Journal website: http://epaa.asu.edu/ojs/ Manuscript received: 28/6/2016 Facebook: /EPAAA Revisions received: 22/1/2017 Twitter: @epaa_aape Accepted: 2/16/2017 education policy analysis archives A peer-reviewed, independent, open access, multilingual journal Arizona State University Volume 25 Number 22 March 13, 2017 ISSN 1068-2341 Exploring School Choice and the Consequences for Student Racial Segregation within Pennsylvania’s Charter School Transfers1 Erica Frankenberg Pennsylvania State University Stephen Kotok University of Texas at El Paso Kai Schafft Pennsylvania State University & Bryan Mann University of Alabama United States Citation: Frankenberg, E., Kotok, S., Schafft, K., & Mann, B. (2017). Exploring school choice and the consequences for student racial segregation within Pennsylvania’s charter school transfers. Education Policy Analysis Archives, 25(22). http://dx.doi.org/10.14507/epaa.25.2601 1 We thank the Pennsylvania Department of Education for providing the data used within this article. This project began as part of a larger study of charter schools in rural areas in Pennsylvania supported by Center for Rural Pennsylvania. We appreciate the assistance of Yosef Bodovski with GIS analysis. All opinions are solely those of the authors. http://epaa.asu.edu/ojs/ http://dx.doi.org/10.14507/epaa.25.2601 Education Policy Analysis Archives Vol. 25 No. 22 2 Abstract: Using individual-level student data from Pennsylvania, this study explores the extent to which charter school racial composition may be an important factor in students’ self-segregative school choices. Findings indicate that, holding distance and enrollment constant, Black and Latino students are strongly averse to moving to charter schools with higher percentages of White students. Conversely, White students are more likely to enroll in such charter schools. As the percentage and number of students transferring into charter schools increases, self-segregative school choices raise critical questions regarding educational equity, and the effects of educational reform and school choice policies on the fostering of racially diverse educational environments. Keywords: charter schools; racial segregation; educational equity; diversity; school reform La exploración de la elección de escuela y los consequências para la segregación racial de los estudiantes dentro de las transferencias escuelas charter de Pennsylvania Resumen: El uso de datos de los estudiantes a nivel individual de Pennsylvania, este estudio explora la medida en la composición racial escuela autónoma que puede ser un factor importante en la selección de escuelas auto-segregadora de los alumnos. Los resultados indican ue, la celebración de la distancia y la inscripción constante, los estudiantes latinos y negros son fuertemente reacios a trasladarse a las escuelas autónomas con mayor porcentaje de estudiantes blancos. Por el contrario, los estudiantes blancos son más propensos a inscribirse en este tipo de escuelas autónomas. Como el porcentaje y el número de estudiantes que se transfieren al charter escuelas aumenta, opciones de escuelas auto-segregadora plantean cuestiones críticas con respecto a la equidad educativa, y los efectos de las políticas de reforma y de elección de escuelas de educación sobre el fomento de diversas razas entornos educativos. Palabras-clave: escuelas charter; segregación racial; equidad educativa; la diversidad; reforma de la escuela A exploração de escolha da escola e as consequências para a segregação racial de alunos em escolas charter transferências Pennsylvania Resumo: A utilização de dados de estudantes a nível individual da Pensilvânia, este estudo explora a extensão a composição racial escola charter pode ser um fator importante na seleção de escolas os alunos auto-segregação. Os resultados indicam ue, mantendo constante a distância e inscrição, os estudantes latinos e negros são fortemente relutantes em mudar -se para as escolas charter com o maior percentual de alunos brancos. Por outro lado, os alunos brancos são mais propensos a se inscrever neste tipo de escolas charter. Como a percentagem eo número de alunos que transferir para as escolas charter aumenta, a escolha da escola de auto -segregação levantar questões críticas sobre a equidade educacional, e os efeitos das políticas de reforma e escolha da escola de promoção de educação racialmente diversos ambientes educacionais. Palavras-chave: escolas charter; segregação racial; equidade educacional; diversidade; reforma da escola Introduction Although charter schools have very high levels of segregation as compared to traditional public schools (Cobb & Glass, 1999; Frankenberg, Siegel-Hawley, & Wang, 2011), some have argued this is due to their disproportionate location in urban areas (see Carnoy, Jacobsen, Mishel, & Rothstein, 2005). Because of the continuing importance of school racial composition for students’ experiences and outcomes in schools and beyond (e.g., Linn & Welner, 2007; Mickelson & Nkomo, School choice and racial segregation in Pennsylvania 3 2012), federal law and policy endorse the importance of pursuing diverse student bodies and/or eliminating racial segregation for traditional public schools (hereafter, TPS) and charter schools alike. At a time of persisting neighborhood segregation, it is critical to assess how student movement to and from charter schools affects school segregation, as well as whether students and parents make these choices while more integrative options exist. Few studies have been able to ascertain how movements to charter schools affect segregation, but those that have suggest that the moves to charter schools for Black students result in more segregated school contexts; findings for White and Latino students, where studied, are more mixed (Bifulco & Ladd, 2007; Kotok, Frankenberg, Schafft, Fuller, & Mann, 2015; Stein, 2015; Zimmer et al., 2009). This study focuses on the extent to which transfers from TPSs to charter schools in Pennsylvania metropolitan areas provide students with opportunities to enroll in racially diverse educational environments, and how the racial composition of “destination” charter schools compares to racial composition of the TPS the student left. Our study uses school-level and individual-level data from Pennsylvania to examine the extent to which student choices influence racial segregation in 10 Pennsylvania metropolitan areas with more than one charter school so that students have multiple charter school options. Specifically, this study addresses two main research questions: 1) Does the racial segregation of Black, White, and Latino students differ between the TPS they exit and the charter schools they enroll in? If so, to what extent? 2) Do students of different races choose charter schools where the racial composition is similar to theirs and to what extent does the school’s distance influence enrollment? We answer these questions by analyzing the transfers of more than 8,000 students who left a Pennsylvania TPS in 2010–11 and transferred to charter schools the following year within the same metropolitan area. Although several other studies use individual data to examine charter school racial composition and segregation, our study makes a unique contribution by considering students’ enrollment decisions in the context of the supply of charter school options of differing composition available to them. As we describe below, Pennsylvania is an ideal site to study these questions because of its legal and demographic context pertaining to charter schools. The number of charter schools has increased 54% between 2006–07 and 2010–11, and charter schools are located disproportionately in the state’s metropolitan areas (Schafft et al., 2014). Perspectives Theory of School Choice Assumptions. Historically, public school choice in the United States has long been geographically constrained. For many students, the only choices they have regarding where they can attend school are tied to their families’ residential locations within certain neighborhoods, which are associated with specific school districts and often a particular school.2 Initially advocated by Milton Friedman in the 1950s who proposed school vouchers, the market theory of school choice suggests that allowing families more choice would spur competition among schools. Schools are also freed from certain governmental regulations. Together these processes would improve educational outcomes for all students. Friedman (1955) and others (e.g., Hoxby, 1996; Moe, 2001) argued that while all students would benefit, disadvantaged students would benefit the most from a more market-oriented system of schooling since markets are color-blind (Henig & McDonald, 2002). 2 Of course, given discrimination in the housing market, and community zoning restrictions, residential choice is not equally available for all (e.g., Rothwell, 2012). Education Policy Analysis Archives Vol. 25 No. 22 4 School choice today takes many forms (Orfield & Frankenberg, 2013). One of the oldest and most widespread forms of choice has been magnet schools and some districts also operate choice systems that govern the assignment of students to all schools. While both magnet schools and controlled choice policies began to try to create more integration, not all magnet schools or districtwide choice or transfer policies are now aimed primarily to achieve diversity. Charter schools began in the early 1990s as a means to provide more flexibility than TPS in exchange for accountability. While charter schools are publicly-funded schools of choice, they are often separate from public school districts and may have their own admissions process. More recently, implementations of vouchers have provided public funding for students to use at private schools. The implementation of school choice during the last several decades has not delivered improved outcomes for all students as Friedman and Moe predicted, although school choice advocates might argue that the reason choice in practice has not delivered is that it has not been implemented fully (Scafidi, 2015). Charter schools have emerged as the most rapidly increasing form of school choice, spurred in part by governmental policy incentives in recent decades. The effects of charter schools on academic achievement are vigorously debated, though most analyses suggest that many students’ scores do not increase after moving to charter schools (e.g., Bettinger, 2005; Bifulco & Ladd, 2007; Chingos & West, 2015; Clark, Gleason, Tuttle, & Silverberg, 2015), that charter schools may have mixed or negative effects on a TPS (Bettinger, 2005; Zimmer et al., 2009), and there may be high student attrition from charters (Miron, Urschel, & Saxton, 2011). Moreover, most researchers agree that there is considerable variety in student outcomes within the charter school sector (Berends, 2015; Buddin & Zimmer, 2005). Indeed, Orfield & Frankenberg (2013) argued that market-based school choice is based on several assumptions. First, all potential consumers (e.g., families) have equal access to accurate information about their possible options (Bell, 2009; Holme, 2002). Second, it assumes there is widespread competition with many providers (e.g., schools), which would provide incentives for providers to improve and differentiate their product (Lubienski & Lee, 2016). Third, all consumers must have equal choice opportunities. National and state-level educational policies have often been implemented to increase access to school choice (U.S. Department of Education, 2014). However, for school choice to produce the intended improved outcomes for all students under the market theory, families need access to appropriate information in a means that is understandable. Moreover, competition assumes that there is a shared set of criteria across all groups by which schools are being judged; this is often assumed to be standardized test or accountability scores of schools (Orfield & Frankenberg, 2013). Yet research suggests that higher income families may have more access to information to make choices (Teske & Schneider, 2001). When coupled with other advantages such as reliable transportation, because of the overlap between family socioeconomic status and race, choice may further stratification (among public school districts, see Carlson, Lavery, & Witte, 2011). Further, some forms of academic achievement data are less comprehensible to parents (Hastings & Weinstein, 2008). For example, test scores are not necessarily