial attitudes. The classic presentation ofcontact theory, and the conditions under which contact can lead to reducedprejudice, are given by Allport (1954). As interpreted by St. John (1975, p. 85)for the case of school desegregation, contact between ethnic groups can lead toreduced prejudice if the contact is prolonged, if it is between equals who arepursuing a common goal, and if it is sanctioned by authorities. All of theseconditions would appear to be met by active school-sponsored organizations,especially school teams. Those who have examined the role of extracurricularorganizations in the personal development of adolescents appear to agree on theimportance of interracial contact, though views diverge on exactly what socialmechanisms are most important. From one perspective, extracurricular activitiesare valuable to the extent that they form a bridge to interracial friendships.Hallinan and Williams (1989, p. 68) emphasize the importance of groups forfostering friendships among students:

Since interaction, whether by chance or by choice, generally leads to positive senti-ment, students who are assigned or choose to belong to the same instructional groupsor participate in joint co-curricular and extracurricular activities are more likely tobecome friends than those who are in different groups.10

To observers such as Ellison and Powers (1994), contact is important chiefly tothe extent that it leads to friendships, which in turn are central to positiveeffects on attitudes.

A contrasting perspective on the importance of contact in extracurricularorganizations holds that the “weak ties” built in such groups are more importantto interracial relations because of their role as a bridge between the largelyhomogeneous social circles of school. Not only are cross-race friendships rela-tively rare, their presence may jeopardize same-race friendships. Mere acquain-tances have the virtue of linking otherwise separate groups (Granovetter, 1986,p. 87). Because they foster such weak ties, extracurricular organizations take onspecial significance:

In junior high and high school settings, extracurricular activities become a particu-lar focus of out-of-class student interaction, and may take on greater significance formany students than the formal educational process. (Granovetter, 1986, p. 83)

Page 6: Interracial Contact in High School Extracurricular Activities

30 THE URBAN REVIEW

Although they differ in the importance they attach to friendships, both of theseperspectives support the commonsensical proposition that, other things beingequal, more interracial contact is better than less, in the sense of underminingnegative stereotypes and fostering positive racial attitudes. Thus, the kind ofinformal contact provided by teams and other extracurricular organizationsholds the promise for attitudes conducive to tolerance and cooperation amongmembers of different racial groups. Interracial contact may also have an impor-tant role in integrating students into the social networks that play a large role ingenerating subsequent job opportunities.11

As a vehicle for interracial contact, clubs and sports teams exist on a contin-uum somewhere between the externally imposed grouping of classroom assign-ments and the purely voluntary grouping created by friendships. Membership insome high school organizations—such as competitive sports teams and per-forming-arts groups subject to tryouts—necessarily depends on skill and effort.Team membership may also be conditioned on grades and could as well, insome instances, be influenced by discriminatory practices of selection. By con-trast, membership in many high school clubs is almost purely a matter of self-selection. Still other organizations, such as student government or academichonor societies, rely on voting and other formal selection procedures.12 In judg-ing their importance for fostering interracial contact, however, the distinctionsamong these selection criteria seem less important than the resulting patterns ofmembership themselves.

DATA AND MEASUREMENT

The principal data for the study are derived from a sample of 1997–1998high school yearbooks taken from a collection of sample copies produced byJostens, Inc., a large yearbook publisher in Winston-Salem, North Carolina.Limited only by the space available to transport them and the time required toexamine them, research assistants took for study about 200 of these samplecopies, making no effort to choose particular books or to balance the sample inany way. After elimination of yearbooks for elementary and middle schools,there remained 193 books for schools that covered high school grades. Becauseit was born of a unique opportunity to obtain a large number of books, and notfrom a random sample of all high schools, the sample and the averages calcu-lated from it cannot be taken as representative of all American high schools. Byregion, owing undoubtedly to the location of the publisher, the sample ofschools covered is heavily weighted toward the South, East, and Midwest.13 Bytype of school, the percentage of schools in the sample that are private is closeto that of all high schools in the nation, but the average size of the sampleschools is larger than average.14 Slightly more than half (14) of the privateschools in the sample are Roman Catholic, roughly in line with the percentageof private secondary schools in the nation that are Catholic.15

Page 7: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 31

Data for the study were obtained from a detailed examination of these year-books. For every sports team or other organization having a formal group pic-ture in the yearbook, a count was made of the white, black, Hispanic, and otherstudents. Ideally, the classification of students into racial and ethnic categorieswould be done by self-identification, the method used today in most socialsurveys and public documents, but that approach was obviously infeasible inthis case. Thus, classifications had to be based on visual inspection, using bothfacial appearance and names to indicate ethnicity. This approach is not whollysatisfactory, both because it is subject to error and because it raises the thornyquestion of the very definition and validity of racial categories. Still, most peo-ple in contemporary America do commonly rely on such visual categorization.Although some ambiguity existed about the proper categories for a few of thefaces pictured, readers generally had little difficulty classifying the vast major-ity of students pictured. In addition, efforts were made to ensure that readers’judgments were reasonably consistent with each other.16

For the purposes of presentation, the organizations were grouped into athle-tic teams, performing-arts groups, publications, and other organizations. Al-though the inclusion of pictures in a school’s yearbook is typically the decisionof one adviser and a group of students working under a constraints of time andmoney, making it unlikely that consistent criteria were followed for the inclu-sion of groups across these schools, there is no evidence of systematic biasregarding which organizations were included. No attempt was made to compilerosters of organizations or to record instances of multiple memberships.

