Order 1261668: Schooling Difference

Social Problems

, Vol. 52, Issue 3, pp. 398–418, ISSN 0037-7791, electronic ISSN 1533-8533. © 2005 by Society for the Study of Social Problems, Inc. All rights reserved. Please direct all requests for permission to photo- copy or reproduce article content through the University of California Press’s Rights and Permissions website, at http://www. ucpress.edu/journals/rights.htm.

School Choice, Charter Schools, and White Flight

LINDA A. RENZULLI,

University of Georgia

LORRAINE EVANS,

University of Georgia

The “choice” movement of the 1990s culminated in a proliferation of charter schools. However, school choice and charter school options may have future consequences for racial segregation given the potential for white flight similar to that which occurred in the 1960s and 1970s. Drawing from racial competition theory, this article contributes to literature on education and stratification in a broader sense by examining white enrollment in charter schools and its possible consequences for racial segregation. Data are drawn from the Schools and Staffing Survey (SASS), the Common Core of Data (CCD), and a unique dataset on district academic quality. Analyses suggest that relatively even distributions of white and nonwhite students within districts and corre- sponding competitive pressures spur white charter school enrollment. We suggest that such racial competition within the educational arena may indeed be bolstering the “return to school segregation.”

Political and public debates concerning the implementation of school choice policies are more often based on beliefs than on empirical evidence (Fowler 2003; Manno, Vanourek, and Finn 2000). Supporters of school choice often suggest that school options will create a system where all parents can select “effective” schools to educate their children. Moreover, supporters argue that schools of choice can be effective regardless of their racial composi- tion or access to resources (Chubb and Moe 1990; Coons and Sugarman 1978; Holt 2000). In contrast, opponents of choice in the public school system suggest that choice policies will only exacerbate inequalities already manifest in our schools and threaten the very fabric of public education (Henig et al. 1999; Saporito and Lareau 1999; Wells 1993).

Charter schools are clearly an important case in point.

Although they are public and secular, they elude the bureaucratic constraints of school districts, thereby evoking contro- versy.

1

Proponents argue that charter schools significantly improve public education because they create: (1) choice in curriculum, structure, and discipline; (2) accountability for educa- tional outcomes and student progress; and (3) autonomy for teachers, parents, and administra- tors (Center for Educational Reform 1999; Nathan 1996). Proponents suggest that bolstering choice, accountability, and autonomy will result in high quality schools for all children, most notably those of poor and minority backgrounds (Nathan 1996). Opponents, in contrast, fear

The authors wish to thank E. M. Beck, Jeremy Reynolds, Elizabeth Stearns, Jody Clay-Warner, and Natalie Lacireno- Paquet for their thoughtful comments on earlier drafts of this article. This research was supported by a grant from the American Education Research Association, which receives funds for its “AERA Grants Program” from the U.S. Department of Education’s National Center for Education Statistics and the Office of Educational Research and Improvements under NSF Grant #REC-9980573. Opinions reflect those of the authors and do not necessarily reflect those of the granting agen- cies. Direct correspondence to: Linda Renzulli, University of Georgia, Department of Sociology, Baldwin Hall, Athens, GA 30602. E-mail: [email protected].

1. The precise definition of charter schools—as well as their accountability measures, teacher certification, and enrollment guidelines—vary by state. Although there are different types of charter schools, in this article, we do not dis- tinguish between market and non-market charter schools.

Though charter schools are public schools, we parsimoni- ously refer to traditional public schools as public schools and public charter schools as charter schools.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight

399

that charter schools cannot fix broader educational problems and, if anything, have become instruments of elitism that deplete public school resources (Alexander 1997; Berliner and Biddle 1995; Cobb and Glass 1999; Henig et al. 1999; Weiher and Tedin 2002; Wells 1993). Recent work by the American Federation of Teachers (AFT) concurs: charter schools may be no better able to educate students than are traditional public schools (Nelson, Rosenberg, and Van Meter 2004).

Despite conflicting interpretations over their impact, charter schools have emerged as the fastest growing educational innovation in the United States. As of 2004, charter schools were operating in 38 of the 42 states with charter school legislation (Renzulli and Roscigno Forthcoming). According to the Center for Educational Reform, charter schools educate nearly 700,000 students in 2,996 charter schools. As they continue to grow in number and as the percentage of children who attend them swells, the potential for charter schools to change the system of public education in the United States increases (Manno et al. 2000). One such change may be the resegregation of schools and school districts, as indicated by some initial evidence from the Civil Rights Project (Frankenberg and Lee 2003).

This article examines more closely the conditions under which we might see “pockets of white segregation” (Frankenberg and Lee 2003) and, in particular, instances wherein charter schools become primarily populated by white students. However, rather than simply confirm- ing that whites attend relatively segregated charter schools, we question the circumstances under which charter schools enroll a high percentage

of white students. For example, does the educational arena witness processes of racial competition and inequality that researchers have documented in other contexts. Analyses of race-specific outcomes such as lynching (Beck and Tolnay 1990), riots (Olzak, Shanahan, and McEneaney 1996), and (un)employ- ment (Tomaskovic-Devey and Roscigno 1996) indicate that economic and political competi- tion between whites and blacks cause responses such as social control and discrimination (Olzak, Shannahan, and West 1994). Schooling may be no different.

Drawing on recent racial competition research, particularly that of Olzak (1990, 1992; Olzak et al. 1996; Olzak et al. 1994), we argue that racial competition in school districts will affect white enrollment in—and thus white flight into—local charter schools. Analyses of data gathered from various sources allow us to assess how the distribution of nonwhite students in a school district influences patterns of white enrollment in charter schools. Data are drawn from the charter school component of the Schools and Staffing Survey (SASS), the Common Core of Data (CCD), a collection of charter school legislative statutes (Jennings et al. 1998), and a dataset we gathered that includes a measure of district academic quality. We conclude by discussing the implications of our findings for understanding racial competition, white flight, and stratification in public education more generally.

Racial Dynamics, White Flight, and School Segregation

Competition theorists argue that the roots of collective action and resource mobiliza- tion lie in competition for scarce resources, including status and institutional access (Olzak 1992; Olzak et al. 1994). According to Douglas S. Massey and Nancy A. Denton’s landmark book

American Apartheid

(1993), “White apprehension about racial mixing is associated with the belief that having black neighbors undermines property values,” which implies that “whites perceive blacks to be a direct threat to their social status” (p. 94). Whites get social status not only from the neighborhoods they live in (Massey and Denton 1993), but also from the quality of the schools their children attend (Bankston and Caldas 2002). Within this context of education, competition “can be generated objectively by growing minority enrollments in schools” and can be seen as a “threat to the status of the majority ethnic community” (Olzak et al. 1994:196–97). Though competition theory has not been used

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

400

RENZULLI/EVANS

extensively to understand white flight, it would suggest that for whites, nonwhite enroll- ments lower the status and desirability of schools, and ultimately cause whites to flee.

Whites may avoid individuals they perceive to be low-status, whether by moving to all-white or mostly white neighborhoods or enrolling their children in all-white or mostly white schools. Researchers in the 1970s began to examine the unintended consequences of mandated busing and desegregation on neighborhood schools (Coleman, Kelly, and Moore 1975; Farley 1975; Giles 1978; Sly and Pol 1978). Such analyses found, by and large, a strong relationship between black enrollment in public schools and white flight.

Schools and districts with black populations over 30 percent experienced greater loss of white students than those below 30 percent. Moreover, segregation between districts was greater than segregation within districts, indicating that the response to desegregation was the out-migration of whites to the suburbs, leaving primarily black central-city school districts (Wilson 1987).

Contemporary investigations of white flight and its relation to increased patterns of segregation reveal that the mechanisms visible in the 1960s and 1970s continued to operate into the 1980s and 1990s. Even today, much of the research on white flight supports David R. James’s (1989) finding that “white parents make decisions based on the actual or potential exposure of their children to blacks” (cited in Clotfelter 2001:202).

2

In fact, Charles T. Clot- felter (2001) and Kyle Crowder (2000) suggest that white flight is positively related to the minority student population and the ability to find desirable residential areas with a lower proportion of minority residents. The ease with which whites can find predominately white schooling in nearby areas increases the likelihood that they will exercise school choice through residential mobility, despite the fact that they may incur costs such as increasing the commute time to work (Clotfelter 2001; Frey 1979; Morgan and England 1984).

However, some white families can exercise flight without residential mobility by moving their children from racially heterogeneous public schools into more racially homogenous private schools, as has been the case historically. For instance, Kenneth Andrews (2001) reviews the desegregation policies of Mississippi from 1968 to 1971 and concludes that private acad- emy attendance increased dramatically as the proportion of black school-age children in public schools increased. Research generally shows that, like residentially mobile families, those who transfer schools but do not change homes tend to be affluent and white. Moreover, they maintain status by avoiding nonwhite schools (Fairlie and Resch 2002; Lankford and Wyckoff 2001; Levin 1999; Morgan and England 1984; Saporito and Lareau 1999).

Charter Schools’ Racial Enrollment: White Flight without Residential Mobility?