significant predictors of parents’ school choices (Butler, Carr, Toma, & Zimmer, 2013), nor do they always choose the “optimal” school (Villavicencio, 2013). Recent research also suggests that while some charter schools are located in “high need” areas, they are rarely located in the “highest need” areas (LaFleur, 2016). Given the consistent enrollments of some charter schools that have weak academic outcomes, it is questionable whether these assumptions of the market theory hold in a complex system like public schools. In this case, school choice might result in increasingly segregated school environments instead of a competitive “market-driven” system of school choice that spurs educational innovation through open competition. School choice and racial segregation in Pennsylvania 5 Charter Schools and Segregation. Early advocates of charter schools argued that these schools would yield educational competition and institutional innovation (Chubb & Moe, 1990), while other charter school advocates later argued that choice afforded by charter schools would in fact foster educational equity and racial diversity (Finn, Manno, & Wright, 2016). The logic of the latter argument was that families from racially segregated, high-poverty neighborhoods would no longer be bound by traditional school attendance zones when evaluating options for the education of their children. That is, new choice options afforded by charter schools would create new opportunities for students to attend schools with more racially integrated student bodies than existing options within TPS (Frankenberg, et al., 2011; Lubienski, Gulosino, & Weitzel, 2009). While some students would remain within racially segregated TPSs, the overall charter “school choice” outcome could well be the creation of more racially diverse schools without having to address the politically thornier and far more complex issues of deeply entrenched racial and socioeconomic housing segregation, or contentious school catchment boundaries. It is exactly the enduring school segregation (Orfield & Frankenberg, 2014) coupled with the suggestion that charters can yield educational equity that at least partially explains why many proponents frame school choice as a civil rights issue (Scott, 2011). This underscores the importance of looking more systematically at how charter school choice may, in fact, yield integrative or segregative outcomes with regard to school enrollments. Despite the optimism among some proponents that school choice would create racially diverse student bodies, charter schools have been found to be more racially homogenous than TPSs (Frankenberg et al., 2011; Garcia, 2008). For example, in several studies, Black students attended charter schools where the percentage of Black students was 15 or more percentage points greater than the TPS they exited (Kotok et al., 2015; Weiher & Tedin, 2002). Additionally, studies of individual charter movers in Arizona (Garcia, 2008), North Carolina (Bifulco & Ladd, 2007), Texas and California (Booker, Zimmer, & Buddin, 2005), and Ohio (Zimmer et al., 2009) support the claim that students—especially Black students—tend to leave more racially diverse TPSs for more segregated charter schools. Some studies also find that White students’ transfers to charter schools are also segregative (Zimmer et al., 2009). Some scholars (Ritter, Jensen, Kisida, &McGee, 2010) argue that racial segregation in charter schools simply reflects their disproportionate location in high-minority neighborhoods within metropolitan areas and suggest school comparisons only in central cities where segregation is perceived to be the highest. This argument, while provocative, tends to ignore: a) how the charter marketplace theoretically empowers parents to exit their neighborhood schools and even their district in some cases; and b) increasing levels of segregation outside of central cities. Ultimately, a more accurate means of examining the effect of school choice on racial composition involves using individual data on students enrolling in charter schools and examining the school choices that these students make.3 Understanding School Choices Researchers have suggested several factors to explain de facto charter school racial segregation, including strategic location of charter schools (Lubienski, Gulosino, & Weitzel, 2009), lack of transportation among parents, particularly for low-income families (Siegel-Hawley & Frankenberg, 2011), school closings (Zimmer et al., 2009), and the continued role of housing segregation despite the assumption that school choice policies would decrease the segregating effect 3 Considerable debate exists about how to measure charter school segregation because of concerns about the appropriate geographic scale of comparison between TPSs and charter schools (see Frankenberg et al., 2011; Garcia, 2008). Because of our access to longitudinal student-level data, we avoid this methodological debate. Education Policy Analysis Archives Vol. 25 No. 22 6 of neighborhood boundaries on school choice (Booker, Zimmer, & Buddin, 2005). Although in some instances parents may transport their children long distances to schools, parents typically make school enrollment decisions based heavily on proximity and are therefore still constrained to a certain set of choices (Lubienski & Dougherty, 2009).4 In fact, Bifulco & Ladd (2007) found that parents rarely travel more than 10 miles to a charter school and usually far less (see also Butler et al., 2013). It should be noted that, nationally, high school students typically travel further to school than elementary and middle school students (U.S. Department of Transportation, 2008). However, high school students able to choose their school often consider younger siblings school location, their neighborhood friends’ school choice, and they simply may be more aware of nearby school choices. Additionally, state charter school laws may restrict enrollment eligibility based on school district boundaries. Therefore, boundaries still matter in a school choice context.5 Some studies have sought to understand the role in which preference for similar-race students influences families’ school choice decisions, including the choice of charter school enrollment. In general, studies find discrepancies between parents’ reported preferences and the options they eventually take when choosing a school. In one of the earlier studies of preferences among charter school enrollees, Weiher & Tedin (2002) found that charter school parents, regardless of race, reported it was not important to put children in schools that are majority of the same race (see also Schneider, Marschall, Teske, & Roch, 1998). Yet, they found that students of each race (Black, White, and Latino) transferred to charter schools in which their own racial group was a higher percentage than the public school from which they transferred. Indeed, they concluded that race was one of the strongest predictors of charter school choices, despite parents’ reported preferences on the survey.6 The researchers, however, were not able to ascertain whether families had more diverse options available or whether they made segregative choices because those were the only charter school options they had. Indeed, a subsequent study of charter school movers in North Carolina found that only 19% of Black movers in five large metropolitan areas had integrative charter school options (Bifulco & Ladd, 2007). Other research has found links between school racial composition and parents’ school choices. In Washington, DC, parents’ preferences of charter schools were also more racially-based than their responses to a survey in which parents reported that school composition did not matter (Buckley & Schneider, 2007). Research in other contexts found that a school district’s diversity was associated with higher enrollment by Whites into disproportionately White charter schools, a trend that did not seem to be significantly affected by a school’s academic quality (Renzulli & Evans, 2005; see also Saporito & Sohoni, 2006 for school choice more generally). A study of upstate New York metropolitan areas—demographically similar in many respects to Pennsylvania—found that White parents’ school choices were strongly influenced by wanting their children to attend a school with fewer non-Whites (Lankford & Wyckoff, 2005). A more recent study echoes these issues, showing that given hypothetical school choices, White parents are less likely to select a school if it has a high percentage of Black students, even if the school is rated as high performing (Billingham & Hunt, 2016; see also Holme, 2002). Finally, a national analysis of students’ choices in 2003–04 found that 4 Miron, Urschel, & Saxton (2011) found differences by grade level in students’ exit rates from KIPP charter schools, suggesting that school choice preferences may operate differently based on whether a student attends primary or secondary school. 5 Even with the growing number of cyber charter schools, the majority of school choice options remain in brick and mortar schools, which logistically have some geographic constraints. 6 Academic achievement was likely not an explanatory factor as students, on average, would have been in schools with a higher proficiency rate had they remained in their public school—even though this was selected as one of the most important factors on the parent survey. School choice and racial segregation in Pennsylvania 7 school racial composition was not a significant predictor for students moving to charter schools. Yet, higher percentages of Latino and Black students made magnet schools and other public schools less desirable (Butler et al., 2013).7 This body of research largely focuses on the demand-side of charter school selection; on the supply side, charter schools may employ a variety of practices that limit the availability of seats to interested students, including in ways that may disproportionately advantage certain groups over others and which would also affect racial stratification. Some researchers (Jabbar, 2015a; Lubienski, 2007) find that charter schools market themselves to appeal to a certain type of student or are placed in strategic locations that hinder enrollment from certain student demographic groups (LaFleur, 2016). Jabbar (2015b) shows that in New Orleans, a context where virtually all schools are charter schools, school leaders’ perception of their competitive networks reflect and often reproduce a market hierarchy within the charter school system. Administrators used strategies such as cream- skimming and targeted marketing as recruitment strategies (Jabbar, 2015a). Other work suggests a level of back-filling seats with high performers in the KIPP charter chain (Nichols-Barrer, Gleason, Gill, & Tuttle, 2016). Overall, strategies such as these have the potential to affect charter school racial composition if school leaders selectively recruit and enroll students who belong to particular racial or ethnic groups. While the consensus of research finds that parents’ actions, if not reported preferences, in making school choices tend to exacerbate segregation, it is unclear whether those decisions are being made in the presence or absence of more diverse charter options. Such information about school choices with respect to the supply of charter school options is the contribution we seek to make with this analysis. If we find that parents have few charter school choices that would provide more integrative experiences, then this suggests segregated choices are being made because of limited supply of charter schools with varied racial composition. On the other hand, if segregative choices are being made even as there are more integrative choices as part of a student’s choice set, this implies that such choices are not being driven by supply. Each has different policy implications for educators and advocates. The Pennsylvania Context Pennsylvania has 500 school districts, so that there can be many relatively small school districts within metropolitan areas. While this provides choice for families who are willing to (and have the means to) relocate residentially, it may also mean there are relatively few public schools for students to select within a given district. Pennsylvania law also provides for voluntary participation in interdistrict open enrollment, but it is unclear how many districts, if any, participate given that both sending and receiving districts must agree.8 The Pennsylvania legislature first authorized charter schools in 1997, and charter school enrollments have grown rapidly since, enrolling 6% of all students in Pennsylvania in 2011–12. There are two types of charter schools: brick and mortar and cyber schools. The latter provide for virtual instruction of pupils and are not limited by practical considerations of distance as are brick and mortar schools. Approximately 32,000 of the 105,000 charter school students in 2011–12 were enrolled in the state’s 12 cyber charter schools; the other 73,000 students attended 150 brick and 7 The lack of significant effect among charter school choosers could be due to the small number of students (n=80) in the sample. 