Besides information on clubs and teams, it was important to know abouteach school’s overall racial composition. For the public high schools in thesample, school racial composition was calculated from the National Center forEducation Statistics (NCES) Common Core of Data for 1997–1998, the yearcorresponding to the yearbooks. For the private schools in the sample, compara-ble data were taken from the NCES Private School Universe Survey for 1997–1998. As a check on the comparability of those official data with estimatesobtained from examining pictured students in yearbooks, comparisons weremade for a random subsample of 10 public high schools in the sample for totalenrollment and racial mix. These comparisons showed that calculations basedon the yearbooks understated total enrollment on average by 8.4% and overs-tated the percentage of whites by less than 0.4 percentage points. An inspectionof the private school data also showed that the yearbooks yielded enrollmentsand racial compositions close to the survey numbers. Estimates of grade 9–12enrollment in the private schools, calculated by multiplying by four the numberof 11th-graders pictured, understated actual enrollment on average by 7.2%,while estimates of racial composition based on the 11th grade-understated theshare of whites in the school by an average of 2.8 percentage points.

Table 1 presents a summary of the data collected for the 193 sample highschools, divided by region and type of school. Not surprisingly, the public

Page 8: Interracial Contact in High School Extracurricular Activities

32 THE URBAN REVIEW

TABLE 1.Means of Selected Variables, by Region and Type of School

Non-South South All

Private Public Private Public

1 Number of schools 20 89 6 78 193Mean valuesa

2 School enrollment 704 1,419 825 1,275 9733 Number of organizations 43 57 64 49 53

School racial composition (%)4 Black 10.3 6.4 3.4 31.8 17.55 Other nonwhite 8.0 8.8 8.0 5.8 7.46 Total nonwhite 18.3 15.2 11.4 37.6 24.9

Organizations’ racial composition (%)7 Black 8.6 5.5 2.9 27.4 15.18 Other nonwhite 8.3 6.7 4.2 4.0 5.69 Total nonwhite 16.9 12.2 7.1 31.4 20.7

10 Underrepresentation of nonwhites inorganizationsb �1.4 �3.0 �4.3 �6.2 �4.2

aMeans are weighted by school enrollment.bDifference between nonwhite percentage in organizations and the nonwhite percentage in school.

schools on average had larger average enrollment, but the average number oforganizations did not clearly differ by size of school. Among the public highschools, those in the South had much higher percentages of black students thanthose elsewhere, although other nonwhites were more common outside theSouth; these differences are very much in line with aggregate enrollment statis-tics for the nation. These differences are reversed among private schools, butthe much smaller number of private schools may not yield representative racialcompositions. Below the racial composition for schools is displayed the compo-sition of school organizations. These data reveal that, on average, nonwhitesparticipated proportionately less (based on numbers of members by race) thantheir white compatriots. For all the schools in the sample, the weighted averagepercentage nonwhite in organizations was 20.7%, somewhat less than the24.9% share of nonwhites in total enrollment.17

PATTERNS OF INTERRACIAL CONTACT IN CLUBS AND TEAMS

The degree of interracial contact in school organizations depends on threefactors. First, the racial composition of school organizations is necessarily afunction of that of the school from which they draw members. A school with nononwhites can obviously have no nonwhites in its teams and clubs, so there

Page 9: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 33

obviously can be no interracial contact. Second, interracial contact depends onthe degree to which students of different racial and ethnic groups participate inextracurricular activities. If students of any group join organizations at a lowerthan average rate, the potential for interracial contact is necessarily lessened.Using data from a national survey of high school students, McNeal (1998, p.187) found that whites participated in most types of extracurricular activities ata higher rate than black or Hispanic students, the exceptions being cheerleadingand vocational activities. The rate for Asian-Americans exceeded that forwhites in academic clubs, publications, and, to a small degree, in athletics.18

Table 1 reveals that, for the current sample, nonwhites comprise a smaller shareof the membership of organizations than they do of school enrollment. Thisdisparity could arise if a smaller percentage of nonwhites participated in extra-curricular activities or if they joined fewer organizations on average than whitestudents, or it could result from a combination of these two effects. Because thecurrent data set does not allow the tracking of multiple memberships, it isimpossible using it to distinguish between these two possible effects. Suffice itto say that the average rate of memberships per student is higher for whitestudents than nonwhite students.

A third factor affecting interracial contact is the evenness with which stu-dents of the various groups are distributed across organizations. If a significantnumber of clubs or teams are composed entirely of students of one racial group,the potential for interracial contact obviously will be lessened. Indeed, an in-spection of the sample of yearbooks reveals that such homogeneous groups areby no means uncommon. Out of the entire sample of 8,849 organizations in 193schools, 3,114 (or 35.1%) were exclusively white and 253 (or 2.9%) were ex-clusively nonwhite. Even if one looks only at high schools with nonwhite en-rollments between 10% and 90% (covering 101 schools), fully 19.0% of theorganizations were all-white, and another 3.6% were exclusively nonwhite.

To illustrate what kinds of organizations these were, Table 2 lists the all-white and all-nonwhite groups in 10 selected high schools in the sample. Whilesome of the listed organizations—especially clubs—appear to have race as anorganizing principle, most do not. A listing of the homogeneous organizationsin all 101 of the 10–90% nonwhite schools (not shown in the table) reveals thatthe most common all-white groups were golf, cheerleading, softball, and base-ball. The all-nonwhite groups appearing most often were boys’ basketball, girls’track, black student union,19 and gospel choir. The larger list of homogeneousorganizations confirms what is suggested in Table 2: A much higher proportion ofall-nonwhite organizations were designed ostensibly for specific racial or ethnicgroups, as compared to all-white organizations. Of course, organizations need notbe so thoroughly segregated for the degree of interracial contact to be affected. Ingeneral, the degree of segregation will depend on how the distribution of studentsacross organizations differs from a racially balanced distribution.