Charter schools add another option to the school choice menu, one that helps parents avoid residential mobility costs and private school fees. Furthermore, national-level research offers potentially encouraging evidence for those who hoped that charter schools would provide choice to those who historically have been unable to choose their schools: 52 per- cent of students in charter schools are nonwhite compared to 41 percent in traditional pub- lic schools (Frankenberg and Lee 2003; Gill 2001). However, national-level demographics tell us little about (1) the local concentrations of whites and nonwhites in charter schools and (2) how the racial composition and distributions of charter schools compare to the racial composition and distribution of local schools. Even if some charter schools are serving nonwhites, other charter schools may be largely populated by white students (Frankenberg and Lee 2003; Gill 2001). Below we outline the empirical findings about race and charter schools conducted at the state- and district-levels of analysis, which help supplement the potentially misleading findings of national-level studies.

2. Declining white populations in urban schools may also be attributed to both immigration rates and differences in birth rates (Frey 1995).

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight

401

Since charter school laws and other educational reforms are delegated to states, evaluat- ing charter schools’ racial enrollment at the state level may be more accurate than national averages. In fact, some researchers who have limited the scope of their analyses to the state level have found that charter schools tend to segregate students by race (Frankenberg and Lee 2003; Wamba and Ascher 2003). They show that, in some cases, white charter school students attend charter schools that are more homogenously white than are their public school counterparts. In California, Amy Stuart Wells (1998) found that Latinos were underrepresented in charter schools. Furthermore, although blacks went to charter schools in the same proportion as they attended public schools, charter schools themselves are often segregated: thirty-seven percent of charter schools were predominantly white (that is, 80 to 100 percent white). Robert E. Crew and Mary R. Anderson (2003) report that Florida charter schools are more segregated than traditional public schools (82 percent white compared to 51 percent white). Similar patterns were found in Arizona school districts where Casey D. Cobb and Gene V. Glass (1999) found charter school enrollment was 20 percent more white than traditional schools (see also Miron, Nelson, and Risley 2002 on Pennsylvania). Lance Fusarelli (2002) found that Texas charter schools disproportionately served minorities (43 of 89 schools) because the charter schools there were created explicitly to meet the needs of minority and at-risk children. Accordingly, Texas charter schools mostly serve minorities in segregated contexts and do not serve whites and minorities

together

at the same rate as the public schools. In sum, state analyses largely suggest that charter schools create greater segregation of whites and nonwhites. They do indeed serve minorities, but mostly in segregated contexts.

Charter school laws vary by state, so state analyses may accurately reflect conse- quences of state policy. Nevertheless, many state laws require charter schools to consider district-level characteristics when accepting and recruiting students. For example, in 1998, 15 of the 34 states with charter school legislation had direct provisions for district racial bal- ance (Jennings et al. 1998). The racial balance provisions specifically assert that charter schools must reflect the racial make-up of their district rather than the racial make-up of their state. Research conducted at the district level, therefore, may paint a more accurate picture of the racial composition of charter schools given their local context. In Texas, researchers found that there was more segregation among charter schools than among the schools within the district and thus did not necessarily reflect district racial composition (Weiher and Tedin 2002). Wells (1998) and Wells and associates (1999a; 1999b) found sim- ilar segregation patterns in 17 school districts they examined in California where racial groups were either overrepresented or underrepresented in charter schools compared to their local district demographics.

Taken together, the three levels of analyses—national, state, and district—offer differ- ent depictions of charter schools. At the national level, charter schools seem to be serving minorities adequately; while state- or district-level analyses reveal that minority charter school enrollments occur in largely segregated contexts. Nevertheless, little previous work investigates why charter schools attract the student populations they do, how racial bal- ance among charter schools has been affected (Gill 2001), or the ways in which racial com- petition within local educational contexts may be shaping the enrollment and segregation patterns we find.

Racial Competition and Charter Schools

Racial competition theory provides a useful, although typically overlooked, framework for addressing educational segregation processes in general, and the more specific foci and questions pertaining to charter schools raised previously. As noted by competition scholars, the dynamics of competition typically occur in localized contexts. For example, lynchings

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

402

RENZULLI/EVANS

in the South were a result of economic threats felt by whites from blacks in the local labor market (Beck and Tolnay 1990). Furthermore, interracial exposure in schools in the 1960s was associated with race riots in some cities (Olzak et al. 1994). In both cases, threat and competition occurred in contexts in which whites and blacks encountered one another. We believe that interracial competitive processes in the educational arena, including those relating to enrollment patterns, are similarly created at the local level. Because charter schools operate within a limited school catchment area in an educational market, competitive process will occur at the district and school level. While school choice decisions may be influenced by parental resources and based on calculated decisions by parents pertaining to what constitutes a high-achieving school, recent work suggests that cultural processes and perceptions—even those that are localized—may ultimately have a large impact. Certainly, race perceptions, local racial competition, and any ensuing antagonism may play a role (Bulman 2004).

Our analyses take this insight into account by examining charter schools in their local context. First, we use a national dataset of charter schools to analyze the schools in relation to their school districts. Rather than using case studies of districts, we use data that allow us to look more systematically at the racial enrollment of charter schools across districts in the United States, thus bolstering our ability to generalize. Second, our school district data gauge the degree to which nonwhites and whites attend schools together. We do not simply compare the racial make-up of all charter schools in a district with the racial composition of the school districts in which these schools are situated. Instead, we push the analysis fur- ther by attempting to predict racial enrollment for individual charter schools embedded within school districts that have unique school-level patterns of enrollment, while control- ling for important district, state, and charter school characteristics. Furthermore, we ana- lyze the distribution of nonwhite students among schools and utilize competition theory— specifically, expectations regarding white flight in racially competitive environments—to explain why charter schools vary in their racial composition.

Building on competition theory, we suspect that some of the very same competitive pres- sures that prompted white parents to move their children out of integrating schools and dis- tricts in the 1960s will affect enrollment choices and white utilization of charter schools today. Specifically, where white and nonwhite students are distributed equally among schools (i.e., a more racially competitive environment), we expect to find greater white utilization and enrollment in charter schools. In contrast, white flight into charter schools will be less pro- nounced in districts within which schools are already significantly racially separated. Contact and integration, two structural attributes of districts and their schools discussed in the white flight literature (Clotfelter 2001; Coleman 1975; Taeuber and James 1982), may increase white enrollment in charter schools.

Data and Measurement

One of the main limitations of charter school research is the inability to match charter schools to their surrounding school districts (Lin 2001). We overcame this obstacle by link- ing the Schools and Staffing Charter School Survey 1999–2000 (SASS) to the Common Core of Data 1998–1999 (CCD); both are national datasets collected by the National Center for Education Statistics (NCES). The SASS provides restricted-use data on charter schools operating in the United States during the 1999–2000 school year. The sampling frame was the population of 1,100 charter schools. A response rate of about 79 percent yields a sample of 870 charter schools (see NCES 2004). The CCD is an annual, national database of the uni- verse of public elementary and secondary schools and school districts. Each charter school in the SASS data has a zip code reflecting the physical address of the school. Charter school zip codes were linked to districts by using the CCD’s district identification numbers and zip

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight

403

codes. In most cases, the geographic location of a charter school makes it local to the school district in which it is physically embedded.

3

We supplement the combined data from the SASS and the CCD with a collection of charter school legislative statutes (Jennings et al. 1998).

We also supplemented our data with information about district academic quality as measured by standardized test scores. The 1998–1999 CCD do not contain information about district academic quality, but we were able to obtain test-score data for the districts repre- sented in our analysis by using websites or other means of contacting states and districts. Within states, the type of test used and the scoring methods are consistent, but between states there is some variability in test type and grade level. First, for example, some states used

Stanford 9

tests while others used the

Iowa Achievement Test

or other exams.

However, since parents use whatever test scores available in their districts to make decisions pertaining to school quality and where to enroll their children, the type of test is not relevant. Second, the grade levels tested varied, but all are at the elementary level.

4

In our sample of charter schools, 83 percent serve the elementary grades (K–5); therefore, using elementary tests scores is a reasonable proxy for school quality. In the appendix, we provide a table of the test type, grades, and scoring method by state.

5

We established two criteria for including charter schools in this analysis. First, schools had to be brick-and-mortar rather than home-school associations. Second, schools had to have students who could freely choose to attend the charter school. Thus, charter schools that were created solely for students who had been suspended or expelled, who have dropped out, or who have been referred for behavioral or adjustment problems were not included.

6

The final dataset includes 688 charter schools, distributed throughout 338 U.S. school dis- tricts.

7

The mean number of charter schools in a district is five with a range of one to 27. Detroit, Los Angeles, Houston, Tucson, and the District of Columbia have the most charter schools. This sample of districts enrolls a greater percent of nonwhite students than does the average regular school district in the United States (52 percent compared to 21 percent,

3. Only three states (Colorado, Florida, and Minnesota) have laws that restrict charter school admission to stu- dents who reside in the

sponsoring district.