8 24 PS 13–1316. See http://ecs.force.com/mbdata/mbtab8OE?sid=a0i70000006fu14&rep=OE132T. Pennsylvania also offers a limited voucher program that low-income students in low-performing schools are eligible for and with which they can transfer to private or public participating schools. http://ecs.force.com/mbdata/mbtab8OE?sid=a0i70000006fu14&rep=OE132T Education Policy Analysis Archives Vol. 25 No. 22 8 mortar charter schools, the vast majority of whom were concentrated in the state’s metropolitan areas (Pennsylvania Department of Education, 2016). By 2011–12, the year of this analysis, every student had the choice of enrolling in a cyber charter school in lieu of a TPS. There are no enrollment restrictions, costs to the student, or geographic limitations, and cyber charter schools provide Internet and computers to their students. This means any student can enroll in any cyber charter school that offers their grade-level. Although cyber charter students make up a growing share of the overall charter school enrollment in Pennsylvania and are disproportionately White in comparison to TPS and brick and mortar charter schools (Mann, Kotok, Frankenberg, Fuller, & Schafft, 2016), the effects of cyber charter enrollments are beyond the scope of this article, which is principally concerned with the segregative processes within brick and mortar educational contexts (see Kotok et al., 2015, for more details). Additionally, while we focus here on school choice contexts in metropolitan areas, including physical locations of charter schools, cyber charters do not have a geographic location making such analyses impossible. Pennsylvania is a useful state to examine charter school segregation because it contains two largely minority central cities, racially mixed suburban rings, and vast swaths of rural areas that are largely White. Like most states with substantial charter school enrollments, Pennsylvania’s charter schools have a disproportionately lower percentage of White students than TPSs (Frankenberg et al., 2011). Yet, charter school enrollment is racially diverse (just under half of students were White) thereby enabling an analysis that examines the transfer patterns of White students as well as Blacks and Latinos. Approximately half of charter school students attended schools in central cities in 2007–08 while another 30% were in suburban areas, and 15% in towns or rural areas (Frankenberg et al., 2011). With its complex history of housing segregation, partial and unsuccessful legal remedies for school segregation, and metropolitan fragmentation, Pennsylvania offers an ideal opportunity to test proponents’ claims that charter schools will foster desegregation given persistent TPS segregation (Kotok & Reed, 2015; Morrison, 2004). In this paper, we focus on the transfers of students from a TPS to brick and mortar charter schools. An important aspect of Pennsylvania’s charter school law is that public school districts are required to transport students residing in their district to any charter school within the district or to any charter within 10 miles of the district’s boundary. This provision may enable students to select schools with an important barrier to access—transportation—greatly mitigated (see generally Orfield & Frankenberg, 2013). Taken together, these legal and demographic reasons make Pennsylvania an informative context to study how students’ transfer patterns to charter schools may be more or less segregative. Importantly, we are also able to examine whether these transfers are being made in the presence or absence of charter schools that have different racial composition from one another. Namely, if students are making segregative transfers to charter schools, we will know whether they had less segregative options available than the school they ultimately ended up attending or whether only segregative charter school options existed. Each scenario has different policy implications. Data and Methods This analysis utilizes a dataset that allows us to examine our research questions with individual student-level data across consecutive years in Pennsylvania. We use two main data sources. First, we use individual data for 2010–11 and 2011–2012 provided to us from the Pennsylvania Department of Education (PDE) indicating what school a student attended in each year, their race, and their grade. We linked the individual data with school-level data from the School choice and racial segregation in Pennsylvania 9 Common Core of Data (CCD) collected by the U.S. Department of Education to provide the student demographics of the school (racial and socioeconomic composition), school type, and location. We limit our analysis to Black, White, and Latino students and schools in the 10 metropolitan areas in Pennsylvania with more than one charter school (see Table 1 for list of metropolitan areas). Because of Pennsylvania’s law providing for transportation across district boundary lines, we selected metropolitan areas instead of urban districts only for our analysis of transferring students. In sum, we analyze the records of 8,056 individual students.9 The analysis was a multi-stage process. First, we determined which students in the state of Pennsylvania changed from TPSs to brick and mortar charter schools between school years. Some students also switched schools one or more times within the school year. Because we do not know the order of schooling within the year and because they comprised less than 5% of all movers, we have eliminated them from our analysis.10 To understand the racial segregation that students experience within school settings, we rely on the exposure/isolation index as our measure of segregation because of its ease of interpretation that allows for comparison regarding the school-level racial context that students encounter (Massey & Denton, 1988).11 Isolation refers to the extent to which the enrollment in a given student’s school matches that student’s own racial group. The exposure index (P*) is a weighted average of the school-level composition that the “typical” student of group x experiences. We calculated exposure as: 𝑃∗ = ∑( 𝑥𝑖 𝑋 ∗ 𝑦𝑖 𝑡𝑖 ) 𝑛 𝑖=1 where n is the number of schools, 𝑥 is the number of the first racial group of students in the school i, X is the total number of the first racial group of students in the larger geographical area (e.g., district, state), 𝑥i is the number of the second racial group of students in the school i, and 𝑥i is the total number of students in the school i. If all schools were perfectly integrated, student exposure would be identical across all groups, and would reflect the racial composition of the larger unit (Massey & Denton, 1988). For isolation, in the above equation, 𝑥i is equal to xi. We calculate the isolation in the student’s sending and receiving school separately for Black, Latino and White students who transferred from a TPS to a charter school. It is useful to interpret isolation in relation to a group’s share of the overall population. Since the majority of students in each metropolitan area in our sample is White, the fact that Black students leave TPSs and transfer to charter schools in which they comprise a majority of the enrollment is indicative of segregation. Since there is no universally agreed upon threshold for high racial isolation, we examine how racial isolation values compare in between TPS and charter sector to ascertain the “net effect” of the transfers on student segregation.12 9 An additional 300 students in these metropolitan areas have race/ethnicity that is not Black, White or Latino and thus were not part of our analysis. 10 Virtually all of such within-year movers were not enrolled in a brick and mortar charter school and they were demographically similar to between-year movers. 11 While other segregation measures, such as H, may be favored in other types of analyses because of their decomposition properties, this analysis did not necessitate such decomposition because we were looking only at charter schools. 12 Such an approach, we believe, is methodologically superior to those comparing whether charter schools’ composition reflects that of the surrounding community for several reasons: there are debates in the literature Education Policy Analysis Archives Vol. 25 No. 22 10 For the second research question, we used ArcGIS software to locate all of the charter schools within a 10-mile distance from each TPS as a means of determining the “choice set” for public school students who would transfer to charter schools, which other research suggested encompassed the vast majority of choosers (Bifulco & Ladd, 2007; Butler et al., 2013).13 Since charter schools in Pennsylvania (and elsewhere) tend to use diverse grade configurations (e.g. K–8, 5–8, K–12), we separated our analysis into two subsets: K–grade 7 students and grade 8–12 students.14 For K–7 students, any school that had any elementary or middle school grades was included in their choice set while the latter group included any school enrolling high school students. The purpose of separating the sample into two grade-level groups was to approximate the number of choices available to students in each age group. For example, a fourth grader may be choosing between K–5 schools, K–8 schools, and K–12 schools, but a high school with only grades 9-12 would not be part of the student’s choice set. While this analytic strategy intentionally minimizes structural moves, we focused on whether students switched to charter schools for any reason. In our analysis, we focus on the 10 metropolitan areas within Pennsylvania where students had more than one potential brick-and-mortar charter school option in the 2011–12 academic year (see Table 1). Each metropolitan area has a majority of White students in TPSs, but this ranges from a slight majority to 90% of the total enrollment. We kept the subset of transferring students in metropolitan areas where there were two or more charter schools in their “choice set” and examined the distribution of the racial composition of these charter schools using quintile groupings for the percentage of White students. Table 1 Student Transfers from TPS to Charter Schools in Selected Pennsylvania metropolitan areas, 2010–11 to 2011–12 Metropolitan Statistical Area (MSA) # of Charter Schools # of Students Transferring Racial Composition of Public Schools in Each Metropolitan Area % White % Black % Latino % Other Allentown 3 183 65.20 8.36 22.40 4.20 East Stroudsburg 2 33 58.35 20.17 18.60 2.88 Erie 4 209 78.70 12.28 4.65 4.36 Gettysburg 2 16 85.60 2.70 10.41 1.28 Harrisburg-Carlisle 2 18 69.86 16.33 7.04 6.78 Philadelphia 94 7,164 52.36 30.09 9.56 7.99 Pittsburgh 22 659 81.61 13.41 0.93 4.04 Scranton 2 15 82.17 5.17 10.21 2.45 State College 4 34 90.24 2.04 1.85 5.87 York 5 25 78.77 9.41 8.09 3.73 Total 140 8,356 862,100 252,081 107,604 74,564 Note: The table only includes students who move between TPS to charter and remained in same metropolitan area. An MSA is a metropolitan area that has an urban core with more than 50,000 residents. as to what is the appropriate scale of the public school comparison for charter schools (Gulosino & d’ Entremont, 2011) and the fact that in many areas there is high and increasing segregation among TPSs. 13 Ideally, it would be useful to construct a choice set based on a student’s home address, but we do not have that information. Thus, we use the TPS the student attended as a proxy for student’s residential location. 