Page 10: Interracial Contact in High School Extracurricular Activities

34 THE URBAN REVIEW

TABLE 2.All-Nonwhite and All-White Organizations, 10 Selected High Schools

School StateSchool

type

Percentagenonwhitein school

All-nonwhiteorganizations

All-whiteorganizations

1 NY Private 20 Black Action StudentAssociation(BASA)

Fitness Club

French Club Gaelic ClubLa Nacion Latina Math Team

Varsity VolleyballVarsity BasketballTennisGolf

2 MI Public 21 (None) BaseballSoftballGolfVolleyballJV VolleyballFreshman Volleyball

3 OK Public 24 (None) CheerleadersFreshman Cheer-

leadersJV BaseballGirls’ GolfBoys’ Golf

4 MA Public 37 Latin Club YearbookLatino-American Club Girls StateAsian-American Club Interact Club

Drama ClubField HockeyIndoor Cross Country

5 SC Public 43 Aframhis Club National HonorSociety

Interact ClubFellowship of Chris-

tian AthletesGirls’ Soccer

6 CA Public 44 Drill Team Student Council(Junior)

Movimiento Estudian-til Chicano DeAzlan (MECHA)

Student Council(Freshman)

Page 11: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 35

TABLE 2. (Continued )

School StateSchool

type

Percentagenonwhitein school

All-nonwhiteorganizations

All-whiteorganizations

Black Student Union Women’s Water PoloHispanic Youth

LeadershipGerman

7 TX Public 49 Multi-CulturalAssociation

Girls’ Cross Country

Fellowship of Chris-tian Athletes

8 NC Public 49 JV Cheerleading JV SoftballGirls’ Track Girls’ Tennis

JV VolleyballVolleyballBoys’ Tennis

9 SC Public 54 Alpha Delta Rho Astronomy ClubChorus GolfKappa Gamma Psi Boys’ TennisSoutheastern Consor-

tium of MinorityEngineers (SCME)

Girls’ Tennis

Boys’ Basketball10 GA Public 68 Vocational Industrial

Clubs of America(VICA)

Thespian Society

Vocational Oppor-tunities Clubs ofAmerica (VOCA)

JV Cheerleading

Spanish Club Girls’ BasketballGirls’ Track JV Girls’ Basketball

Golf

In order to look more closely at aspects of interracial contact in school orga-nizations, Table 3 groups the 193 schools by racial composition. To examinethe relative rates of membership between white and nonwhite students, ColumnD shows the gap between the percentage of nonwhites enrolled and the percent-age of nonwhites in school organizations. Whereas nonwhites were slightlyoverrepresented in organizations in the 84 most preponderantly white schools(those in the first two rows), they were underrepresented in schools with 10%or more nonwhites. Measured by the percentage point difference, this gap tends

Page 12: Interracial Contact in High School Extracurricular Activities

36 THE URBAN REVIEW

TABLE 3.Interracial Contact by School Racial Compositiona

Percentagenonwhite Average number

Percentage inorganizations

where they are25% or less

minority

Percentagenonwhite inschool

Numberof

schools School

Allorgani-zations

Under-repre-senta-tion ofnon-

whitesb

Exposurerate tonon-

whites

Member-ship seg-regationindexc

Non-whites

exposedto

whites

Whitesex-

posedto non-whites Whites

Non-whites

Under 5% 55 3.0 3.7 0.7 3.4 8.1 1 28 0.0 92.75–10% 29 7.0 7.4 0.4 6.3 14.9 2 27 0.0 81.710–15% 19 11.6 9.9 �1.7 8.0 19.2 3 24 0.1 67.715–20% 11 17.4 14.8 �2.6 11.4 23.0 3 20 0.2 49.920–25% 14 21.9 15.3 �6.6 12.4 19.0 4 21 0.2 51.325–30% 9 27.1 23.3 �3.8 17.4 25.3 5 17 0.4 26.630–40% 14 34.7 24.5 �10.2 18.8 23.3 6 17 0.5 28.140–50% 15 45.0 36.6 �8.4 25.9 29.2 10 18 1.8 15.650–60% 13 56.1 43.4 �12.7 29.3 32.5 9 11 2.6 8.860% and over 14 80.9 69.6 �11.3 51.0 26.7 17 7 26.5 2.1All 193 24.9 20.7 �4.2 15.3 26.1 5 21 2.7 53.5

aMeans weighted by school enrollment.bDifference between nonwhite percentage in organizations and the nonwhite percentage in school.c100*(percentage nonwhite—exposure rate)/percentage nonwhite. See text.

to rise with the percentage nonwhite in the school; measured by the propor-tional difference between the two rates, however, the degree of underrepresenta-tion neither increases nor decreases systematically.

Also of interest is the calculated exposure rate in organizations, shown inColumn E. This rate is calculated for each school as

E � (1/W) �i

Wi [Ni /(Wi � Ni)], (1)

where Wi and Ni are the number of whites and nonwhites, respectively, in orga-nization i and W is the total number of white members in all organizations inthe school. This rate is interpreted as the percentage nonwhite in the averagewhite student’s school organization. Two schools with the same overall racialmix, for example, would have different exposure rates if all the organizations inone school had the same racial makeup as the school while organizations in theother school differed in racial composition. In the extreme, a school whose

Page 13: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 37

organizations were entirely segregated by race would have an exposure rate ofzero, meaning that the average white student was in an organization with nononwhites and the average nonwhite student was in an organization with nowhites. If school groups in each school all had the same racial composition, thisexposure rate would equal the percentage nonwhite in all organizations (shownin Column C). For all the organizations in the sample, the average exposure ratewas 15.3, meaning that the average white member was in an organization thatwas 15.3% nonwhite. Of course, this exposure rate differs across schools, tend-ing to rise with the percentage nonwhite in the school.