We were able to check that we placed the charter schools for these three states in the correct sponsoring district by matching the charter school characteristics available in the SASS data to information found in the Charter School Directory (Center for Education Reform 2000). Fifteen states, on the other hand, restrict admissions to students residing within the district where the charter school is located (Jennings et al. 1998). Therefore, using physical zip codes of the charter school to place them in a district allows us to find their catchment area. For nine states with open enrollment polices for charter schools, the district served as a proxy for the students who are likely to attend.

4. We used the fourth-grade test scores because they were most frequently and consistently available across states and districts. However, if fourth-grade test scores were missing, we used third- or fifth-grade test scores.

5. The majority of states in this analysis (21) reported scores that indicated the percent of students in the district that passed the test at grade level. In three states, (Arizona, Hawaii, and Georgia) the test data were reported in terms of national percentile rankings (NPR) rather than the percent passing, and in one state (Kansas) no test data were reported. Since NPR and percent passing are not equivalent measures and cannot be compared, we used multiple impu- tation methods for these states and for Kansas in which no data are available. Multiple imputation introduces less bias to our analysis than would omitting the 190 charter schools with missing data (Little and Rubin 1987).

6. The charter schools that do not meet these criteria do not offer the same options as “regular” charter schools. Thus, they do not add the same set of choices to the landscape of educational options and may not reflect the same white flight mechanisms as the other charter schools (Vergari 2002). We dropped 107 charter schools from the analysis because their students were primarily schooled at home and did not meet the criteria for inclusion in this analysis. These home-based charter schools enroll a greater percent of whites students than do other charter schools, but their racial composition is similar to other home-based schools (Stevens 2001). We dropped 55 other charter schools because they were created for students excluded from their school districts. Statistical tests of means reveal that these schools are no different than other charter schools in terms of racial enrollment.

7. Of the total number of qualified charter schools we eliminated 2.8 percent due to missing data (20 of 708). We also eliminated three charter schools because we could not match them to a district (see footnote three). Finally, we eliminated 17 charter schools because they were in districts that had missing data from the CCD for the percent of enrolled students who were poor (an important independent variable in the final analysis).

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

404

RENZULLI/EVANS

respectively).

8

These districts are somewhat unusual as they have charter schools, an educa- tional innovation, and also high nonwhite student enrollment. Thus, it appears that district- level student demographics may be relevant in the development of charter schools, but that relationship is beyond the scope of these analyses and data (but see Renzulli 2002, 2005).

Dependent Variable

Our dependent variable is the percentage of white students in a charter school. Compar- ing mean percent white across our sample, we find that charter schools enroll the same per- cent of white students, on average, as the districts in which they are located (48.12 percent for charter schools and 48.46 percent for districts).

However, averages can be misleading. Instead of simply looking at means, we must examine the entire distribution of charter schools and districts to provide a richer picture of the distribution of these two populations. When we do, our data confirm what Erica Frankenberg and Chungmei Lee (2003) have shown: the percent white in many charter schools is often greater than the percent white in their respective districts. At the same time, however, the percent white in many charter schools is

lower

than in the schools’ respective districts. The average percent white, thus, com- prises schools with very high or very low percent white students. The average mirrors the dis- trict-wide percent white, even as the charter schools individually vary from the average.

Integration and Contact

Although the dependent variable was collected in 1999–2000, our key independent vari- ables were collected in 1998–1999. By using 1998–1999 school year data, we have lagged mea- sures of district characteristics. This allows us to examine how racial enrollment in the district in one year can affect racial enrollment in a charter school in the following year. Such a lag structure in the data helps ensure that results reflect more than associations between simulta- neously occurring events, and strengthens our confidence in the causal nature of the processes we uncover. Table 1 reports variable definitions and descriptive statistics.

The white flight literature features two measures of racial distribution that capture racial competition and the mechanisms that lead to white flight (Clotfelter 1978, 2001; Taeuber and James 1982). The first commonly used measure of racial distribution is

contact

—the exposure rate of whites to nonwhites within their schools across a district. In light of competition the- ory and other theories of prejudice, we might assume that if whites go to school with non- whites, this contact and the resulting tensions may push whites toward homogenously white environments (Becker 1971; Olzak 1992).

More technically, Clotfelter, describes Coleman’s (1975) measure of contact as the “per- centage of nonwhite students in the average white student’s school” (quoted in Clotfelter 2001:205) within a district (see also Clotfelter 1978). Contact is calculated for each district with the following equation (multiplied by 100 for ease of interpretation):

where:

W

i

5

number of whites in a school,

N

i

5

number of nonwhites in a school, and

W

5

number of whites in a district.

Contact C( ) 1 W⁄( ) Wi Ni Wi 1 Ni( )⁄[ ] i

∑=

8. Data on percent nonwhite come from the Common Core of Data, 1998–1999. A regular school district is defined as “local school district” rather than other types defined by the CCD. Although 37 percent of all public school students are nonwhite, the average school district is 21 percent nonwhite due to the uneven distribution of nonwhites stu- dents among districts.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight

405

Contact simultaneously reflects the racial composition of the district

and

the level of exposure of whites to nonwhites in their schools. A contact value of “0” indicates that no nonwhites go to schools with white students. However, the maximum possible value of contact is achieved when all nonwhites are evenly distributed among schools in a district. The maximum pos- sible value varies from district to district because it cannot be greater than the percentage of nonwhites in the district. For example, if a district were 40 percent nonwhite and every school in the district was 40 percent nonwhite, then contact would be at its maximum pos- sible value for that district (i.e., 40). If contact were only 30 in that same district (10 below its maximum), then whites and nonwhites would have less than maximal contact, given the percentage of nonwhites in the district (for similar examples, see Armor and Rossell 2002).

The second measure—

integration

—represents the extent to which the distribution of whites and nonwhites across schools in a district mirrors the racial distribution of students in the district. This measure of integration is not dependent on the racial composition of the dis- trict (Clotfelter 1978, 2001; Coleman 1975; Taeuber and James 1982:134) because it reflects only how closely individual schools match district-wide figures, regardless of the district-wide racial composition.

Integration is calculated for each district by using the following equation:

Integration (

I

)

5

1 – [(

N

/(

W

1

N

)] –

C

)/[

N

/(

W

1

N

)]

where:

C

5

contact,

W

5

number of whites in a district, and

N

5

number of nonwhites in a district.

Integration ranges from 0 to 1 where “1” is perfect integration and “0” is perfect segregation. The meaning of integration,

I

, here differs from the colloquial usage of the word “integra- tion.” We often use the word integration to mean that a district has substantial numbers of nonwhites or a notable balance of different races. In this measure, however, a district will be perfectly integrated (i.e.,

I

5

1), as long as nonwhite students in the district are distributed evenly among the schools. A district could be perfectly integrated even if every school were one percent, 30 percent, or 80 percent nonwhite, as long as these percentages match those of corresponding districts. For example, if a school district is 30 percent nonwhite and all individ- ual schools are also 30 percent nonwhite, then it would have a perfect integration score, if all individual schools were also 30 percent nonwhite. The integration score would be less than 1 to the degree that individual schools deviate from the racial composition of the district. For example, if all the nonwhite students were in one school and all the white students were in the other schools,

I

would be “0.”

9

It is important to note that contact can be low even if integration is high. If one per- cent of a district is nonwhite and nonwhites are distributed evenly among the schools so that each school is one percent nonwhite, integration will be high despite the low per- centage of nonwhites. Contact, however, will still be low because it depends in part on the percent of nonwhite students in the district. In a district with one percent nonwhites, the measure of contact cannot achieve a value greater than one, even if nonwhites are distributed evenly.

As shown by the descriptive statistics in Table 1, the mean integration score in our sample of districts is high—at about 91—indicating that the racial composition of schools, on

9. Though readers may be more familiar with segregation, we subtracted one from segregation,

S

, to create our measure of integration

I

, in order to remain consistent with our hypothesis. Moreover, using

I

instead of

S

facilitates easier interpretation, as a positive coefficient for integration indicates an increase in white enrollment at charter schools. Finally, we multiplied integration by 100 for ease of interpretation in our models.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

406

RENZULLI/EVANS

T a b

le 1

•

V a ri

a b le

D es

cr ip

ti o n

s a n

d M

et ri

c (

N

5

6 8 8 )

D es

cr ip

ti on

M ea

n S D

M et

ri c

D e p

e n

d e n

t v

a ri

a b

le

P e rc

e n

t w

h it

e i n

C S

C o n

ti n

u o u

s v a ri

a b le

: P e rc

e n

t o f

w h

it e s

tu d e n

ts i n

a

ch a rt

e r

sc h

o o l.