14 Eighth graders were included in high school subset since they would be changing to a high school during the transfer year. Note, grade 12 movers only included if they repeated grade 12 in 2011–12. School choice and racial segregation in Pennsylvania 11 To augment our descriptive analysis of transferring students’ choice sets, we used conditional logistic regression to infer the degree to which White, Black, and Latino families transferred to schools based on charter school racial composition and distance. The conditional logistic model, also known as McFadden’s choice model (1973), fit our analysis well since it allows multiple alternatives whereas basic logistic regression models only consider one outcome, and analyzes choice specific details such as distance whereas multinomial logit only considers case specific variables. The unit of analysis in conditional logit models is the set of alternatives (e.g., racial composition, distance) and is appropriate to use when individual choices are “a function of the relevant characteristics of [available] alternatives, rather than the attributes of the individuals” (Hoffman & Duncan, 1988, p. 425). We estimated the following conditional logit model for an individual with J set of choices where 𝑃𝑖𝑗 represents the probability that person i selects the jth alternative: 𝑃𝑖𝑗 = exp (𝑍𝑖𝑗 ∝ ) /∑ 𝐽 𝑘=1 exp(𝑍𝑖𝑗 ∝) 𝑍𝑖𝑗 represents the characteristics of the jth alternative for person i; ∝ represents the corresponding parameter vector. To estimate the conditional logit, we first organized the data into pair-wise matches for charter movers with each school composition option (e.g. 0 to 20% White). If a student did not have one of the five options located within 10 miles, that particular alternative was not included amongst their choices. Distance was calculated as an average of all available schools for each alternative. For instance, if a student had two charter school options with 0–20% White enrollment with the first school being three miles away and the second school being five miles away, the average distance for that alternative would be four miles. Similarly, we also included the average logged enrollment for each alternative. Finally, we estimated the conditional logit model separately for each racial group and grade type.15 After extensive analysis, we decided not to include a measure of low-income students in our analyses for several reasons. The most significant reason for this decision was that several charter schools were missing FRL data. While our dataset includes individual race, it did not provide individual information on FRL status. Second, as is the case among schools nationally, a strong negative correlation exists between the percent of White students and percent of free/reduced price lunch students (FRL) in Pennsylvania charter schools (r=-0.70). When examining the distribution of the percentage of FRL students, the distribution of this variable was not normal, but rather had a bimodal distribution clustered around (1) 40–50% FRL and (2) 85–100% FRL. Finally, there was extremely little FRL variation within each racial composition quintile (e.g., very high FRL% for choices in the 0–20% White category and very low FRL% for the 80–100% White schools).16 Given the irregularities with the FRL data and the overall correlation between race and poverty, we ultimately elected to focus on race for the purpose of this study. However, the high correlation demonstrates that the two factors are strongly related to one another. 15 We separately estimated models for students only in the Philadelphia metropolitan area and in all non- Philadelphia metropolitan areas because of the large percentage of students in that metropolitan area to ensure that there are not different patterns obscured by the large number of charter movers. We also ran the model separately for students with and without an IEP (see appendix C). 16 When we did run the analysis including FRL for those schools with data, we find that the results are similar for Blacks and Latinos while including the percent of free/reduced lunch students reversed some of our findings on White students’ preference for majority White schools likely due to the lack of variation of FRL percentage within each racial quintile. However, given various issues with the FRL variable, we lacked confidence to make sound inferences and we ultimately chose to present the more parsimonious model. Education Policy Analysis Archives Vol. 25 No. 22 12 Limitations Because each state’s charter school law is unique, as are the demographic contexts and school boundary configurations, our findings in Pennsylvania cannot be generalized to other states or other forms of school choice than brick and mortar charter schools. However, because of the explosive growth of charter schools across the country, we think this focus is important for informing our understanding of these federal and state policies. Further, there are other individual characteristics that might be salient to explore in terms of the transfers students make to charter schools, such as individual students’ socioeconomic status. Ideally, we would like to include both racial and economic composition of charter schools to understand how that relates to students’ transfers, but were unable to because of the aforementioned data concerns and the fact that some schools were missing FRL data. We also don’t have information on the extent to which spaces are available at the particular grade level that students are in, which may limit their consideration of some choices. Finally, these analyses do not include students transferring to cyber charter schools, which are a growing share of students in Pennsylvania. Findings Racial Isolation of Charter School Choosers We focus here on analyzing the choices of students transferring to charter schools from TPSs between 2010–11 and 2011–12. Importantly, our analysis focuses on the movers within the selected 10 metropolitan areas, and their racial composition is almost identical to that of all students moving to charter schools in Pennsylvania after the 2010–11 school year (Figure 1). Source: Pennsylvania Department of Education Figure 1. Racial Composition of Pennsylvania Students Moving from TPS to Charters, 2010–11 to 2011–12. Note: All Charter Movers refers to any student who attended a TPS in 2010–11 and a charter school in 2011–12 (n=10,169). Only Metropolitan Areas is a subsample of these students representing only charter movers who lived in the same metropolitan both years of attendance (n=8,356). 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 70.00% 80.00% 90.00% 100.00% All Charter Movers Only Charter Movers Within Metropolitan Areas Other Asian White Latino Black School choice and racial segregation in Pennsylvania 13 Between the two school years (2010–11 to 2011–12), Black and Latino students moved into charter schools where they were more racially isolated than the already segregated TPS they left (see Table 2).17 For example, the typical Black mover left a TPS in 2010–11 with around 68% Black students to attend a charter with 80% Black students—an increase of nearly 12 percentage points. These patterns hold regardless of grade level, with younger students experiencing slightly more segregative settings, post-school transfer.18 Table 2 Racial Isolation for Students Moving from a TPS to Charter School in Selected Metropolitan Areas, 2010–11 to 2011–2012 Number of Students Sending TPS Receiving Charter Difference in isolation experienced by transferring student Black Grades K–7 2,673 68.54 80.66 12.12 Grades 8–12 2,827 67.49 78.89 11.40 Total 5,500 68.00 79.75 11.75 White Grades K–7 596 63.41 61.96 -1.45 Grades 8–12 349 62.39 63.81 1.42 Total 945 63.03 62.64 -0.39 Latino Grades K–7 684 43.87 53.62 9.75 Grades 8–12 927 39.06 53.21 14.15 Total 1,611 41.10 53.38 12.28 Source: Pennsylvania Department of Education; NCES Common Core of Data Note: Student grade level is from 2010–11. Latinos did not end up in charters as racially isolated as Blacks and did not leave TPSs that were as racially isolated. These lower isolation numbers are probably a function of being a lower proportion of overall student population. Still, Latino movement patterns were similar to that of Blacks in that they enrolled in charter schools with higher racial isolation than the public schools they left. The typical Latino mover left a TPS with 41% other Latinos for a school with over 53% other Latinos—an increase of 12 percentage points, as was the case for Black students (who left and transferred to more isolated schools). Latino students enrolled in charter schools where a majority of students were also Latino. These trends were similar for Latino students regardless of grade level, although older students experienced a larger difference in racial isolation than younger students The trend for White movers to charter schools was mixed with the net effect overall of White students moving to charter schools with roughly the same racial isolation, on average. 17 These trends are similar to statewide trends in Pennsylvania in 2011–12 and earlier years (Kotok et al., 2015). 18 One possible explanation for these segregative moves is that charters with high Black isolation had a mission focused on African American history. We did not systematically analyze charter schools’ missions since other work had found ethnocentric themes to not be significant predictors in addition to racial composition of student transfers (Bifulco & Ladd, 2007). Education Policy Analysis Archives Vol. 25 No. 22 14 Younger students moved to charter schools with slightly lower isolation while high school White students’ moves were to charter schools that resulted, on average, in slightly higher isolation. Like Black students, White students in the 10 metropolitan areas examined had attended TPSs with fairly high White isolation, similar to national trends (e.g., Orfield & Frankenberg, 2014). Because most of the sample lived in the Philadelphia metropolitan area, we separately analyzed students in Philadelphia and those in all other metropolitan areas to see if statewide patterns were influenced by dynamics unique to Philadelphia (see Appendix A and B). We found that the trends for Black and Latino students were consistent regardless of whether they were in Philadelphia or another metropolitan area: students’ transfers were to charter schools with more same-race peers. The pattern was mixed among White students. Transfers by White students in the Philadelphia metropolitan area were segregative in that they enrolled in charter schools that, on average, had a higher percentage of White students while White students in non-Philadelphia metropolitan areas moved to slightly more diverse charter schools than the TPSs they left. Types of Choices Available to Charter School Choosers We examine the types of charter school choices available to movers in more detail, first for K–7 students and then for high school movers (grades 8–12). We first examine descriptive data to understand whether students are making segregative choices because those are the only options available to them. Since the majority of the movers lived in the Philadelphia metropolitan area, we provide an analysis of all “Pennsylvania Metropolitan Areas” and a subsample of movers in “Other Metropolitan Areas” not including Philadelphia. Regardless of whether choosers are in Philadelphia or elsewhere, K–7 students transferring to charter schools usually had a substantial number of choices (see Table 3). White students were mostly likely to have only one choice; approximately one- eighth of White elementary school movers had only one charter school option. Conversely, around 75% of Black and Latino K–7 students had more than 20 charter school options.19 Even in the non- Philadelphia metropolitan areas—which have fewer choices—nearly half of White movers had six or more charter school options in their choice set; higher percentages of Black and Latino students in these nine other metropolitan areas also had a similarly large number of choices. In addition to having a number of charter school choices, most student movers, regardless of race and metropolitan area, had charter school choices of varying compositions. Across all 10 metropolitan areas, at least two-thirds of students, regardless of race, had one or more charter school in each of the quintiles of charter school racial composition. White students were less likely to have heavily non-White charter school options (83% overall had a charter school choice with fewer than 20% White students) as well as overwhelmingly White charter schools (68% lived near an 80–100% White school). Higher percentages of Black and Latino students had school choices in each quintile than did White students, indicating they had more diverse sets of choices. This somewhat surprising finding is likely due to a few overwhelmingly White charter schools in urban centers. In non-Philadelphia metropolitan areas, the pattern is more like what would be expected: Blacks and Latinos, in comparison with White students, were less likely to have a charter school option that was 80% or greater White enrollment. In sum, students of all races had charter schools of varying racial composition in their choice set. White students lived further, on average, from the nearest charter school and also furthest, on average, from their chosen charter schools. The difference between the nearest charter school and chosen charter school for White students was shorter than the difference for Black students. 