A convenient way to summarize the gap between the exposure rate and theoverall nonwhite percentage in organizations is to define an index of member-ship segregation, based on the percentage divergence between this rate of expo-sure and the racial composition of all organization members. This index isdefined as

S � 100(PCN � E)/PCN, (2)

where PCN is the overall proportion nonwhite in organizations.20 This measureranges from zero, signifying racially balanced organizations, to 100, signifyingthat the school’s organizations are completely segregated by race. Note that thisis a measure of racial balance across a school’s organizations and is indepen-dent of the degree of over- or underrepresentation of nonwhites in organiza-tions. A striking finding, shown in Column F of the table, is the tendency formembership segregation to be higher in schools with higher percentages ofnonwhites. For the whole sample, the actual degree of interracial exposure(15.3% nonwhite in the average white student’s organization) is some 26% lessthan it would be if all school organizations within each school had the sameracial composition (20.7% nonwhite). Combining the effect of this membershipsegregation with the lower rate of participation in organizations among minoritystudents reveals an even larger gap between the degree of actual interracialcontact in organizations and the rate that would exist if participation rates wereequal and all organizations were racially balanced—about 39% (100(0.249–0.153)/0.249).

Instead of looking at percentages, one might instead like to know the numberof students from another racial group any student member is typically exposedto in school organizations. Columns G and H of Table 3 give these averages.Based on the sample of yearbooks, white students who were members of schoolgroups could expect on average to be in an organization with 5 nonwhites. Bycontrast, the average nonwhite member was in an organization with 21 whites.Not unexpectedly, the number of nonwhites for the average white increasedwith the school’s overall nonwhite percentage, while the number of whites forthe average nonwhite fell.

Page 14: Interracial Contact in High School Extracurricular Activities

38 THE URBAN REVIEW

One other question of interest is whether students are in organizations inwhich they are significantly outnumbered by other racial groups. This situationwould be uncommon if school organizations were characterized by “tippingpoints,” thresholds beyond which members of one group will tend to abandonan organization.21 In his study of a desegregating high school in Memphis dur-ing the 1970s, Collins (1979) reports that, except when they were starters,whites stopped participating on school teams when they were no longer in themajority on the team. To see whether a phenomenon like this might be at workinside high schools, Columns I and J of the table give the percentage of whiteand nonwhite members, respectively, who belong to in an organization in whicha student’s own racial group is 25% or less of the total membership. Column Ishows quite clearly that this situation almost never occurs for white high schoolstudents. Only about 3% of the white members were in organizations wherewhites were a quarter or less of the group. For nonwhites, the experience wasquite different: over half of the nonwhite members were in groups where theywere outnumbered by whites by at least three to one. While these data do notprove there is a tipping point for whites in school organizations, they are con-sistent with the existence of one somewhere below 75%.

Does interracial contact differ by type of organization? Is there any factualbasis to the stereotype of the all-white soccer team? Do nonwhites predominateon basketball teams? Table 4 divides the 8,849 organizations in the sampleschools into six categories for sports teams and three for other types of organi-zations, with sports teams accounting for roughly half of all organizations in thesample. The table does indeed reveal a striking degree of divergence in racialcomposition across these types. Among the listed categories, the organizationswith the highest percentages of white membership were the baseball and soccerteams, both over 90% white. At the other extreme were teams for two othersports, basketball and football, with slightly less than three-quarters of theirmembers being white. To see whether whites or nonwhites were systematicallyunderrepresented in organizations of various types, one needs to compare theracial composition of organizations with that of the schools with which eachwas affiliated. This comparison is presented in the table’s next column, whichshows a weighted mean for the difference between organization and schoolracial composition. These figures show that, on average, all but two of the ninetypes of organizations had a preponderance of whites. Those with the largestgaps were baseball, soccer, and publications. In two types of organizations,basketball and football, nonwhites were overrepresented. In this case, then, ste-reotypes do contain an element of truth.

A question of historical and policy importance to be posed in the currentcontext is whether there is more or less interracial contact in the South, theregion that until the 1960s maintained state-sanctioned segregation in publicschools. One might reasonably conjecture that school organizations in the South

Page 15: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 39

TABLE 4.Summary Statistics by Type of Club, All High Schools

Percentage of Members

Type oforganization

Number oforganiza-

tions

Number ofschoolswith at

least one

Averagenumber ofmembers White Black

Othernonwhite

All non-white

Under-repre-senta-tion ofnon-

whitesa

All organiza-tions 8,849 193 20 82.8 12.1 5.1 17.2 �4.2

Sports teams 4,431 193 17 84.5 11.7 3.8 15.5 �4.5Baseball 548 155 15 91.5 5.9 2.6 8.5 �14.9Basketball 758 186 12 73.8 23.7 2.5 26.2 7.0Cheerlead-

ing 434 173 12 86.3 10.6 3.0 13.7 �7.3Football 297 165 36 73.2 24.5 2.3 26.8 2.6Soccer 465 158 19 91.3 3.2 5.5 8.7 �11.6Other sports 1,929 191 17 87.3 8.0 4.7 12.7 �6.9