4 8 .1

2 3 7 .2

2 0 .0

0 —

1 0 0 .0

0 N

o n

w h

it e : B

la ck

, H

is p a n

ic , A

si a n

/P a ci

fi c

Is la

n d e r,

A m

e ri

ca n

-I n

d ia

n /A

la sk

a n

In d

e p

e n

d e n

t v

a ri

a b

le s

M e a su

re s

o f

d is

tr ic

t ra

ci a l d is

tr ib

u ti

o n

In te

g ra

ti o n

C o n

ti n

u o u

s v a ri

a b le

: D

is tr

ib u

ti o n

o f

w h

it e a

n d n

o n

w h

it e s

tu d e n

ts

a cr

o ss

s ch

o o ls

i n

a d

is tr

ic t

(s e e C

lo tf

e lt

e r

2 0 0 0 :2

0 6 )

9 1 .4

4 9 .0

2 4 7 .2

5

—

1 0 0 .0

0

C o n

ta ct

C o n

ti n

u o u

s v a ri

a b le

: P e rc

e n

t o f

n o n

w h

it e s

tu d e n

ts i n

a v e ra

g e w

h it

e s

tu d e n

t’ s

sc h

o o l (s

e e C

lo tf

e lt

e r

2 0 0 0 :2

0 6 )

4 6 .2

2 2 9 .4

0 0 .5

7

—

9 9 .7

7

D is

tr ic

t a n

d s

ta te

c h

a ra

ct e ri

st ic

s

P e rc

e n

t b e lo

w t

h e p

o v e rt

y l in

e C

o n

ti n

u o u

s v a ri

a b le

: P e rc

e n

t o f

st u

d e n

ts i n

t h

e d

is tr

ic t

w h

o a

re

b e lo

w t

h e p

o v e rt

y l in

e i n

1 9 9 8 – 1 9 9 9 s

ch o o l y e a r.

1 9 .2

0 1 3 .3

8 0 .0

0

—

6 8 .2

0

A ca

d e m

ic q

u a li ty

C o n

ti n

u o u

s v a ri

a b le

: P e rc

e n

t o f

st u

d e n

ts i n

t h

e d

is tr

ic t

th a t

a t

m in

im u

m p

a ss

t h

e ir

s ta

n d a rd

iz e d t

e st

s (s

e e a

p p e n

d ix

1 )

5 7 .1

0 1 9 .6

9 5 .6

7

—

9 7 .6

7

R e g io

n D

u m

m y v

a ri

a b le

s: r

e fe

re n

ce g

ro u

p

5

n o rt

h e a st

(b a se

d o

n C

e n

su s

B u

re a u

R e g io

n s)

M id

w e st

1

5

st a te

s w

it h

in M

id w

e st

r e g io

n , 0

5

o th

e r

0 .2

4

S o u

th 1

5

st a te

s w

it h

in S

o u

th r

e g io

n , 0

5

o th

e r

0 .2

7

W e st

1

5

st a te

s w

it h

in W

e st

r e g io

n , 0

5

o th

e r

0 .4

0

R a ce

l e g is

la ti

o n

f o r

C S

D u

m m

y v

a ri

a b le

: 1

5

st a te

c h

a rt

e r

sc h

o o l le

g is

la ti

o n

s p e ci

fi e s

d is

tr ic

t ra

ci a l p ro

p o rt

io n

a li ty

, 0

5

n o t

st a te

d o

r n

o c

la u

se 0

.3 7

# o

f ch

a rt

e r

sc h

o o ls

i n

d is

tr ic

t C

o n

ti n

u o u

s v a ri

a b le

: N

u m

b e r

o f

o p e ra

ti n

g c

h a rt

e r

sc h

o o ls

i n

a d

is tr

ic t.

4 .2

3 6 .3

1 0 .0

0 —

2 6 .0

0

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight

407

C h

a rt

e r

sc h

o o l ch

a ra

ct e ri

st ic

s

U rb

a n

s ch

o o l d is

tr ic

t D

u m

m y v

a ri

a b le

: 1

5

c e n

tr a l ci

ty (

m in

im u

m n

u m

b e r

o f

p e o p le

e q u

a l o r

g re

a te

r th

a n

5 0 ,0

0 0 w

it h

a C

e n

su s

u rb

a n

iz a ti

o n

c o d e )

0

5

se rv

e s

a n

M S A

b u

t n

o t

a c

e n

tr a l ci

ty , o r

d o e s

n o t

se rv

e a

n M

S A

0 .5

7

A g e o

f ch

a rt

e r

C o n

ti n

u o u

s v a ri

a b le

: A

g e o

f ch

a rt

e r

sc h

o o l

(1 9 9 8 – Y

e a r

ch a rt

e r

w a s

o p e n

e d )

2 .5

9 1 .5

0 1 .0

0 —

8 .0

0

C h

a rt

e r

sc h

o o l cu

rr ic

u lu

m D

u m

m y v

a ri

a b le

s: R

e g u

la r

e le

m e n

ta ry

o r

se co

n d a ry

c h

a rt

e r

sc h

o o l is

t h

e r

e fe

re n

ce g

ro u

p .

S p e ci

a l cu

rr ic

u la

r fo

cu s

1

5

sp e ci

a l cu

rr ic

u la

r fo

cu s

(a ca

d e m

ic ),

0

5

o th

e r

0 .0

2

S p e ci

a l e d u

ca ti

o n

1

5

c h

a rt

e r

sc h

o o l d e si

g n

e d f

o r

sp e ci

a l e d u

ca ti

o n

, 0

5

o th

e r

0 .2

3

T e ch

e d u

ca ti

o n

1

5

te ch

n ic

a l/

v o ca

ti o n

a l tr

a in

in g s

ch o o ls

, 0

5

o th

e r

0 .0

1

A lt

. 1

5

a lt

e rn

a ti

v e t

y p e c

u rr

ic u

lu m

, 0 5

o th

e r

0 .1

9

E le

m e n

ta ry

D

u m

m y v

a ri

a b le

: 1 5

c h

a rt

e r

sc h

o o l se

rv e s

K – 5 ,

0 5

c h

a rt

e r

sc h

o o l se

rv e s

6 – 1 2

0 .8

3

S ta

rt -u

p s

ta tu

s D

u m

m y v

a ri

a b le

s: r

e fe

re n

ce g

ro u

p 5

p u

b li c

sc h

o o l co

n v e rs

io n

.

N e w

s ch

o o l

D u

m m

y v

a ri

a b le

: 1 5

c h

a rt

e r

sc h

o o l is

a n

e w

s ch

o o l, 0

5 ch

a rt

e r

sc h

o o l

co n

v e rt

e d f

ro m

p u

b li c

sc h

o o l o r

p ri

v a te

s ch

o o l

0 .7

4

C o n

v e rt

e d p

ri v a te

s ch

o o l

D u

m m

y v

a ri

a b le

: 1 5

c h

a rt

e r

sc h

o o l is

c o n

v e rt

e d f

ro m

a n

e x is

ti n

g p

ri v a te

s ch

o o l, 0

5 c

h a rt

e r

sc h

o o l

co n

v e rt

e d f

ro m

p u

b li c

sc h

o o l o r

is a

n e w

s ch

o o l

0 .1

0

P e rc

e n

t o f

st u

d e n

ts e

li g ib

le f

o r

sc h

o o l lu

n ch

C o n

ti n

u o u

s v a ri

a b le

: P e rc

e n

t o f

st u

d e n

ts i n

t h

e c

h a rt

e r

sc h

o o l

e li g ib

le f

o r

sc h

o o l lu

n ch

— re

d u

ce d o

r fr

e e .

4 2 .7

9 3 3 .0

5 0 .0

0 —

1 0 0 .0

0

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

408 RENZULLI/EVANS

average, matches the racial composition of the districts in which they are located.10 These find- ings, in combination with Clotfelter’s (2001), imply that greater segregation between dis- tricts than within districts (Clotfelter 2001; Orfield, Frankenberg, and Lee 2003).

District and State Attributes

In addition to race, class also appears to be an important, albeit seldom examined, component of white flight. To ensure that the white flight we hypothesized is due to racial preferences and not class preferences, we control for the poverty rate of the district or, more specifically, the percent of students below the poverty line.

Another important control we utilize is academic quality in the feeding school district. We operationalize this as the percent of students passing standardized achievement tests. White parents often claim that low quality of education, not racial demographics, influ- enced their decisions to move schools (Goldring and Hausman 1999; Kleitz et al. 2000; Schneider, Teske, and Marschall 2000). Without a control for quality in the district, it would be difficult to delineate whether race matters because of racial preferences/prejudices or because race tends to be associated with school quality (Condron and Roscigno 2003).

In much of the white flight and racial competition literature, region of the country plays an important role. Researchers have found a greater tendency toward white flight in the Northeast relative to the South (Clotfelter 2001; Rivkin 1994). Region, measured by Census Bureau state classification, can capture any unmeasured heterogeneity that might be present due to historical residential and educational segregation.

White enrollment in charter schools may also be sensitive to racial policies. Research on magnet schools suggests that policy can affect who enrolls. Salvatore Saporito’s (2003) research on magnet schools in one city suggests that when minority enrollments of magnet schools are unregulated, the racial patterns mirror the racial composition of neighborhoods. Charter schools may be no different. Charter school legislation varies significantly by state, but within a state, the legislative guidelines consistently apply to all charter schools. Half the states in our data have a clause in their law that requires charter schools to comply with desegregation policies or to create racial balance per the district in which it is located (about 37 percent of charter schools are in states with a race clause) (Jennings et al. 1998). Fran- kenberg and Lee (2003) suggest that more research on the effects of state regulations on charter school composition at the district level is needed. We fill that gap by including a measure of a race regulation clause.