19 Though not a direct focus of this analysis, the considerable variation between White and non-White students’ options implies a fairly high level of residential separation for these students. School choice and racial segregation in Pennsylvania 15 Table 3 Choice set characteristics for K–7 school students All Pennsylvania Metropolitan Areas Non-Philadelphia Metropolitan Areas Black White Latino Black White Latino Total number of choices observed 2,673 596 684 415 236 135 # of charter schools in choice set % of students with: 1 choice 0.88 12.54 0.87 1.41 11.34 0.74 2 choices 1.17 6.03 2.61 4.92 9.31 11.85 3–5 choices 4.48 16.20 6.37 20.06 30.37 23.70 6–10 choices 5.45 11.89 12.74 30.21 24.69 60.00 11–20 choices 6.93 11.90 0.86 42.85 24.30 3.70 Over 20 choices 81.09 41.42 76.47 0 0 0 Avg. # of charters to choose from 43.55 23.26 41.48 8.74 6.72 5.24 Avg. enrollment 543.78 531.30 546.38 302.85 294.98 359.16 Racial composition of schools in choice set % of students with at least one choice with: 0–20% White 99.51 83.39 97.66 98.31 78.81 91.11 20–40% White 96.03 75.50 83.63 89.64 69.07 31.85 40–60% White 93.75 72.99 90.79 73.49 51.27 65.93 60–80% White 94.65 81.21 91.23 79.04 71.61 68.89 80–100% White 84.18 68.29 83.77 19.52 33.90 25.93 Distance for choice Avg. distance to nearest charter 0.70 1.86 0.76 1.62 2.35 1.55 Avg. distance to furthest charter (within 10 miles) 8.98 8.31 8.56 7.84 7.17 5.74 Avg. distance to chosen charter 2.36 2.96 1.65 3.17 3.20 2.48 Chosen vs. nearest difference 1.66 1.10 0.89 1.55 0.85 0.93 Distance to each choice set 0–20% White 4.46 5.50 4.26 4.27 4.56 3.21 20–40% White 5.42 4.47 3.97 4.63 4.64 3.10 40–60% White 5.40 5.18 4.95 5.04 5.41 4.82 60–80% White 5.47 4.78 5.67 5.67 5.08 4.92 80–100% White 5.97 5.58 5.28 4.76 4.66 4.49 Note: Students who transferred to charter schools in 2011–12 in the same metropolitan area as their former TPS. Education Policy Analysis Archives Vol. 25 No. 22 16 Notably, schools with between 20% and 40% White students were, on average, closer for Whites than any of the other choice sets with different racial composition. Black students overall lived the closest to the nearest charter schools while Latino students outside of Philadelphia were most likely to live closest to and attend charter schools located the closest to them. Although Black students typically had charter school options in all five quintiles of racial composition, they would have to travel about a mile further, on average, to attend a charter school with more than 20% White students. However, outside of the Philadelphia metropolitan area, the differences in distance to charter schools of different quintiles were less pronounced for Black and White students. Latino students lived further, on average, from charter schools with more than 40% White students. In general, there are fewer charter high schools than K–7 schools but they typically have higher enrollments. Thus, overall, we see fewer choices for high school students, particularly outside of the Philadelphia metropolitan area (Table 4). Black and Latino students, however, still had many choices at the high school level, with over 90% having more than 20 choices, although this was due to most of these students living in the Philadelphia area. White students had fewer choices than their Black and Latino peers. In comparison to younger students described above, lower percentages of White high school students had only one choice. On average, Black high school students had around 33 choices, Latino high school students had more than 35 choices, and White high school students had around 19 choices. However, outside of the Philadelphia metropolitan area, the difference between racial groups was very minor with Black high school students having an average of six choices followed by White students with almost five choices and Latinos with fewer than four choices. Overall, many high school movers to charter schools also had potential charter school choices with a range of demographics. In non-Philadelphia metropolitan areas, there were very few students who had a choice set including overwhelmingly White charter schools (those with more than 80% White students), but most students had both majority White (60–80% White) and majority non-White schools (<40% White) in their choice sets. Ironically, when contemplating the choices of all students, few students from any race had the most diverse schools, schools that were 40–60% White, in their choice sets which indicates the relative polarization of charter school choices in these metropolitan areas. White students were most likely to have such diverse schools to choose from (13%). As with younger students, White high school students lived, on average, further from the closest charter school, including the charter school they eventually chose. Latino students were closer to charter school options when including those in the Philadelphia metropolitan area. In contrast to the K-7 results, the difference between the nearest and chosen charter school was approximately one mile with Latino students having the smallest discrepancy and White students having the largest. Although, most high school students had varying choices (with the exception of the middle quintile, 40-60% White)), Black and Latino students typically had to travel considerably further than Whites if they wanted to attend a school with between 60-100% White students. It should be noted, that the differences in travel distance among charter school racial composition quintiles were not as large when analyzing students outside the Philadelphia metro area. School choice and racial segregation in Pennsylvania 17 Table 4 Choice set characteristics for high school students All Pennsylvania Metropolitan Areas Non-Philadelphia Metropolitan Areas Black White Latino Black White Latino Total number of choices observed 2,827 349 927 185 135 26 # of charter schools in choice set % of students with: 1 choice 0.21 3.72 0.11 0.54 3.7 3.85 2 choices 1.31 11.75 1.40 19.46 28.15 50.00 3–5 choices 0.68 15.48 1.29 8.65 33.33 42.31 6–10 choices 4.60 14.04 0.22 70.26 34.81 3.85 11–20 choices 0.53 4.31 0.54 1.08 0 0 20+ choices 92.61 50.07 96.44 0 0 0 Avg. # of charters to choose from 32.97 19.39 35.37 5.99 4.58 3.46 Avg. enrollment 643.53 593.83 669.49 643.53 593.83 669.49 Racial composition of schools in choice set % of students with at least one choice with: 0–20% White 98.62 85.39 98.60 82.70 73.33 53.85 20–40% White 99.01 74.79 98.60 89.73 47.41 50.00 40–60% White 0.32 13.18 1.08 3.78 38.46 32.59 60–80% White 94.98 92.26 98.38 85.41 87.41 57.69 80–100% White 56.03 48.71 85.87 2.70 11.85 0.00 Distance for choice Avg. distance to nearest charter 0.57 1.95 0.43 2.12 2.82 2.21 Avg. distance to furthest charter (within 10 miles) 9.12 8.46 8.71 7.23 7.34 4.96 Avg. distance to chosen charter 1.62 3.34 1.17 3.49 4.23 3.45 Chosen vs. nearest difference 1.05 1.39 0.74 1.37 1.41 1.24 Distance to each choice set 0–20% White 4.50 5.38 4.17 4.19 4.58 3.12 20–40% White 5.03 4.52 5.00 4.62 4.45 2.46 40–60% White 4.81 5.79 3.56 4.71 6.14 4.85 60–80% White 6.67 5.58 5.69 5.56 5.22 5.19 80-100% White 8.57 5.24 8.07 4.79 4.43 4.98 Note: students who transferred to charter schools in 2011–12 in the same metropolitan area as their former TPS. Education Policy Analysis Archives Vol. 25 No. 22 18 For both elementary and high school charter choosers, the average distance to their chosen charter was usually a further distance than the average distance to the closest charter, suggesting that proximity is not the only factor influencing decision making. As described above, most choosers have charter schools of varying racial composition in their choice set. Yet, large percentages of Black and Latino movers chose schools with few White students while most White choosers chose majority White charter schools. For White students of all grade levels, the closest charter schools, on average, were 20-40% White schools. Black and Latino K-7 students typically had to travel a little further to attend schools with larger percentages of White students. High schools with greater shares of White students were also further for Black and Latino students, but these longer distances may be less daunting for older students (and their families) given expanded modes of transportation such as personal cars and public transportation. This is at least suggestive of a possible role that charter school racial composition might play. The next stage of our analysis helps us to understand the influence of proximity, school size, and racial composition in the choices being made by students leaving a TPS for a charter school. Examining Students’ Charter School Enrollment Patterns We now try to ascertain the enrollment behavior of students who have at least two charter school quintile types (e.g., by racial composition, 0–20% White, 20–40% White, etc.) in their choice set. For ease of interpretation, we have focused on three factors: racial composition (measured in quintiles of White percentage), distance, and school enrollment (measured by logged enrollment).20 While there are undoubtedly additional factors that influence student and families’ charter school choices and ultimate enrollment, other work suggests that school mission added little value to predictive models and did not affect the relationship between our variables of interest and student preferences (Bifulco & Ladd, 2007). Results are presented in a conditional logistic regression where positive coefficients reflect a greater likelihood and negative coefficients represent a lower likelihood. In some instances, we have added odds ratios to the text where it helps with interpretation. Across all analyses by student race and grade, the relationship with distance to the school is negative, indicating that, unsurprisingly, students enrolled in charter schools that were closer rather than further away (Table 5). Additionally, older students had fewer choice types (as measured by the quintiles of White percentage of students) in their choice sets than younger students did. For instance, on average, Black, K–7 students had an average of 4.8 choice types compared to the typical Black high school student who had an average of 3.5 choice types. Likewise, Latino K–7 students had 0.6 more choice types than their high school peers while White K–7 students had 0.8 more choice types than their high school peers. These similarities aside, we see distinct patterns in students’ choice of charter schools by their racial composition particularly when comparing Black and Latino students to White students. Black student movers to charter schools were much less likely to enroll in schools with higher percentages of White students (in comparison to our base category of 0–20% White students), holding distance and enrollment constant.21 These patterns hold for younger and older 20 We also checked to see whether the relationship with distance was curvilinear by including a variable for distance squared in our models. It was not significant, and thus we only include distance in our final models. 