Other organi-zations 4,418 188 23 81.7 12.3 6.0 18.3 �4.3Performing

arts 844 167 29 84.9 11.0 4.1 15.1 �6.0Publications 248 133 17 87.6 6.2 6.2 12.4 �9.3All other 3,326 184 22 80.3 13.1 6.6 19.7 �3.1

Source: Yearbook data and Common Core of Data; author’s calculations.aNonwhite percentage in clubs minus nonwhite percentage in schools, weighted by school totalenrollment.

would be more racially segregated than those elsewhere, perhaps as a mani-festation of more virulent racial animosity or of informal pressures brought tobear by school authorities seeking to thwart desegregation efforts. In order totest this conjecture, regressions were run with data on organizations using as adependent variable the percentage nonwhite in the organization and data onschools with a dependent variable the average exposure rate of white to non-white members. Equations 5.1 and 5.2 in Table 5 are based on the regressionsfor organizations, the second one adding indicators for the various types ofschool organizations listed in the previous table. To reflect the necessary depen-dence of organization racial composition on school racial composition, all ofthe right-hand-side variables are interacted with school racial composition,meaning that for an all-white school the regression will necessarily yield aprediction of all-white organizations. To allow the relationship to be nonlinear,school racial mix is entered as a cubic function.22

Equation 5.1 implies that, holding constant the school’s racial composition,organizations in the South tended to have higher, not lower, percentages ofnonwhites than those outside the South. No significant difference was found forprivate schools, however. Equation 5.2 adds dummy variables for eight of the

Page 16: Interracial Contact in High School Extracurricular Activities

40 THE URBAN REVIEW

TABLE 5.Regressions Explaining Racial Composition in Organizations and

Interracial Exposurea

Equation (5.1) (5.2) (5.3)Unit of observation Organizations Organizations Schoolsd

Dependent variable Percentagenonwhite inorganization

Percentagenonwhite inorganization

Exposure rate ofwhites tononwhites

CoefficientStd.error Coefficient

Std.error Coefficient

Std.error

VariablePercentage nonwhite 0.936 0.030 0.997 0.030 1.054 0.056Pct. NW squared �0.886 0.110 �0.839 0.105 �1.290 0.160Pct. NW cubed 0.902 0.091 0.837 0.087 1.133 0.123Pct.NW*SOUTH 0.080 0.016 0.036 0.016 �0.007 0.013Pct.NW*PRIVATE �0.013 0.028 0.013 0.027 0.028 0.030NW representationb 0.713 0.037Intercept �0.006 0.004Organization type dummy

variablesc

Performing arts �0.042 0.017Publications �0.271 0.040Baseball �0.411 0.027Cheerleading �0.134 0.034Football 0.234 0.022Basketball 0.365 0.025Other sports �0.138 0.017Soccer �0.374 0.031

R-square 0.705 0.732 0.973N 8,849 8,849 192

aFor purposes of estimation, percentages nonwhite in organizations and schools are expressed as aproportion. Unweighted sample means are 0.174 and 0.206, respectively. Regressions 5.1 and 5.2are weighted by organization size; 5.3 is weighted by school enrollment. Coefficients in bold aresignificant at the 95% level.bPercentage nonwhite in all organizations minus percentage nonwhite in school.cExcluded category is other organizations.dOne school with no whites excluded since exposure rate is undefined.

Page 17: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 41

nine organization types, each interacted with school racial composition. Thesevariables reflect the differences in the propensity of whites and nonwhites tojoin various kinds of organizations, as revealed in Table 4 above. The additionof these variables has the effect of decreasing the regional effect, so that it is nolonger significantly different from zero at the 95% level.

Equation 5.3 takes on the question of regional difference by focusing onaverage interracial exposure in schools. It adds a measure of relative participa-tion rate, the nonwhite representation gap used in previous tables (the differencebetween the percentage of nonwhites in organizations minus that in schools).The coefficient estimates in this regression reflect the importance of schoolracial composition and relative rates of participation. As in the previous twoequations, percent nonwhite in the school has a positive but nonlinear effect.Exposure is also enhanced by greater participation by nonwhites. However, theestimated regional difference is quite small in relation to its standard error. Insum, none of the estimated regressions suggests that school organizations areany more a vehicle for segregation in the South than in the rest of the country.

DISCUSSION

Should one look upon these findings with optimism or despair? Measured bythe racial composition of their enrollments, public schools in the United Stateshave become more integrated since the late 1960s. However, there is evidencethat actual interracial contact is not as great as would be suggested by enroll-ments, due to academic tracking and self-segregation in friendship patterns.This paper examines interracial contact along another important but largely un-explored dimension: high school extracurricular organizations. The data aretaken from a collection of 193 high school yearbooks. Tabulations based onthese data reveal that sports teams and clubs tend to be racially imbalanced.The degree of interracial exposure in these school organizations is thereforetypically less (averaging about 26% less) than it would have been had all theorganizations in each school been racially balanced. If the comparison is to theracial composition of the schools themselves, the gap is even larger (39%). Thisincreased gap is the result of the lower rate of participation by minority studentsin extracurricular groups.23

These calculations arise from two features of participation in extracurricularactivities. First, high school organizations tend not to be racially balanced. Theresult is that these organizations do not offer the same degree of interracialcontact as they would if their memberships had reflected the enrollments oftheir respective schools. The second contributing factor is the higher rate ofparticipation, noted elsewhere, of students from higher socioeconomic back-grounds and, correspondingly, of white students. In these extracurricular activ-ities, one sees another example of how parents and their children can blunt the

Page 18: Interracial Contact in High School Extracurricular Activities

42 THE URBAN REVIEW

impact of public policies through private, often collective, action. Althoughmuch less dramatic than moving their children from public to private schools,affluent parents are able to lessen the impact of desegregation by having theirchildren participate more often than other students in extracurricular activities,by giving them opportunities to learn sports skills outside of school and beingable to provide them with transportation to practices and after-school meetings.