Finally, districts with many charter schools may have different patterns of racial enroll- ment than districts with only one or few charter schools. In fact, Rossell (2003) found that as the percentage of magnet schools in a voluntary desegregation plan increased and as schools of choice became part of the fabric of the district, they became more racially similar to regular schools. Therefore, we control for the number of charter schools in a district.

Charter School Characteristics

Including controls for charter school characteristics is also necessary, particularly if par- ents are making decisions and actively choosing a school for their children. Attributes of schools may influence the types of students that get drawn to or are targeted by a school, which likely would affect white enrollment. Urbanicity is especially important for charter

10. Our integration scores differ from Clotfelter’s (2001) because his measure of integration was calculated for metropolitan areas not individual school districts. Clotfelter found integration in the Detroit Metropolitan Area was .27. We find that the Detroit School District proper is more integrated (I5.60). Though using metropolitan areas is appropri- ate for Clotfelter’s analysis and research question, we are concerned with how districts (not metropolitan areas) are inte- grated and therefore use individual districts.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight 409

schools and white flight because those who change schools may be more likely to reside in urban areas where schools have a reputation for being subpar (Howard 2003; Roscigno and Condron 2003). If that is the case, charter schools that serve urban areas may have differ- ent enrollment patterns than those that serve suburban and rural areas.

We control for the number of years the charter school has been operating to account for the possible networks and reputation it may gain, which in turn may affect the racial composition of the school. If access to charter schools is largely by word of mouth, then we would expect charter schools to be fairly homogenous racially. If affluent white mothers share information about charter schools with each other, as Cobb and Glass (1999) found, then we would expect charter schools to be relatively more white the longer the charter school has been open.

Charter schools vary in their function and form (Lin 2001), and ignoring these differ- ences can hinder our evaluation of their racial enrollment. The type of curriculum and the type of school may affect the likelihood of having minority enrollment (Cobb and Glass 1999). Although charter schools cannot have selective admission requirements,11 they can specify their mission and establish unique curricula. Some charter schools provide special- ized educational curricula such as marine biology or fine arts; others provide “Direct Instruction” and “Back to the Basics” curricula. Still others are schools for special education students. Nonwhites may be attending charter schools designed for special education stu- dents and not attending schools with college preparation as their focus, as was found in Arizona (Cobb and Glass 1999) and Texas (Wamba and Ascher 2003).12 See Table 1 for a description of the four dummy variables: special program curriculum, special education, technical education, and alternative education (regular charter schools is the reference category).

Variation in grade level and in conversion status of a charter school may affect white enrollment. Elementary schools tend to be more homogenous than high schools. Charter schools converted from private schools may tend to be more segregated because private schools typically have higher white enrollment than traditional counterparts (Fairlie and Resch 2002). Conversion schools typically must give preference to those who already attended the school, and schools themselves are not always representative of the district’s overall demographics, as in California (Wohlstetter et al. 2002).

In addition, the white flight literature suggests that the socioeconomic status associ- ated with the destination school is also an important factor in decisions of a residential or educational move (Andrews 2002; Clotfelter 2001). The socioeconomic status of the charter school—as measured by the percent of students eligible to receive reduced or free lunch— may contribute to the racial composition of the school, net of district characteristics.13

Analytic Strategy and Results

The charter schools in our data are not independent because 475 of the 688 charter schools are clustered within 123 districts. Without adjusting the modeling, this characteristic of the data would violate the assumption of independence in ordinary least squares regres-

11. The SASS data have information on charter school admission requirements such as the use of admissions tests, standardized tests, special talents, letters of recommendation, interviews, etc. We considered using these as control variables, but we found neither a significant effect on the percent white nor a change in our models’ overall fit. More research on admission requirements is necessary, however, beyond the scope of this work.

12. We prefer to measure “Afro-centric” programs as one school type because such a curriculum may attract more black students than other curricular programs. However, no respondents in the SASS data described their schools’ cur- riculum as Afro-centric.

13. The eligibility for school lunch was a reliable measure in the Schools and Staffing Survey for charter schools. A measure of the percent of students who are below the poverty line (comparable to our district-level measure) is not available. The students who are eligible for school lunch are 185 percent or more below the poverty line.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

410 RENZULLI/EVANS

sion and bias the standard errors. We thus correct for correlated errors by using robust clus- tering in STATA. We also tested for multicollinearity by examining variance inflation factors: all were far below the accepted value of 10, and none was above three in all our models.

We have argued that district demographics are not enough to understand the context in which charter school enrollment emerges because they do not reveal the racial patterns that competition theory would suggest cause conflict between groups—whites and nonwhites attending schools together. We expect that the racial make-up of schools and distribution of students will affect white enrollment at charter schools, and we test this hypothesis by regress- ing the percent white in a charter school on our two measures of racial distribution in the dis- trict: integration and contact. Our analyses proceed in three steps. First, we introduce integration—our main variable of interest—and contact in the model. Then we add district and state controls. In our final model, we add charter school controls.

Results

As three generations of research on white flight shows, residential mobility allows whites to avoid nonwhites in their schools. We find that in the current era of school choice, charter schools are also a white flight option. In fact, we find that relative integration is a significant determinant of white enrollment in charter schools. Table 2, Model 1 shows that as the level of integration increases in a school district, the percent of white students in local charter schools increases as well. Whites’ and nonwhites’ equal or proportional distri- bution among schools positively affects white enrollment in charter schools, net of con- tact. One-percent increases in integration introduce a .73 percentage point increase in the percent of white students in a charter school.14 These findings suggest that when districts do not have schools that serve as white enclaves, a greater percent of white students will attend local charter schools.

The effect of our second measure of racial distribution—contact—is contrary to that which we expected. Charter schools have smaller percentages of whites when the schools are located in districts where whites have more contact with nonwhites. Given past research, we would expect whites with more contact with nonwhites to choose charter schools, but our data do not show this. We believe, however, that the negative coefficient for contact is consistent with the nature of the measure, the location of charter schools, and our findings regarding integration. Contact increases as the numbers and proportions of nonwhites in a district increase (Armor and Rossell 2002; Clotfelter 2001; Taeuber and James 1982). Contact, therefore, partially reflects the percentage of nonwhites in a district. As the percentage of nonwhites in a district increases, the percent of nonwhites in charter schools will be driven up, and the percent of whites who attend them will be driven down as a function of the pool from which the charter schools draw students (i.e., increased con- tact implies a corresponding increase in the pool of nonwhite students in a district and eligi- bility to enroll in a charter school). This statistical pattern, however, does not affect the contribution of our other measure of racial distribution.15 Integration continues to predict an increase in the percent white in charter schools, despite the negative effect of contact.16

14. We also created and tested models with an index of dissimilarity score. Results were consistent across models; therefore, we choose to use the integration measure since it consistent with the research on white flight (Armor and Rossell 2002; Clotfelter 1978, 2001; Coleman 1975).

15. We also estimated two additional models, one with integration only and the other with contact only, to see if including both in the same model affected their coefficients. We found similar results (positive for integration and nega- tive for contact at the .05 level of confidence, two-tailed tests). Thus, we can conclude that the two variables did not affect each other in the model.

16. In our final model, after the specification is complete, we control for the percent of nonwhites in the district. We find that it has a negative effect on white enrollment, and our integration hypothesis is still confirmed.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight 411

To ensure that our models were not affected by misspecification or unobserved hetero- geneity, we controlled for a number of factors at the state and district levels in Models 2 and 3. Several noteworthy relationships emerge, although integration notably remains a robust and significant predictor (see Table 2, Model 2).

As other research on racial integration has found, region of the country plays a role in the percent of whites who go to charter schools, controlling for other state and district char- acteristics (Table 2, Model 2). Charter schools in the West, South, and Midwest have a higher percent of white enrollment than do schools in the Northeast. According to Steven G. Rivkin (1994), from 1968 to 1988 the percentage of whites with no black schoolmates in the North-

Table 2 • Unstandarized OLS Estimates of Racial Distributions and District and Charter School Characteristics on Charter School White Enrollment

Model 1 Model 2 Model 3

Coef SE a

Coef SE Coef SE

Measures of district race distribution Integration 0.73* 0.14 0.65* 0.14 0.60* 0.13 Contact 20.86* 0.03 20.82* 0.05 20.65* 0.05

District/state characteristics Percent of student below the poverty line — — 20.17** 0.09 0.04 0.09 Academic quality — — 20.12 0.09 20.14*** 0.08

Region (ref 5 Northeast) Midwest — — 12.29** 4.96 10.02** 4.29 South — — 12.60** 5.91 12.06** 5.06 West — — 18.55* 4.15 14.22* 3.62

Race legislation for CS — — 22.88 2.29 22.68 2.03 Number of charter schools in district — — 20.10 0.21 20.15 0.19

Charter school characteristics Charter school serves urban area

(ref 5 suburban and rural) — — — — 23.21 2.26

Number of years charter has operated — — — — 0.76 0.64 Curriculum

Special curricular focus CS — — — — 14.75** 6.95 Special education CS — — — — 1.60 2.34 Tech ed CS — — — — 4.53 9.78 Alt. CS — — — — 2.67 2.44

Elementary CS (ref: middle and high school) — — — — 4.11*** 2.16

Startup status (ref 5 public conversion) New school — — — — 24.52** 2.19 Converted from private school — — — — 23.08 4.35

Percent eligible for school lunch — — — — 20.37* 0.04 Constant 20.57 13.39 25.30 12.66 35.30* 12.12

N 688 688 688

Number of clusters (districts) 338 338 338

F 407 135 114

R2 0.55 0.57 0.66

a All standard errors are calculated using robust standard errors.