21 All results should be interpreted as being compared to the reference group of 0–20% White. We originally used 40–60% White as a reference group under the assumption that such a balanced student composition would reflect racial diversity and would be easier to interpret preferences in relation to this category. Unfortunately, there were very few K–7 schools and almost no high schools that fit this category making it a poor choice for a reference category. School choice and racial segregation in Pennsylvania 19 students and all results are statistically significant. The elementary schools with the highest percentage of White students, 60–80% and 80–100% White, were those in which there were the largest negative coefficients. When the coefficients are transformed into odds ratios, we find that that Black students were about 99% less likely to attend a charter school with between 60–100% White students compared to a charter school with 0–20% White students. These large disparities suggest that Black students were either resistant to choosing such schools or those schools did not recruit or admit a diverse applicant pool. Although we observe a similar pattern with high schools, Blacks were the least likely to enroll in a school that was between 40–60% White, though this is likely due to the relative scarcity of such options. Notably, on average, an increase in enrollment size was positively associated with enrolling in a particular high school for Black students. The patterns for Latino students were very similar to those of Black students. Simply put, Latino charter school choosers were significantly less likely to enroll in charter schools with a higher percentage of White students in comparison when holding distance and enrollment constant. As the percentage of White students increased, Latino high school students were less likely to enroll in a given school. However, for K–7 students, all four alternatives were less chosen in comparison to schools that were 0–20% White. In fact, a Latino child was between 95% and 98% less likely to enroll in any of the choices where more than a fifth of the students were White. Like all other students, Latino students enrolled in schools that were closer, but in comparison to students of other racial groups, the magnitude of the coefficient was higher, meaning that distance was more influential for them. And similar to Black and White students, the magnitude of distance was higher for older students. The patterns for White students who moved from TPSs to charter schools are quite distinct from Black and Latino students. In contrast to their other-race peers, White students were significantly more likely to enroll in schools with higher shares of White students in comparison to the reference group schools of 0–20% White students, holding distance and enrollment constant. The only exception is 40–60% White schools, which is not significant for K–7 or high school students; the lack of statistical significance may be due to the relatively few diverse charter schools. Even when students had a charter school option in the 40–60% White category, it was likely to only be one or two schools. White students were significantly more likely to enroll in majority White schools, although slight variation in racial composition existed based on age of student. Among younger students, the largest value indicates a preference for 80–100% White schools. In fact, the odds of a White student enrolling in a charter school with 80–100% other Whites is six times as high as that of the odds of them enrolling in a school with 0–20% Whites. For older students, 60–80% White schools have the highest positive coefficient indicating particular value by potential White transferees. The odds of a White student enrolling in a charter school with between 60% and 80% other White students is 4.8 times as likely as the odds of them attending a school with fewer than 20% White students. Thus, despite the slight differences, the overall pattern is clear: White students who are transferring to charter schools enrolled in schools with more White students than the reference group (less than 20% White students) and were especially likely to enroll schools with a majority of White students. In terms of enrollment, K–7 Whites were more likely attend a school as elementary enrollment increased, but size was not a significant factor for high school students. Education Policy Analysis Archives Vol. 25 No. 22 20 Table 5 Conditional Logit for Students Choosing Charter Schools with Different Racial Composition, by Race and Grade Black Students Latino Students White Students K–7 Grade 8–12 K–7 Grade 8–12 K–7 Grade 8– 12 0–20% White 20–40% White -2.13*** (0.10) -3.54*** (0.16) -3.42*** (0.21) -3.14 *** (0.27) 0.93*** (0.17) 1.30*** (0.22) 40–60% White -4.00*** (0.18) -20.30 *** (2.41) -3.96*** (0.32) -7.33 ** (2.62) -0.19 (0.21) -1.53 (0.86) 60–80% White -4.38 *** (-0.18) -10.78 *** (-0.84) -3.87*** (-0.28) -4.86*** (-0.92) 0.77*** (0.17) 1.57*** (-0.25) 80–100% White -5.06*** (0.34) -13.37 *** (1.13) -3.58*** (0.36) -9.33*** (1.56) 1.80*** (0.19) 0.87* (0.38) Distance (miles) -0.48*** (0.03) -0.64*** (0.05) -0.54*** (0.07) -0.86*** (0.08) -0.35*** (0.03) -0.40*** (0.05) Enrollment (logged) -0.45 (0.25) 14.47*** (1.67) 2.08*** (0.29) 4.29* (1.84) 0.63*** (0.15) 0.10 (0.34) Cases 2,618 2,793 671 915 538 319 Observations 12,458 9,831 3045 3,534 2215 1,067 Average Number of choices 4.8 3.5 4.5 3.9 4.1 3.3 *** p<.001; ** p<.01; * p <.05 Note: Student had to have lived in the same metropolitan in consecutive years, moved from TPS to charter in the second year, and had at least 2 charter composition alternatives in their choice set. Discussion This research illuminates the extent to which the rapidly growing number of students transferring to charters in Pennsylvania are making choices that are more segregative, and begins to explore reasons for such choices. Research confirms the importance of attending diverse schools for students of all racial groups. Minority students in more diverse school settings have higher short- term and long-term academic outcomes than those who attend racially isolated minority schools (Bohrnstedt, Kitmitto, Ogutm Sherman, & Chan, 2015; Johnson, 2012; Mickelson, 2015; Saatcioglu, 2010). A range of benefits in diverse schools for White students as well as students of color include reduced prejudice and higher likelihood of living and working in integrated environments as adults (Goldsmith, 2010; Stearns, 2010; Wells & Crain, 1994). Because these benefits have important implications for communities’ well-being and social cohesion within a democratic society (Mickelson & Nkomo, 2012), federal law and policies have endorsed the importance of policies aimed at creating diverse schools and/or reducing racial isolation (U.S. Departments of Education and Justice, 2011). In 2016, the U.S. Department of Education strongly endorsed using socioeconomic diversity as a school turnaround strategy and requested $120 million for a new competitive grant program to support local diversity initiatives. Additionally, some states have explicit racial diversity provisions in their charter school laws, which may reflect both this understanding of the benefits of integrated schools and the ways in which school choice was used in the aftermath of Brown to avoid desegregation (Frankenberg et al., 2011). More recently, there has been growing attention to School choice and racial segregation in Pennsylvania 21 questions of equity and access in charter schools. One such indicator of that interest is the formation of a network of charter schools that seek to implement practices in order to draw a diverse group of students (Kahlenberg & Potter, 2012; Wohlstetter, Smith, & Farrell, 2013). Additionally, the federal government recently released guidance about how charter schools should protect the civil rights of all interested students (U.S. Department of Education, 2014). The federal government has also included funding preferences for diversity in charter school funding as a way to incentivize less segregated charter schools. This article advances our understanding of charter schools and racial segregation in four ways. First, as the percentage of students transferring to charter schools has increased rapidly in recent years, Black and Latino students tend to move to charter schools with majorities of same race students. Moreover, Black and Latino students experience particularly large increases in racial isolation when transferring to charter schools. Given findings about the importance of reducing racial isolation and increasing school diversity, these patterns are concerning especially considering the high levels of segregation already existing in TPSs in Pennsylvania. Second, in seeking to understand why such segregation is occurring, we find that most students transferring to charter schools had a demographic range of charter school options (although, notably, students were least likely to have schools with the most diversity, 40–60% White, as part of their choice sets). Third, while distance is certainly an influential factor in students’ charter school choices—closer schools are more preferred—it is not the only factor given that the average distance to students’ charter school choice is further than their closest charter school option. For students in charter-dense metropolitan areas, the marketplace may give them a number of options within a relatively short distance (Henig & MacDonald, 2002). In fact, the average distance to a chosen charter school was at least twice as far as the nearest charter school for Black and Latino students regardless of age group. In the case of Black K–7 students, the average distance of a chosen charter school was almost four times greater than the nearest charter school, which is especially noteworthy considering the challenges of transporting younger children. Additionally, White students traveled, on average, an additional mile for their chosen school compared to their closest TPS. Finally, the enrollment patterns in terms of charter school demographics indicate sharp differences between Black and Latino choosers who tend not to enroll in schools with higher percentages of White students and White choosers who were more likely to enroll in charter schools that are at least 60% White. Such diverging behaviors, holding other factors constant, complicate efforts to create diverse schools when allowing for extensive family choice—and may help to explain the relatively few racially diverse charter options that exist. These findings have important policy implications including the need to consider opening charter schools that provide demographic diversity for potential students and to consider efforts to inform students more broadly of their charter school options. One possible policy mechanism would be to encourage the use of weighted lotteries that would prioritize certain groups in charter school admissions that would enhance diversity where schools have more demand than available seats. An encouraging finding, for integration efforts, is that proximity is not an overwhelmingly decisive factor in charter school choice. A state with a charter school context like Pennsylvania’s allows the opportunity to overcome district boundary lines that has left the state’s public schools deeply segregated (Kotok & Reed, 2015). This presumes, of course, that both urban and suburban parents know what charter school options are available to them, particularly when located in other districts. A key implication is the need to educate the public about the benefits to students of all races/ethnicities that stem from attending racially diverse schools, whether a charter school or a TPS. Further research should examine how other factors influence how students determine school choice decisions, including qualitative work to understand the decision-making contexts related to Education Policy Analysis Archives Vol. 25 No. 22 22 charter school choice. For instance, to what extent do parents consider achievement and to what degree do certain schools target certain families through advertising, outreach, or even a school name/theme (e.g., a classical curriculum)? How do charter school selection practices, where charter schools are over-subscribed, contribute to the patterns described here? However, at least in Pennsylvania, it is difficult to identify oversubscribed charter schools, as the Pennsylvania Coalition of Charter Schools acknowledges that there is no universal system for accurately measuring waiting lists and some charter schools misrepresent this data as to seem more popular (Mezzacappa, 2014). Further, given the preference for selecting charter schools with same-race peers, future research should examine how strong this preference is to have children attend racially segregated schools. While some qualitative research has explored why White parents make self-segregative choices for their children’s schooling (e.g., Roda & Wells, 2013), research on why Black and Latino parents are also making segregating choices is needed. Such research has implications for advocates who seek to further diverse schooling options. These findings help us understand how this