Interpreting the evidence here on segregation in extracurricular activities hassomething of a half-full/half-empty character to it. If one takes as a point ofcomparison the racial composition that would exist if all organizations were toreflect the overall makeup of schools, then these figures suggest that the pat-terns of organization membership in these high schools has a sizable segrega-tive effect, reducing the exposure that might otherwise have existed. This stan-dard seems overly demanding, however, in part because academic trackingreduces interracial contact from the levels suggested by a school’s overall racialcomposition. As noted above, the little empirical that has been done on within-school segregation suggests that tracking does have the effect of reducing inter-racial contact inside schools. If one settles instead for the less stringent standardof the racial composition of those who actually participate, there is still imbal-ance across organizations, which again reduces the amount of interracial con-tact.

Instead of using any of these racial-balance standards for judging the expo-sure rates for school organizations, it may be more realistic to take as thebenchmark of comparison the extent of contact that would have existed if theonly interracial contact outside of class arose from friendships. As noted above,interracial friendships are relatively uncommon. Using this benchmark as thestandard of comparison, one would conclude that interracial contact has beenenhanced rather than retarded as a result of extracurricular activities. Indeed, ifthe personal relationships that develop among members of clubs and teams aremore significant for the formation of interracial tolerance and understandingthan those in classrooms, the existence of these organizations may constitute aninstrument with significant potential for more tolerant racial attitudes.

SUMMARY AND CONCLUSION

Based on an examination of yearbooks for 193 high schools, this paper pre-sents descriptive measures of interracial contact in 8,849 sports teams and otherextracurricular organizations. The findings reveal that the extent of such contactin these organizations is markedly less than would be the case if every organi-zation reflected the racial mix of its school. The gap arises because nonwhitestend to be underrepresented in these organizations and because those nonwhiteswho are members tend to be distributed unevenly among the organizations.More than one-third of the 8,849 groups were all-white while only about 3%

Page 19: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 43

were exclusively nonwhite. Owing in large part to their overall numerical pre-ponderance, white students rarely found themselves outnumbered in groups byas much as three to one; by contrast, nonwhites often were in this position. Bytype of organization, nonwhites tended to be most severely underrepresented onbaseball and soccer teams and on the staffs of school publications. They wereoverrepresented on football and basketball teams. Controlling for the type oforganization, there was no statistical difference in the extent of underrepresenta-tion by region or between public and private schools.

Because extracurricular activities appear to play an important part in thepersonal development of high school students and their racial attitudes, patternsof contact such as these are worthy of continued attention. Although the mea-sured rates of contact are lower than their theoretical maximums, they provideopportunities for interracial contact that would surely not exist in their absence.

NOTES

1. I am grateful to Thomas Anderson, Chi Leng, Cathleen McHugh, Robert Malme, Karen Price,Martin Steinmeyer, and Jennifer Sturiale for research assistance; to Terri Finn for her assistancein making data available; to Philip Cook, Harriet Morgan, Theresa Newman, William Rodgers,and John Wilson for useful comments; and to Duke University for financial support.

2. For an examination of interracial contact along a number of dimensions, see Sigelman et al.(1996).

3. In this paper, the term racial is used as shorthand to refer to both racial and ethnic distinctions,white refers to non-Hispanic whites, and nonwhite refers to all others.

4. For a recent study of contact in one high school over time, see Rodgers (1999).5. For a detailed analysis of the role of organizations in the social network of one high school, see

Quiroz, Gonzalez, and Frank (1996).6. For a description of recent trends, see Orfield et al. (1997). Greatly affecting these measured

rates of integration in the public schools have been decisions by white families about where tomove or whether to enroll their children in private schools. Whether these moves conform tothe stereotype of “white flight” that occurs in direct response to changing racial compositionsin the public schools or instead arise from more complicated decision making based on avariety of factors, the effect of these moves has been to reduce the number of white students inmany districts that enroll a significant number of minority students. For an analysis of whiteenrollment patterns in the face of public school desegregation, see Clotfelter (2001).

7. See, for example, Clotfelter (1999).8. Gamoran’s (1992) study of placement in ninth-grade honors English classes in several mid-

western school districts reveals significant underrepresentation of minority students, even whenachievement test scores were controlled for. For a study of within-class grouping in elementaryschool, see Epstein (1985). For a summary of the arguments against tracking, including itsracial aspects, see Oakes (1987), for example.

9. Using the same measure of segregation given in Equation 2 of this paper, based on black-whiteexposure indices (calculated as in Equation 1), Morgan and McPartland find that the averageindex for their sample of high schools was 0.11, implying that the interracial exposure rate inclassrooms was 11% less than it would have been had classrooms been racially balanced withineach school. While the average high school in their sample was 66.04% white, the averageblack student was in a classroom that was 58.98% white. The index was calculated as 66.04–

Page 20: Interracial Contact in High School Extracurricular Activities

44 THE URBAN REVIEW

58.98)/66.04 � 0.11. They note that such segregation indices are equivalently based on whiteexposure to blacks or vice versa. The average divergence was 13% (S � 0.13) in high schoolsin the South, compared to about 0.09 in the Northeast and Midwest and 0.07 in the West.Among all high schools, the segregation rate was highest for those with 40–70% white enroll-ment (0.15) and lowest in schools with less than 10% or more than 90% white (0.06) (Morganand McPartland 1981, Tables 3, 4). For a detailed study of interracial contact within a publicmiddle school, see Schofield (1982).