* p , .01, ** p , .05, *** p , .1; two-tailed.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

412 RENZULLI/EVANS

east increased from 5.9 percent to 7.4 percent. In fact, the white flight literature has shown more white flight in the Northeast than in other regions (Rivkin 1994). We speculate that whites may have already fled to other school destinations, and therefore we see a lower percent of white students in charter schools in the Northeast. It is also a possible that charter schools in the Northeast do a better job teaching white and nonwhite students in their charter schools. Charter schools may not be created equal across regions and more research should consider the effects of past educational policies—both de jure and de facto—across U.S. regions, to evaluate and understand charter school enrollment.

Models 2 and 3 also show that the racial distribution of charter school enrollment does not depend on the academic quality of school districts (i.e., district-wide test scores). People may be using academic rationalizations to justify moving their children to charter schools, but there is little evidence to support this. Even when controlling for academic quality, we find that integration remains statistically significant. In the conclusion, we will discuss the implications of finding that academic quality is not a significant predictor of white enroll- ment in charter schools.

In our final model, we control for a host of charter school-level characteristics. This model confirms that, even when controlling for charter school characteristics that may either draw or divert white students, levels of integration and the threats of racial competition that they arguably breed continue to be a significant predictor of whites’ behavior (see Table 2, Model 3). Nevertheless, the reduction in the size of the coefficient does indicate that char- ter school white enrollment is partly associated with charter school characteristics.

In addition to the level of integration in school districts, studies of white flight also examined the receiving districts’ desirability. That is, parents assess their current district for desirability along with the desirability of the one to which they will move their children. We control for the desirability of the charter school (the receiving schools) by including the type of charter school and the poverty level of the charter school. We find that charter schools for special education students positively affect white enrollment. As we discuss in the conclusion, parents of white special needs students may be especially likely to send their children to charter schools because special education classrooms in public schools tend to have higher proportions of nonwhites than the schools in which they are located (Losen and Orfield 2002). Therefore, white parents of children with special needs may search for options that remove their children from highly nonwhite classrooms. In addition, the greater the percentages of students who are eligible for subsidized lunch, the lower the percentage of white students who attend the charter school. This result indicates that charter schools with more white students tend to have lower poverty rates, which may reflect the demo- graphic fact that white students are less likely to be poor than nonwhite students (U.S. Census Bureau 2002).

Ultimately, we conclude that whites move to charter schools when whites and non- whites are distributed proportionally among their district’s schools. In fact, even when whites are the majority in a school district and among their schools, whites who attend schools with nonwhites continue to look for options that are even more white. These models can be interpreted within the competition theory framework, which suggests that nonwhite enrollment in public schools is perceived as a threat to the status of whites in the same schools and that whites will escape the threat by retreating to white schooling options. Integration remains an important predictor of the percent of whites who attend a charter school net of state policy and other district and charter school characteristics.

Conclusion

We have used new data on charter schools to better understand which student popula- tions they serve and where they are served. We have learned that white flight persists

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight 413

within our public school system and that racially competitive processes occur within the contemporary educational process. These processes have been linked to employment dis- crimination (Tomaskovic-Devey 1993), historical battles over bussing (Olzak et al. 1994), and the social control of racial and ethnic minority groups (Tolnay, Beck, and Massey 1992). Our work has implications for theories of white flight and for our efforts to under- stand how school choice affects racial integration. Our work also has policy implications. Phrased most starkly, charter schools left to their own devices may promote racial segrega- tion in the public schools.

Researchers have argued that relative integration is as important as composition for racial competition (Blau 1977; Quillian 2002). Our work confirms that hypothesis. We offer a theoretical rationale for why racial integration is such an important predictor of the “whiteness” of charter schools. Racial competition over school status seems to contribute to white families’ decisions to send their children to charter schools. The very title “charter schools” suggests to some an educational option that is accountable, autonomous, and innovative (Sarason 1998). Thus, charter schools may have a high status connotation to some parents. Conversely, nonwhite enrollment in regular public schools may decrease the status of the schools. Therefore, one way to interpret our results is that competition over school status finds whites avoiding their nonwhite counterparts.

Racial integration has been a key element in white flight and does indeed remain a key factor in the racial composition of charter schools and other schools of choice (Saporito and Lareau 1999). Research finds that integration in school districts affects people’s deci- sions to move residences or to move their children to another school (Fairlie 2002; Gold- ring and Hausman 1999; Saporito and Lareau 1999; Swanson and Schneider 1999). Our research shows that the same patterns exist for movement into charter schools. The inte- gration level of the district affects the whiteness of a charter school even after we control for academic achievement and other factors. Thus, we conclude that charter schools pro- vide a public school option for white flight without the drawbacks of residential mobility such as job change or commute for parents (Clotfelter 2001). Although some research sug- gests that whites are “under-enrolled in both magnet and charter schools” (Orfield et al. 2003:18), we see that the degree of integration of schools in the district positively affects white enrollment in charter schools.

Three of the control variables in the analysis are worthy of more discussion and future research. First, curricula directed toward special educational students have a strong effect on the whiteness of a charter school. Whites may be reacting to the racial composition of their child’s special education classroom. Nonwhite students are more likely to be placed in special education classes than are their white counterparts (Losen and Orfield 2002). Thus, whites in special education classes may in fact have more interaction with nonwhite stu- dents than white students who are not in special education classes. This interaction may encourage parents to find white options for their children with special needs. In addition, special education is expensive and districts with higher per-pupil spending may be more white and better able to afford to create a charter school for special education. Both these explanation are empirically testable; however, the data we use (i.e., SASS) do not provide the proper measures. Future research on the type of charter school and racial enrollment is clearly warranted.

Second, we do not find a significant association between a race-based clause and white enrollment in charter schools. Even states that have a race clause on the books might not enforce such regulations. Some research, in fact, has shown that the race provisions, when they do exist, are not enforced. For example, California charter schools were not meeting district requirements, but none had been closed for failing to enroll the correct percentage of nonwhite students (Wohlstetter et al. 2002). Therefore, the magnitude and significance of the effect of race legislation on white enrollment in our research might be even greater if charter school laws dealing with race were more strictly imposed. More research is needed

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

414 RENZULLI/EVANS

on the effects of legislation on charter school outcomes including racial enrollment in order to better understand how legislation can shape schooling.

Third, academic quality does not seem to affect white enrollment in charter schools. Though the data we use here cannot test the notion of “creaming”—pulling the best students from dis- tricts into charter schools—the results (or lack of) may lead us to other research questions. For example, are the most academically inclined white students leaving public schools for charter schools? Is it possible that those white students who are going to charter schools had high test scores—thus raising the district average—but their leaving the district for a charter school would decrease the district average? On the other hand, it is possible that academic quality is not the most important factor in determining whether white or nonwhite parents send their children to a charter school. These data imply that racial integration, not poor academics, in the feeding district is associated with an increase in white enrollment in charter schools.

Our analysis has ruled out explanations of white flight based on one objective measure of school quality: whites do not seem to be fleeing schools because of weak academic quality as measured in standardized test scores. However, other dimensions of quality—such as safety, library resources, and teacher quality—could also account for the competitive process between nonwhites and whites in public schools. Future research should examine these issues.

Our study has several limitations. Although we have lagged district data, which allows us to make arguments about the effects of districts on charter schools, our charter school data are cross-sectional. This limits our ability to assess causality in the relationships between char- ter school characteristics and white enrollment. Charter school characteristics might reflect the degree to which charter schools have white enrollment and, thus, might explain white enrollment better than can district integration. Longitudinal data on charter schools and charter school enrollment are needed. Furthermore, if a charter school is primarily white, white parents may be more likely to send their child to the school than if it is a racially hetero- geneous charter school. As more whites send their children to the school, the proportion of white enrollment increases. The present data only allow us to speculate on causal order because they do not have longitudinal charter school composition data. Of course, our data do not allow us to test how the whiteness of a charter school affects subsequent racial enrollment in that charter school. This line of inquiry might help explain white flight into charter schools as well as how the operational age of a charter school affects its racial composition.

Finally, the measure of contact does not measure the degree to which students are inter- acting with nonwhite students in their classes and daily activities. We suspect—both theo- retically (Allport 1954) and empirically (Khmelkov and Hallinan 1999; Oakes 1995)—that interaction in classrooms may encourage whites to leave schools where they attend classes together. However, classroom homogeneity is not captured in this measure; therefore, we cannot rule out the possibility that interaction of whites and nonwhites at the classroom level might encourage whites to attend charter schools.