growing form of choice may further or impede efforts to attain educational equity for all students. While policymakers expand school choice through the provision of charter schools—which is rapidly occurring in Pennsylvania and across the country—such an expansion may involve a tradeoff with other policy goals such as fostering racially integrated educational environments. Policy efforts should focus on how outreach efforts can appeal to families across racial lines, and help to fully inform families of all schooling options, including those that will be effective and offer diverse student bodies. Finally, charter school authorizers may wish to consider charter school effectiveness in attracting (and retaining) diverse groups of students in granting or renewing school charters. References Bell, C. A. (2009). All choices created equal? The role of choice sets in the selection of schools. Peabody Journal of Education, 84, 191–208. https://doi.org/10.1080/01619560902810146 Berends, M. (2015). Sociology and school choice: What we know after two decades of charter schools. Annual Review of Sociology, 41, 159–80. https://doi.org/10.1146/annurev-soc-073014- 112340 Bettinger, E. (2005). The effect of charter schools on charter students and public schools. Economics of Education Review, 24, 133–147. https://doi.org/10.1016/j.econedurev.2004.04.009 Bifulco, R. & Ladd, H.F. (2007). School choice, racial segregation, and test-score gaps: Evidence from North Carolina's charter school program. Journal of Policy Analysis and Management, 26(1), 31–56. https://doi.org/10.1002/pam.20226 Billingham, C. M., & Hunt, M. O. (2016). School racial composition and parental choice: New evidence on the preferences of White parents in the United States. Sociology of Education, 89(2), 99–117. https://doi.org/10.1177/0038040716635718 Bohrnstedt, G., Kitmitto, S., Ogut, B., Sherman, D., & Chan, D. (2015). School composition and the black–White achievement gap (NCES 2015–018). U.S. Department of Education, Washington, DC: National Center for Education Statistics. Booker, K., Zimmer, R., & Buddin, R. (2005). The effects of charter schools on school peer composition. Santa Monica, CA: RAND Corporation. www.rand.org/pubs/working_papers/WR306 Buckley, J., & Schneider, M. (2007) Charter schools: Hope or hype? New Jersey: Princeton University Press. https://doi.org/10.1080/01619560902810146 https://doi.org/10.1146/annurev-soc-073014-112340 https://doi.org/10.1146/annurev-soc-073014-112340 https://doi.org/10.1016/j.econedurev.2004.04.009 https://doi.org/10.1002/pam.20226 https://doi.org/10.1177/0038040716635718 http://www.rand.org/pubs/working_papers/WR306 School choice and racial segregation in Pennsylvania 23 Buddin, R., & Zimmer, R. (2005). Student achievement in charter schools: A complex picture. Journal of Policy Analysis and Management 24(2), 351–371. https://doi.org/10.1002/pam.20093 Butler, J. S., Carr, D. A., Toma, E. F., & Zimmer, R. (2013). Choice in a world of new school types. Journal of Policy Analysis and Management 32(4), 785–806. https://doi.org/10.1002/pam.21711 Carlson, D., Lavery, L., & Witte, J. F. (2011). The determinants of interdistrict open enrollment flows evidence from two states. Educational Evaluation and Policy Analysis, 33(1), 76–94. https://doi.org/10.3102/0162373710388643 Carnoy M., Jacobsen, R., Mishel, L., & Rothstein, R. (2005). The charter school dust-up: Examining the evidence on enrollment and achievement. New York: Teachers College Press. Chingos, M. M., & West, M. R. (2015). The uneven performance of Arizona’s charter schools. Educational Evaluation and Policy Analysis, 37(1 suppl), 120S–134S. https://doi.org/10.3102/0162373715576077 Chubb, J. E., & Moe, T. M. (1990). Politics, markets, and America's schools. Washington, D.C: Brookings Institution. Clark, M. A., Gleason, P. M., Tuttle, C. C., & Silverberg, M. K. (2015). Do charter schools improve student achievement? Educational Evaluation and Policy Analysis, 37(4), 419–436. https://doi.org/10.3102/0162373714558292 Cobb, C. & Glass, G. (1999). Ethnic segregation in Arizona charter schools. Education Policy Analysis Archives, 7(1). https://doi.org/10.14507/epaa.v7n1.1999 Frankenberg, E., Siegel-Hawley, G., & Wang, J. (2011). Choice without equity: Charter school segregation. Education Policy Analysis Archives, 19(1). https://doi.org/10.14507/epaa.v19n1.2011 Friedman, M. (1955). The role of government in education. In R.A. Solo (Ed.), Economics and the public interest (123–144). New Brunswick, N.J.: Rutgers University Press. Finn Jr, C. E., Manno, B. V., & Wright, B. L. (2016). Charter schools at crossroads: Predicaments, paradoxes, possibilities. Cambridge: Harvard Education Press. Garcia, D. R. (2008). The impact of school choice on racial segregation in charter schools. Educational Policy 22, 805–829. https://doi.org/10.1177/0895904807310043 Goldsmith, P.R. (2010). Learning apart, living apart: How the racial and ethnic segregation of schools and colleges perpetuates residential segregation. Teachers College Record 112(6): 1602– 1630. Gulosino, C., & d’Entremont, C. (2011) Circles of influence: An analysis of charter school location and racial patterns at varying geographic scales. Educational Policy Analysis Archives, 19(8). https://doi.org/10.14507/epaa.v19n8.2011 Hastings, J. S. & Weinstein, J. M. (2008). Information, school choice, and academic achievement. The Quarterly Journal of Economics, 123(4), 1373–1414. https://doi.org/10.1162/qjec.2008.123.4.1373 Henig, J. R. and MacDonald, J. A. (2002). Locational decisions of charter schools: Probing the market metaphor. Social Science Quarterly, 83, 962–980. https://doi.org/10.1111/1540- 6237.00126 Hoffman, S. D., & Duncan, G. J. (1988). Multinomial and conditional logit discrete-choice models in demography. Demography, 25(3), 415–427. https://doi.org/10.2307/2061541 Holme, J. J. (2002). Buying homes, buying schools: School choice and the social construction of school quality. Harvard Educational Review, 72, 177–205. https://doi.org/10.17763/haer.72.2.u6272x676823788r Hoxby, C. M. (1996). Are efficiency and equity in school finance substitutes or complements? The Journal of Economic Perspectives, 10(4), 51–72. https://doi.org/10.1257/jep.10.4.51 https://doi.org/10.1002/pam.20093 https://doi.org/10.1002/pam.21711 https://doi.org/10.3102/0162373710388643 https://doi.org/10.3102/0162373715576077 https://doi.org/10.3102/0162373714558292 https://doi.org/10.14507/epaa.v7n1.1999 https://doi.org/10.14507/epaa.v19n1.2011 https://doi.org/10.1177/0895904807310043 https://doi.org/10.14507/epaa.v19n8.2011 https://doi.org/10.1162/qjec.2008.123.4.1373 https://doi.org/10.1111/1540-6237.00126 https://doi.org/10.1111/1540-6237.00126 https://doi.org/10.2307/2061541 https://doi.org/10.17763/haer.72.2.u6272x676823788r https://doi.org/10.1257/jep.10.4.51 Education Policy Analysis Archives Vol. 25 No. 22 24 Jabbar, H. (2015a). “Every Kid Is Money”: Market-like competition and school leader strategies in New Orleans. Educational Evaluation and Policy Analysis, 37(4), 638–659. https://doi.org/10.3102/0162373715577447 Jabbar, H. (2015b). Competitive networks and school leaders’ perceptions: The formation of an education marketplace in post-Katrina New Orleans. American Educational Research Journal, 52(6), 1093–1131. https://doi.org/10.3102/0162373715577447 Johnson, R. C. (2012). The grandchildren of Brown: The long legacy of school desegregation. Unpublished manuscript, Goldman School of Public Policy, University of California, Berkeley. Kahlenberg, R., & Potter, H. (2012). Diverse charter schools: Can racial and socio-economic integration promote better outcomes for students. Washington, DC: Poverty and Race Research Council. Retrieved from http://www.tcf.org/assets/ downloads/Diverse_Charter_Schools.pdf Kotok, S., & Reed, K. (2015). Is opportunity knocking or slipping away? Pennsylvania's increasing racial diversity and segregation in schools, 1989–2010. Los Angeles: Civil Rights Project/ Proyecto Derechos Civiles. Kotok, S., Frankenberg, E., Schafft, K., Fuller, E., & Mann, B. (2015, online first). School choice, racial segregation, and poverty concentration: Evidence from Pennsylvania. Educational Policy. https://doi.org/10.1177/0895904815604112 Lankford, H., & Wyckoff, J. (2005). Why are schools racially segregated? Implications for school choice policy. In J. Scott (Ed.), School choice and diversity: What the evidence says (pp. 9–26). New York: Teachers College Press. LaFleur, J. C. (2016). Locating Chicago's charter schools: A socio-spatial analysis. Education Policy Analysis Archives, 24. https://doi.org/10.14507/epaa.24.1745 Linn, R. L., & Welner, K. G. (Eds.). (2007). Race-conscious policies for assigning students to schools: Social science research and the Supreme Court cases. Washington, DC: National Academy of Education. Lubienski, C. (2007). Marketing schools consumer goods and competitive incentives for consumer information. Education and Urban Society, 40(1), 118–141. https://doi.org/10.14507/epaa.24.1745 Lubienski, C., & Dougherty, J. (2009). Mapping educational opportunity: Spatial analysis and school choices. American Journal of Education, 115(4), 485–491. https://doi.org/10.1086/599783 Lubienski, C., Gulosino, C., & Weitzel, P. (2009). School choice and competitive incentives: Mapping the distribution of educational opportunities across local education markets. American Journal of Education, 115(4), 601–647. https://doi.org/10.1086/599778 Lubienski, C., & Lee, J. (2016). Competitive incentives and the education market: How charter schools define themselves in metropolitan Detroit. Peabody Journal of Education, 91(1), 64–80. https://doi.org/10.1080/0161956X.2016.1119582 Mann, B., Kotok, S., Frankenberg, E., Fuller, E., & Schafft, K. (2016). Choice, cyber charter schools, and the educational marketplace for rural school districts. The Rural Educator, 37(3), 17–29. Massey, D. S., & Denton, N. A. (1988). The dimensions of residential segregation. Social Forces, 67, 281–315. https://doi.org/10.2307/2579183 McFadden, D. (1973). Conditional logit analysis of qualitative choice behavior. In Zarembka, P. (Ed.), Frontiers in Econometrics (pp. 105–142). New York: Academic Press. Mezzacappa, D. (2014, December 12). Head of state charter coalition: Not sure of size of waiting lists. The Notebook. Accessed on April 20, 2016, at http://thenotebook.org/articles/2014/12/12/head-of-state-charter-coalition-not-sure-of- size-of-waiting-lists https://doi.org/10.3102/0162373715577447 https://doi.org/10.3102/0162373715577447 https://doi.org/10.1177/0895904815604112 https://doi.org/10.14507/epaa.24.1745 https://doi.org/10.14507/epaa.24.1745 https://doi.org/10.1086/599783 https://doi.org/10.1086/599778 https://doi.org/10.1080/0161956X.2016.1119582 https://doi.org/10.2307/2579183 http://thenotebook.org/articles/2014/12/12/head-of-state-charter-coalition-not-sure-of-size-of-waiting-lists http://thenotebook.org/articles/2014/12/12/head-of-state-charter-coalition-not-sure-of-size-of-waiting-lists School choice and racial segregation in Pennsylvania 25 Mickelson, R. A. (2015). The cumulative disadvantages of first- and second-generation segregation for middle school achievement. American Educational Research Journal 52, 657–692. https://doi.org/10.3102/0002831215587933 Mickelson, R. A., & Nkomo, M. (2012). Integrated schooling, life-course outcomes, and social cohesion in multiethnic democratic societies. Review of Research in Education, 36, 197–238. https://doi.org/10.3102/0091732X11422667 Miron, G., Urschel, J. L., & Saxton, N. (2011). What makes KIPP work? A study of student characteristics, attrition, and school finance. Kalamazoo: Western Michigan University. Moe, T. (2001). School, vouchers, and the American public. Washington, D.C.: Brookings Institution Press. Morrison, M. (2004). An examination of Philadelphia’s school desegregation litigation. Perspectives on Urban Education, 3(1). Retrieved from http://www.urbanedjournal.org/archive/volume-3- issue-1-fall-2004 Nichols-Barrer, I., Gleason, P., Gill, B., & Tuttle, C. C. (2016). Student selection, attrition, and replacement in KIPP middle schools. Educational Evaluation and Policy Analysis, 38(1), 5–20. https://doi.org/10.3102/0162373714564215 Orfield, G., & Frankenberg, E. (2014). Increasingly segregated and unequal schools as courts reverse policy. Educational Administration Quarterly 50(6), 718–734. https://doi.org/10.1177/0013161X14548942 Orfield, G., & Frankenberg, E. (2013). Educational delusions? Why