To give a sense of how tracking might reduce the rate of interracial exposure in a school,consider a hypothetical school with a 20% black and an 80% white composition. If black andwhite students in this hypothetical school were distributed across four academic tracks in thesame proportions as they are actually distributed nationwide across four performance ranges inthe National Assessment of Educational Progress (NAEP) test of reading proficiency for 17-year-olds (U.S. Department of Education 1999, Table 112, p. 130), the tracks would range from9.1% black in the highest track to 46.4% in the lowest. Whereas blacks would comprise 20% ofoverall enrollment, the exposure rate of whites to blacks would be only 18.6%, implying asegregation index of S � (20–18.6)/20 � 7.0%. This gap is a bit smaller than, but still on thesame order of magnitude as, the findings of Morgan and McPartland (1981) noted above.

10. See also Schofield (1979) for similar findings.11. As Arrow (1998, p. 98) has noted, “Social segregation can give rise to labor market segregation

through network referrals.” By the same token, increased interracial contact has the potential toexert the opposite effect. See also Schofield (1995, p. 610) for a discussion of these effects.

12. Quiroz, Gonzalez, and Frank (1996) argue that, because of school policies and other pressures,membership in high school organizations is neither open nor entirely voluntary. See alsoMcNeal (1998, p. 184) for a discussion of this point.

13. For purposes of this study, I followed Orfield and Monfort (1992, p. 2) in defining the South asfollows: Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, North Carolina, SouthCarolina, Tennessee, Texas, and Virginia.

14. Of the sample high schools, 26, or 13%, were private, compared to 11% among all highschools in the United States (U.S. Department of Education 1999, Tables 5, 60; pp. 14, 71).Regarding school size, as shown in Table 1, the average sizes of the private schools in thesample were 825 and 704 in and out of the South, respectively. For public schools, the compa-rable averages in the sample schools were 1,275 and 1,419. In the fall of 1995, the averageprivate high school in the United States had 320 students, and the average public high schoolhad 595 (U.S. Department of Education 1999, Tables 3, 60; pp. 12, 71).

15. In the year 1993–1994, 46% of U.S. private secondary schools were Catholic (U.S. Departmentof Education 1999, Table 62, p. 73.)

16. As a check of consistency across readers, selected schools were examined by two differentreaders and the results compared. Where one reader’s tabulations differed from those producedby a benchmark reader, the former tabulations were redone.

17. If yearbooks systematically tend to picture whites at a higher rates than nonwhites, the calcula-tions of nonwhite underrepresentation would be understated, as would be the actual gap inexposure. As noted above, the comparisons using the 10-school subsample show a slight ten-dency to include nonwhites less completely than whites in yearbook photos of individual stu-dents. If this tendency extended to pictures of school organizations, then the calculated rates ofnonwhite participation would tend to overstate the true degree of underrepresentation.

18. In multivariate analysis, using two national surveys, McNeal (1998, p. 187; 1999, pp. 300–301)found that participation rises with socioeconomic status and performance on standardized tests.Holding those and other factors constant, the partial effect on participation is actually positivefor blacks for most types of activities. See also Moore and Ehrie (1999) for evidence fromanother national survey that participation in extracurricular activities by 12- to 17-year-oldsrises with socioeconomic status.

Page 21: Interracial Contact in High School Extracurricular Activities

INTERRACIAL CONTACT IN HIGH SCHOOL ACTIVITIES 45

19. Similar organizations included African Experience Club, African Heritage Club, Afro-Ameri-can Club, and Black Action Student Association.

20. An index of this form has been used to measure segregation in schools, for example, in Morganand McPartland (1981) and Clotfelter (1999, 2001). For a discussion of it and its relationship tomeasures of exposure, see Clotfelter (1978). It is useful to note that the value of S is invariantwith respect to which of two groups is used as the basis for calculating the exposure rate. Thatis, S can be calculated using the exposure of nonwhites to whites, where W, the overall percent-age of students who are white, is the maximum for this exposure rate: S � (W � ENW) / W.

21. For a discussion of tipping points in the context of school desegregation, see Clotfelter (1976).22. To reflect the possible influence of average income level on the composition of school organi-

zations, variants of the equations presented in Table 5 were estimated with the addition ofmedian family income in the zip code where the school was located, entered by itself andinteracted with percentage nonwhite. Neither of these variables was statistically significant, andnone of the other coefficients was materially altered.

23. Another qualification to the measures of segregation used in the current paper is the argument thatthe correct benchmark of comparison ought to be not an exactly racially balanced composition butthe exposure rate that would arise if membership in organizations were random. Owing to thesmall size of these school organizations, the laws of chance would determine that their racialcompositions would often deviate somewhat from the racial composition of all members oforganizations or the school, depending on which standard was being used. Using the binomialdistribution, sample calculations were made of expected exposure rates under the assumption ofrandom assignment. When calculated, these rates turned out to be quite close to the racial balancerates, so the latter standard was retained, as is conventional in studies of segregation.

REFERENCES

Allport, Gordon W. (1954). The Nature of Prejudice. Cambridge, MA: Addison-Wesley.Arrow, Kenneth J. (1998). What has economics to say about racial discrimination? Journal of

Economic Perspectives 12(Spring): 91–100.Braddock, Jomills Henry, II, Crain, Robert L., and McPartland, James M. (1984). A long-term view

of school desegregation: Some recent studies of graduates as adults. Phi Delta Kappan 66(De-cember): 259–264.