The history of racial segregation (Massey and Denton 1993) and separatism in education (Kaestle 1983) in the United States requires that research on charter schools focus on the con- sequences for racial equity in our schools. By using theories of school segregation, racial com- petition, and known determinants of white flight, this article extends the understanding of charter school racial composition. Nevertheless, more research on charter schools is needed to clarify their role in the more general resegregation of districts (Orfield et al. 2003).

References

Alexander, Karl L. 1997. “Public Schools and Public Good.” Social Forces 76:1–30. Allport, Gordon W. 1954. The Nature of Prejudice. Reading, MA: Addison-Wesley.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight 415

Andrews, Kenneth T. 2002. “Movement-Countermovement Dynamics and the Emergence of New Insti- tutions: The Case of ‘White Flight’ Schools in Mississippi.” Social Forces 80:911–36.

Armor, David J. and Christine Rossell. 2002. “Desegregation and Resegregation in the Public Schools.” Pp. 219–58 in Beyond the Color Line: New Perspectives on Race and Ethnicity in America, edited by Abigail Thernstrom and Stephan Thernstrom. Stanford, CA: Hoover Institution Press.

Bankston, Carl L., III and Stephen J. Caldas. 2002. A Troubled Dream: The Promise and Failure of School De- segregation in Louisiana. Nashville: Vanderbilt University Press.

Beck, E. M. and Stewart E. Tolnay. 1990. “The Killing Fields of the Deep South: The Market for Cotton and the Lynching of Blacks, 1882–1930.” American Sociological Review 55:526–39.

Becker, Gary. 1971. The Economic of Discrimination. Chicago: University of Chicago Press. Berliner, David C. and Bruce J. Biddle. 1995. The Manufactured Crisis: Myths, Fraud, and the Attack on Amer-

ica’s Public Schools. Reading, MA: Addison-Wesley. Blau, Peter M. 1977. Inequality and Heterogeneity: A Primitive Theory of Social Structure. New York: Free Press. Bulman, Robert C. 2004. “School Choice Stories: The Role of Culture.” Sociological Inquiry 74:492–519. Center for Education Reform. 2000. National Charter School Directory, edited by Angela Dale. Washington,

DC: Center for Education Reform. Center for Educational Reform. 1999. “FAQs.” Center for Educational Reform: Making Schools Work Better for

All Children. Retrieved April 6, 2004 (http://edreform.com/school_reform_faq/charter_schools.htm). Chubb, John E. and Terry M. Moe. 1990. Politics, Markets and American Schools. Washington, DC: The

Brookings Institute. Clotfelter, Charles T. 1978. “Alternative Measures of School Desegregation: A Methodological Note.”

Land Economics 54:373–80. ———. 2001. “Are Whites Still Fleeing? Racial Patterns and Enrollment Shifts in Urban Public Schools,

1987–1996.” Journal of Policy Analysis and Management 20:199–221. Cobb, Casey D. and Gene V. Glass. 1999. “Ethnic Segregation in Arizona Charter Schools.” Education Policy

Analysis Archives 7(1). Retrieved April 6, 2005 (http://epaa.asu.edu/epaa/v7n1/). Coleman, James. 1975. “School Desegregation and the Loss of Whites from Large Central-city School

Districts.” Pp. 96–135 in School Desegregation: The Courts and Suburban Migration, edited by the U.S. Commission on Civil Rights. Washington, DC: U.S. Government.

Coleman, James S., Sara D. Kelly, and John A. Moore. 1975. Trends in School Segregation, 1968–1973. Urban Institute: Washington, DC.

Condron, Dennis J. and Vincent. J. Roscigno. 2003. “Disparities Within: Unequal Spending and Achieve- ment in an Urban School District.” Sociology of Education 76:18–36.

Coons, John. E. and Stephen Sugarman. 1978. Education by Choice: The Case for Family Control. Berkeley: University of California Press.

Crew, Robert E. and Mary R. Anderson. 2003. “Accountability and Performance in Charter Schools in Florida: A Theory-Based Evaluation.” American Journal of Evaluation 24:189–212.

Crowder, Kyle. 2000. “The Racial Context of White Mobility: An Individual-level Assessment of the White Flight Hypothesis.” Social Science Research 29:223–57.

Fairlie, Robert W. 2002. “Private Schools and ‘Latino Flight’ from Black Schoolchildren.” Demography 39:655–74.

Fairlie, Robert W. and Alexandra M. Resch. 2002. “Is There ‘White Flight’ into Private Schools? Evidence from the National Educational Longitudinal Survey.” Review of Economics and Statistics 84:21–33.

Farley, Reynolds. 1975. “Racial Integration in the Public Schools, 1967 to 1972: Assessing the Effects of Governmental Policy.” Sociological Focus 8:1–26.

Fowler, Frances. 2003. “School Choice: Silver Bullet, Social Threat, or Sound Policy?” Educational Re- searcher 32:33–39.

Frankenberg, Erica and Chungmei Lee. 2003. “Charter Schools and Race: A Lost Opportunity for Inte- grated Education.” Educational Policy Analysis Archives 11 (32). Retrieved April 6, 2005 (http://epaa.asu. edu/epaa/v11n32/).

Frey, William H. 1979. “Central City White Flight: Racial and Non-Racial Causes.” American Sociological Review 44:425–48.

———. 1995. “Immigration and Internal Migration ‘Flight’ from U.S. Metropolitan Areas: Toward a New Demographic Balkanization.” Urban Studies 32:733–58.

Fusarelli, Lance D. 2002. “Texas: Charter Schools and the Struggle for Equity.” Pp. 175–91 in The Charter School Landscape, edited by Sandra Vergari. Pittsburgh, PA: Pittsburgh University Press.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

416 RENZULLI/EVANS

Giles, Michael. 1978. “White Enrollment Stability and School Desegregation: A Two Level Analysis.” American Sociological Review 43:848–64.

Gill, Brian P. 2001. Rhetoric vs. Reality: What We Know and What We Need to Know about Vouchers and Charter Schools. Santa Monica, CA: Rand Corporation.

Goldring, Ellen B. and Charles S. Hausman. 1999. “Reasons for Parental Choice of Urban Schools.” Jour- nal of Education Policy 14:469–90.

Henig, Jeffrey R., Richard C. Hula, Marrion Orr, and Desiree S. Pedescleaux. 1999. The Color of School Reform: Race, Politics, and the Challenge of Urban Education. Princeton, NJ: Princeton University Press.

Holt, Mikel. 2000. Not Yet “Free at Last”: The Unfinished Business of the Civil Rights Movements. Oakland, CA: Institute for Contemporary Studies.

Howard, Tyrone. 2003. “Who Receives the Short End of Shortage? Implications of the Teacher Shortage on Urban Schools.” Journal of Curriculum and Supervision 18:142–60.

James, David R. 1989. “City Limits on Racial Equality: The Effects of City-Suburb Boundaries on Public- school Desegregation,1968–1976.” American Sociological Review 54:963–85.

Jennings, Wayne, Eric Premack, Andrew Adelmann, and Debra Solomon. 1998. A Comparison of Charter School Legislation: Thirty-Three States and the District of Columbia Incorporating Legislative Changes through October, 1998. Washington, DC: RPP International.

Kaestle, Carl E. 1983. Pillars of the Republic: Common Schools and American Society. New York: Hill and Wang. Khmelkov, Vladimir T. and Maureen T. Hallinan. 1999. “Organizational Effects on Race Relations in

Schools.” Journal of Social Issues 55:627–46. Kleitz, Bretten, Gregory R. Weiher, Kent L. Tedin, and Richard Matland. 2000. “Choice, Charter Schools,

and Household Preferences.” Social Science Quarterly 81:846–54. Lankford, Hamilton and James Wyckoff. 2001. “Who Would Be Left Behind by Enhanced Private School

Choice?” Journal of Urban Economics 50:288–312. Levin, Betsy. 1999. “Race and School Choice.” Pp. 266–99 in School Choice and Social Controversy, edited by

Stephen Sugarman and Frank Kemerer. Washington, DC: Brookings Institute Press. Lin, Qiuyun. 2001. “An Evaluation of Charter School Effectiveness.” Education 122:166–76. Little, Roderick J. A. and Donald B. Rubin. 1987. Statistical Analysis with Missing Data. New York: Wiley. Losen, Daniel J. and Gary Orfield. 2002. Racial Inequity in Special Education. Cambridge, MA: The Civil

Rights Project at Harvard University and the Harvard Education Press. Manno, Bruno V., Gregg Vanourek, and Chester E. Finn, Jr. 2000. “The Future of Charter Schools: How

Big, How Bright?” Teaching and Change 7:222–34. Massey, Douglas S. and Nancy A. Denton. 1993. American Apartheid: Segregation and the Making of the Un-

derclass. Cambridge, MA: Harvard University Press. Miron, Gary, Chris Neslon, and John Risley. 2002. Strengthening Pennsylvania’s Charter School Reform: Findings

from the Statewide Evaluation and Discussion of Relevant Policy Issues. Kalamazoo, MI: Western Michigan University Evaluation Center.