Clotfelter, Charles T. (1976). School desegregation, “tipping,” and private school enrollment. Jour-nal of Human Resources 22(Winter): 29–50.

Clotfelter, Charles T. (1978). Alternative measures of school desegregation: A methodological note.Land Economics 54(Aug.): 373–380.

Clotfelter, Charles T. (1999). Public school segregation in metropolitan areas. Land Economics75(November): 487–504.

Clotfelter, Charles T. (2001). Are whites still “fleeing”? Racial patterns and enrollment shifts in urbanpublic schools, 1987–1996. Journal of Policy Analysis and Management 20(Spring): 199–221.

Collins, Thomas W. (1978). Reconstructing a high school society after court-ordered desegregation.Anthropology and Education Quarterly 9(Winter); 248–257.

Collins, Thomas W. (1979). From courtrooms to classrooms: Managing school desegregation in aDeep South school. in R. C. Rist (Ed.), Desegregated Schools: Appraisals of an American Exper-iment, pp. 89–113. New York: Academic Press.

Ellison, Christopher G., and Powers, Daniel A. (1994). The contact hypothesis and racial attitudesamong black americans. Social Science Quarterly 75(June): 385–400.

Epstein, Joyce L. (1985). After the bus arrives: Resegregation in desegregated schools. Journal ofSocial Issues 41(3): 23–43.

Gamoran, Adam (1990). Access to excellence: Assignment to honors English classes in the transi-

Page 22: Interracial Contact in High School Extracurricular Activities

46 THE URBAN REVIEW

tion from middle to high school. Educational Evaluation and Policy Analysis 14(Fall 1992):185–204.

Granovetter, Mark (1986). The micro-structure of school desegregation. In Jeffrey Prager, DouglasLongshore, and Melvin Seeman (Eds.), School Desegregation Research: New Directions in Situ-ational Analysis. New York: Plenum Press.

Hallinan, Maureen T., and Williams, Richard A. (1989). Interracial friendship choices in secondaryschools. American Sociological Review 54(February): 67–78.

Joyner, Kara, and Kao, Grace (2000). School racial composition and adolescent racial homophily.Social Science Quarterly 81(September): 810–825.

McNeal, Ralph B., Jr. (1998). High school extracurricular activities: Closed structures and stratify-ing patterns of participation. Journal of Educational Research 91(January/February): 183–191.

McNeal, Ralph B., Jr. (1999). Participation in high school extracurricular activities: Investigatingschool effects. Social Science Quarterly 80(June): 291–309.

Mickelson, Roslyn Arlin (2001). Subverting Swann: First- and second-generation segregation in theCharlotte-Mecklenburg Schools. American Education Research Journal 38 (Summer): 215–252.

Moore, Kristin, and Ehrie, Jennifer. (1999). Children’s environment and behavior: Participation inextracurricular activities. Snapshots of America’s Families, C-5. Washington: Urban Institute.

Morgan, P. R., and McPartland, James M. (1981). The extent of classroom segregation withindesegregated schools. Unpublished manuscript, Johns Hopkins University, Center for Social Or-ganization of Schools, August.

Oakes, Jeannie. (1987). Tracking in secondary schools: A contextual perspective. Educational Psy-chologist 22(2): 129–153.

Oakes, Jeannie. (1993). Ability grouping, tracking, and within-school segregation in the San Joseunified school district. Unpublished report, October.

Orfield, Gary. (1983). Public School Desegregation in the United States, 1968–1980. Washington,DC: Joint Center for Political Studies.

Orfield, Gary, Bachmeier, Mark D., James, David R., and Eitle, Tamela (1997). Deepening segrega-tion in American public schools. Harvard Project on School Desegregation, unpublished paper,Harvard University, April 5.

Orfield, Gary, and Monfort, Frank (1992). Status of School Desegregation: The Next Generation.Cambridge, MA: Metropolitan Opportunity Project, Harvard University, March.

Quiroz, Pamela Anne, Gonzalez, Nilda Flores, and Frank, Kenneth A. (1996). Carving a niche inthe high school social structure: Formal and informal constraints on participation in the extracurriculum. Research in Sociology of Education and Socialization 11: 93–120.

Rodgers, William M., III, and Ragsdale, Michelle R. (1999). Racial segregation in extracurricularactivities: A longitudinal analysis. Unpublished paper, November 9.

Rossell, Christine H., and Armor, David J. (1996). The effectiveness of school desegregation plans,1968–1991. American Politics Quarterly 24(July): 267–302.

St. John, Nancy (1975). School Desegregation: Outcomes for Children. New York: Wiley.Schofield, Janet Ward (1979). The impact of positively structured contact on Intergroup behavior:

Does it last under adverse conditions? Social Psychology Quarterly 42: 280–284.Schofield, Janet Ward (1982). Black and White in School: Trust, Tension, or Tolerance New York:

Praeger.Schofield, Janet Ward (1995). Review of research on school desegregation’s impact on elementary

and secondary school students. In James A. Banks and Cherry A. McGee Banks (Eds.), Hand-book of Research on Multicultural Education. New York: MacMillan Publishing.

Sigelman, Lee, Bledsoe, Timothy, Welch, Susan, and Combs, Michael W. (1996). Making contact?Black-white social interaction in an urban setting. American Journal of Sociology 101(March):1306–1332.

U.S. Department of Education, National Center for Educational Statistics (1999). Digest of Educa-tion Statistics 1998. Washington, DC: Government Printing Office.


Recommended