Morgan, David and Robert England. 1984. “White Enrollment Loss: The Effects of School Desegregation in Perspective.” American Politics Quarterly 12:241–64.

Nathan, Joe. 1996. Charter Schools: Creating Hope and Opportunity for American Education. San Francisco, CA: Jossey-Bass Publishers.

National Center for Education Statistics (NCES). 2004. “Schools and Staffing Survey.” NCES Electronic Cat- alog. Washington, DC: National Center for Education Statistics.

Nelson, F. Howard, Bella Rosenberg, and Nancy Van Meter. 2004. “Charter School Achievement on the 2003 National Assessment of Educational Progress.” American Federation of Teachers. Washington, DC: U.S. Department of Education.

Oakes, Jeannie. 1995. “Two Cities’ Tracking and Within-school Segregation.” Teachers College Record 96:681– 91.

Olzak, Susan. 1990. “The Political Context of Competition: Lynching and Urban Violence, 1882–1914.” Social Forces 69:395–421.

———. 1992. The Dynamics of Ethnic Competition and Conflict. Stanford, CA: Stanford University Press. Olzak, Susan, Suzanne Shanahan, and Elizabeth H. McEneaney. 1996. “Poverty, Segregation, and Race

Riots: 1960 to 1993.” American Sociological Review 61:590–613. Olzak, Susan, Suzanne Shanahan, and E. West. 1994. “School-Desegregation, Interracial Exposure, and

Antibusing Activity in Contemporary Urban America.” American Journal of Sociology 100:196–241. Orfield, Gary, Erica D. Frankenberg, and Chumgmei Lee. 2003. “The Resurgence of School Segregation.”

Educational Leadership 60:16–20.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

School Choice, Charter Schools, and White Flight 417

Quillian, Lincoln. 2002. “Why Is Black-White Residential Segregation So Persistent?: Evidence on Three Theories from Migration Data.” Social Science Research 31:197–229.

Renzulli, Linda A. 2002. “Entrepreneurial Ambitions in the Public Sector: A Random Effects Model of the Emergence of Charter Schools in North Carolina.” Education Policy Analysis Archives 10(19). Re- trieved April 6, 2005 (http://epaa.asu.edu/epaa/v10n19/).

———. 2005. “Organizational Environments and the Emergence of Charter Schools in the U.S.” Sociology of Education 78:1–26.

Renzulli, Linda A. and Vincent. J. Roscigno. Forthcoming. “Charter School Policy, Implementation, and Diffusion in the U.S.” Sociology of Education.

Rivkin, Steven G. 1994. “Residential Segregation and School Integration.” Sociology of Education 67:279–92. Roscigno, Vincent J. and Dennis J. Condron. 2003. “When Busing Ends: Resegregation and Inequality in

an Urban School District.” Pp. 75–96 in The End of Desegregation?, edited by Carl Bankston III and Stephen J. Caldas. New York: Nova Science Publishers.

Rossell, Christine. 2003. “The Desegregation Efficiency of Magnet Schools.” Urban Affairs Review 38:697–725. Saporito, Salvatore. 2003. “Private Choices, Public Consequences: Magnet School Choice and Segrega-

tion by Race and Poverty.” Social Problems 50:181–203. Saporito, Salvatore and Annette Lareau. 1999. “School Selection as Process: The Multiple Dimensions of

Race in Framing Educational Choice.” Social Problems 46:418–39. Sarason, Seymour B. 1998. Charter Schools: Another Flawed Educational Reform? New York: Teachers College

Press. Schneider, Mark, Paul Teske, and Melissa Marschall. 2000. Choosing Schools: Consumer Choice and the Qual-

ity of American Schools. Princeton, NJ: Princeton University Press. Sly, David F. and Louis G. Pol. 1978. “The Demographic Context of School Segregation and Desegrega-

tion.” Social Forces 56:1072–86. Stevens, Mitchell L. 2001. Kingdom of Children. Princeton, NJ: Princeton University Press. Swanson, Chris B. and Barbara Schneider. 1999. “Students on the Move: Residential and Educational

Mobility in America’s Schools.” Sociology of Education 72:54–67. Taeuber, Karl E. and David R. James. 1982. “Racial Segregation among Public and Private Schools.” Soci-

ology of Education 55:133–43. Tolnay, Stewart E., E. M. Beck, and J. L. Massey. 1992. “Black Competition and White Vengeance—Legal

Execution of Blacks as Social-control in the Cotton South, 1890 to 1929.” Social Science Quarterly 73:627–44.

Tomaskovic-Devey, Donald. 1993. Gender & Racial Inequality at Work: The Sources and Consequences of Job Segregation. Ithaca: ILR Press.

Tomaskovic-Devey, Donald and Vincent. J. Roscigno. 1996. “Racial Economic Subordination and White Gain in the U.S. South.” American Sociological Review 61:565–89.

U.S. Census Bureau. 2002. “Poverty.” U.S. Census Bureau. Retrieved April 6, 2005 (http://www.census.gov/ hhes/www/poverty.html).

Vergari, Sandra. 2002. The Charter School Landscape. Pittsburgh, PA: Pittsburgh University Press. Wamba, Nathalis G. and Carol Ascher. 2003. “An Examination of Charter School Equity.” Education and

Urban Society 35:462–76. Weiher, Gregory R. and Kent L. Tedin. 2002. “Does Choice Lead to Racially Distinctive Schools?: Charter

Schools and Household Preferences.” Journal of Policy Analysis and Management 21:79–92. Wells, Amy Stuart. 1993. Time to Choose: America at the Crossroads of School Choice Policy. New York: Hill and

Wang. ———. 1998. “Charter School Reform in California—Does It Meet Expectations?” Phi Delta Kappan

80:305–12. Wells, Amy Stuart, Cynthia Grutzik, Sibyll Carnochan, Julie Slayton, and Ash Vasudeva. 1999a. “Under-

lying Policy Assumptions of Charter School Reform: The Multiple Meanings of a Movement.” Teachers College Record 100:513–35.

Wells, Amy Stuart, Alejandra Lopez, Janelle Scott, and Jennifer Jellison Holme. 1999b. “Charter Schools as Postmodern Paradox: Rethinking Social Stratification in an Age of Deregulated School Choice.” Harvard Educational Review 69:172–204.

Wilson, William Julius. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chi- cago: University of Chicago Press.

Wohlstetter, Priscilla, Noelle C Griffin, and Derrick Chau. 2002. “Charter Schools in California: A Bruis- ing Campaign for Public School Choice.” Pp. 32–53 in The Charter School Landscape, edited by Sandra Vergari. Pittsburgh, PA: Pittsburgh University Press.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018

418 RENZULLI/EVANS

Appendix

Table A1 • Test Score Data Sources

Score Type

State NPR* PP** Test Type Grade

AK X X CALIFORNIA ACHIEVEMENT TEST 5 4 AZ X STANFORD ACHIEVEMENT TEST 9 4 CA X X STANFORD ACHIEVEMENT TEST 9 4 CO X CSAP, COLORADO STUDENT ASSESSMENT PROGRAM 4 CT X CMT, CONNECTICUT MASTERY TEST 4 DC X X STANFORD ACHIEVEMENT TEST 9 4 DE X DSTP, DELAWARE STUDENT TESTING PROGRAM 4 FL X FCAT, FLORIDA COMPREHENSIVE ASSESSMENT TEST 4 GA X STANFORD ACHIEVEMENT TEST 9 3 HI X STANFORD ACHIEVEMENT TEST 9 3 IL X ITBS, IOWA TEST OF BASIC SKILLS 4 KS NO TEST DATA AVAILABLE — LA X LEAP, LOUISIANA EDUCATIONAL ASSESSMENT PROGRAM 4 MA X MCAS, MASSACHUSETTS COMPREHENSIVE ASSESSMENT PROGRAM 4 MI X MEAP, MICHIGAN EDUCATIONAL ASSESSMENT PROGRAM 4 MN X MCA, MINNESOTA COMPREHENSIVE ASSESSMENTS 3 NC X NCEOGT, NORTH CAROLINA END OF GRADES TEST 4 NJ X ESPA, ELEMENTARY SCHOOL PROFICIENCY ASSESSMENT 4 NM X NMAAP, NEW MEXICO ACHIEVEMENT ASSESSMENT PROGRAM 4 OH X ACS, ACADEMIC CONTENT STANDARDS 4 PA X PSSA, PENNSYLVANIA SYSTEM OF SCHOOL ASSESSMENT 5 RI X NSRE, NEW STANDARDS REFERENCE EXAMINATION 4 SC X PACT, PALMETTO ACHIEVEMENT CHALLENGE TESTS 4 TX X TAAS, TEXAS ACHIEVEMENT OF ACADEMIC SKILLS 4 WI X WSAS, WISCONSIN STUDENT ASSESSMENT SYSTEM 4

* National percentile. ** Percent proficient and above indicates the data provided information on how many students were considered to have passed the tests.

Downloaded from https://academic.oup.com/socpro/article-abstract/52/3/398/1631762 by CUNY Graduate Center user on 19 January 2018