choice can deepen inequality and how to make it fair. Berkeley, CA: University of California Press. https://doi.org/10.1525/california/9780520274730.001.0001 Pennsylvania Department of Education. (2016). Enrollment reports and projections. Retrieved from http://www.education.pa.gov/Data-and- Statistics/Pages/Enrollment%20Reports%20and%20Projections.aspx#.V2yNUlNViko Renzulli, L.A. & Evans, L. (2005). School choice, charter schools, and White flight. Social Problems, 52, 398–418. https://doi.org/10.1525/sp.2005.52.3.398 Ritter, G., Jensen, N., Kisida, B., & McGee, J. (2010). A closer look at charter schools and segregation: Flawed comparisons lead to overstated conclusions. Education Next, 10(3), 69. Roda, A., & Wells, A. S. (2013). School choice policies and racial segregation: Where White parents’ good intentions, anxiety, and privilege collide. American Journal of Education, 119(2), 261–293. https://doi.org/10.1086/668753 Rothwell, J. (2012). Housing costs, zoning, and access to high-scoring schools. Washington, D.C.: The Brookings Institution. Saatcioglu, A. (2010). Disentangling school- and student-level effects of desegregation and resegregation in the dropout problem in urban high schools: Evidence from the Cleveland Municipal School District: 1977–1998. Teachers College Record, 112(5), 1291–1442. Saporito, S., & Sohoni, D. (2006). Coloring outside the lines: Racial segregation in public schools and their attendance boundaries. Sociology of Education, 79, 81–105. https://doi.org/10.1177/003804070607900201 Scafidi, B. (2015). The integration anomaly: Comparing the effects of K–12 education delivery models on segregation in schools. Indianapolis, IN: The Friedman Foundation. Retrieved November 16, 2015, from http://www.edchoice.org/research/the-integration-anomaly/ Schafft, K., Frankenberg, E., Fuller, E., Hartman, W., Kotok, S., & Mann, B. (2014). Assessing the enrollment trends and financial impact of charter schools on rural and non-rural school districts in Pennsylvania. Harrisburg, PA: Center for Rural Pennsylvania. Retrieved from http://www.rural.palegislature.us/documents/reports/Charter_School_2014.pdf Schneider, M., Marschall, M., Teske, P., & Roch, C. (1998). School choice and culture wars in the classroom: What different parents seek from education. Social Science Quarterly 79(3): 489–501. https://doi.org/10.3102/0002831215587933 https://doi.org/10.3102/0091732X11422667 http://www.urbanedjournal.org/archive/volume-3-issue-1-fall-2004 http://www.urbanedjournal.org/archive/volume-3-issue-1-fall-2004 https://doi.org/10.3102/0162373714564215 https://doi.org/10.1177/0013161X14548942 https://doi.org/10.1525/california/9780520274730.001.0001 http://www.education.pa.gov/Data-and-Statistics/Pages/Enrollment%20Reports%20and%20Projections.aspx#.V2yNUlNViko http://www.education.pa.gov/Data-and-Statistics/Pages/Enrollment%20Reports%20and%20Projections.aspx#.V2yNUlNViko https://doi.org/10.1525/sp.2005.52.3.398 https://doi.org/10.1086/668753 https://doi.org/10.1177/003804070607900201 http://www.edchoice.org/research/the-integration-anomaly/ http://www.rural.palegislature.us/documents/reports/Charter_School_2014.pdf Education Policy Analysis Archives Vol. 25 No. 22 26 Scott, J. (2011). School choice as a civil right: The political construction of a claim and its implications for school desegregation. In E. Frankenberg & E. DeBray (Eds.), Integrating schools in a changing society: New policies and legal options for a multiracial generation (pp. 32–52). Chapel Hill, NC: University of North Carolina Press. https://doi.org/10.5149/9780807869208_frankenberg.6 Siegel-Hawley, G., & Frankenberg, E. (2011). Does law influence charter school diversity? An analysis of federal and state legislation. Michigan Journal of Race & Law 16(2): 321-376. Stearns, E. (2010). Long-term correlates of high school racial composition: Perpetuation theory reexamined. Teachers College Record 112(6), 1654–1678. Stein, M. L. (2015). Public school choice and racial sorting: An examination of charter schools in Indianapolis. American Journal of Education, 121(4), 597–627. https://doi.org/10.1086/681920 Teske, P., & Schneider, M. (2001). What research can tell policymakers about school choice. Journal of Policy Analysis and Management, 20(4), 609–631. https://doi.org/10.1002/pam.1020 U.S. Department of Education. (2014). Dear Colleague Letter. Washington, D.C. Accessed on March 27, 2015 at http://www2.ed.gov/about/offices/list/ocr/letters/colleague-201405- charter.pdf U.S. Departments of Education and Justice. (2011). Guidance on the voluntary use of race to achieve diversity and avoid racial isolation in elementary and secondary schools. Washington DC. U.S. Department of Transportation. (2008). NHTS Brief. Accessed on January 5, 2016, at http://nhts.ornl.gov/briefs/Travel%20To%20School.pdf Villavicencio, A. (2013). "It's Our Best Choice Right Now": Exploring How Charter School Parents Choose. Education Policy Analysis Archives, 21(81). https://doi.org/10.14507/epaa.v21n81.2013 Weiher, G. R., & Tedin, K. L. (2002). Does choice lead to racially distinctive schools? Charter schools and household preferences. Journal of Policy Analysis and Management, 21(1), 79–92. https://doi.org/10.1002/pam.1041 Wells, A. S., & Crain R. L. (1994). Perpetuation theory and the long-term effects of school desegregation. Review of Educational Research, 6, 531–555. https://doi.org/10.3102/00346543064004531 Wohlstetter, P., Smith, J., & Farrell, C. C. (2013). Choices and challenges: Charter school performance in perspective. Cambridge, MA: Harvard Education Press. Zimmer, R., Gill, B., Booker, K., Lavertu, S., Sass, T.R., & Witte, J. (2009). Charter schools in eight states: Effects on achievement, attainment, integration, and competition. Santa Monica, CA: RAND. https://doi.org/10.5149/9780807869208_frankenberg.6 https://doi.org/10.1086/681920 https://doi.org/10.1002/pam.1020 http://www2.ed.gov/about/offices/list/ocr/letters/colleague-201405-charter.pdf http://www2.ed.gov/about/offices/list/ocr/letters/colleague-201405-charter.pdf http://nhts.ornl.gov/briefs/Travel%20To%20School.pdf https://doi.org/10.14507/epaa.v21n81.2013 https://doi.org/10.1002/pam.1041 https://doi.org/10.3102/00346543064004531 School choice and racial segregation in Pennsylvania 27 Appendix A Racial Isolation for Students Moving from a TPS to Charters in Philadelphia Metropolitan Area, 2010-11 to 2011-2012 Number of Students Sending TPS Receiving Charter Difference in isolation experienced by transferring student Black Grades K-7 2,258 74.53 84.00 9.47 Grades 8-12 2,642 72.42 81.75 9.33 Total 4,900 73.39 82.79 9.39 White Grades K-7 360 58.94 63.67 4.73 Grades 8-12 214 57.10 66.66 9.56 Total 574 58.25 64.79 6.53 Latino Grades K-7 549 42.38 55.00 12.62 Grades 8-12 921 39.07 55.22 16.15 Total 1470 40.31 55.14 14.83 Source: Pennsylvania Department of Education; NCES Common Core of Data; student grade level is from 2010-11. Education Policy Analysis Archives Vol. 25 No. 22 28 Appendix B Racial Isolation for Students Moving from a TPS to Charters in Non-Philadelphia Metropolitan Areas, 2010-11 to 2011-2012 Number of Students Sending TPS Receiving Charter Difference in isolation experienced by transferring student Black Grades K-7 415 50.89 60.87 9.98 Grades 8-12 185 49.44 62.19 12.75 Total 600 50.44 61.28 10.83 White Grades K-7 236 66.84 58.38 -8.46 Grades 8-12 135 66.99 54.85 -12.14 Total 371 66.89 57.10 -9.80 Latino Grades K-7 135 46.21 49.15 2.94 Grades 8-12 26 39.06 44.71 5.65 Total 161 45.06 48.43 3.38 Note: Pennsylvania Department of Education; NCES Common Core of Data; student grade level is from 2010-11. School choice and racial segregation in Pennsylvania 29 Appendix C Conditional Logit for Students with an IEP Choosing Charter Schools with Different Racial Composition, by Race and Grade Black Students Latino Students White Students K-7 Grade 8-12 K-7 Grade 8-12 K-7 Grade 8-12 0-20% White 20-40% White -2.53*** (0.29) -2.81*** (0.23) -6.01*** (1.26) -8.50** (2.41) 0.13 (0.40) 0.72 (0.49) 40-60% White -3.85*** (0.44) -21.86 (26.87) -6.80 ** (2.04) -90.44** (30.10) -1.37* (0.58) -15.26 (1002.21) 60-80% White -7.27*** (1.11) -4.60*** (0.96) -5.08*** (0.97) -38.78** (11.44) -0.46 (0.45) -0.80 (0.60) 80-100% White -6.40*** (1.53) -7.30 (3.88) -3.73** (1.08) -62.81 (9979.43) 1.15 * (0.50) 0.17 (0.85) Distance (miles) -0.70*** (0.10) -0.48*** (0.10) -0.45* (0.19) -1.80** (0.60) -0.36*** (0.09) -0.53*** (0.14) Enrollment (logged) -0.07 (0.65) 0.90 (1.95) 5.10 *** (1.36) 67.38** (22.39) 1.81*** (0.51) 0.09 (0.61) Cases 401 622 83 206 80 46 Observations 1,906 2,199 381 795 305 161 Average Number of choices 3.8 3.5 4.6 3.9 3.8 3.5 *** p<.001; ** p<.01; * p <.05 Note: Student had to have lived in the same metropolitan in consecutive years, moved from TPS to charter in the second year, and had at least two charter composition alternatives in their choice set. Education Policy Analysis Archives Vol. 25 No. 22 30 About the Authors Erica Frankenberg Pennsylvania State University euf10@psu.edu http://orcid.org/0000-0002-9224-7734 Erica Frankenberg is an associate professor of education and demography at the Pennsylvania State University. Her research interests focus on racial desegregation and inequality in K –12 schools, school choice and racial stratification, and the connections between school segregation and other metropolitan policies. Stephen Kotok University of Texas at El Paso sakotok@utep.edu Stephen Kotok is an assistant professor in the Department of Educational Leadership and Foundations at the University of Texas at El Paso. His research focuses on the extent that schools expand opportunity or act as a stratifying mechanism. His specific areas of research include educational leadership, school climate, charter schools, and segregation. Kai Schafft Pennsylvania State University kas45@psu.edu Kai Schafft is an Associate Professor of Education in the Department of Education Policy Studies at Penn State. Trained as a rural sociologist, his research focuses on the relationship between the well-being of rural schools and the communities they serve. Bryan Mann Pennsylvania State University bmann4@gmail.com Bryan Mann is a PhD graduate of the Educational Theory and Policy program at Pennsylvania State University and in August will start as an assistant professor in the Department of Educational Leadership, Policy, and Technology Studies at the University of Alabama. His research focuses on charter school policies and the effects they have on organizational development and leadership, student performance, and school and community demographics. mailto:euf10@psu.edu http://orcid.org/0000-0002-9224-7734 mailto:sakotok@utep.edu mailto:kas45@psu.edu mailto:bmann4@gmail.com School choice and racial segregation in Pennsylvania 31 education policy analysis archives Volume 25 Number 22 March 13, 2017 ISSN 1068-2341 Readers are free to copy, display, and distribute this article, as long as the work is attributed to the author(s) and Education Policy Analysis Archives, it is distributed for non- commercial purposes only, and no alteration or transformation is made in the work. More details of this Creative Commons license are available at http://creativecommons.org/licenses/by-nc-sa/3.0/. All other uses must be approved by the author(s) or EPAA. EPAA is published by the Mary Lou Fulton Institute and Graduate School of Education at Arizona State University Articles are indexed in CIRC (Clasificación Integrada de Revistas Científicas, Spain), DIALNET (Spain), Directory of Open Access Journals, EBSCO Education Research Complete, ERIC, Education Full Text (H.W. Wilson), QUALIS A2 (Brazil), SCImago Journal Rank; SCOPUS, SOCOLAR (China). Please send errata notes to Audrey Amrein-Beardsley at audrey.beardsley@asu.edu Join EPAA’s Facebook community at https://www.facebook.com/EPAAAAPE and Twitter feed @epaa_aape. http://www.doaj.org/ mailto:audrey.beardsley@asu.edu https://www.facebook.com/EPAAAAPE Education Policy Analysis Archives Vol. 25 No. 22 32 education policy analysis archives editorial board Lead Editor: Audrey Amrein-Beardsley (Arizona State University) Editor Consultor: Gustavo E. Fischman (Arizona State University) Associate Editors: David Carlson, Margarita Jimenez-Silva, Eugene Judson, Mirka Koro-Ljungberg, Scott Marley, Jeanne M. Powers, Iveta Silova, Maria Teresea Tatto (Arizona State University) Cristina Alfaro San Diego State University Ronald Glass University of California, Santa Cruz R. Anthony Rolle University of Houston Gary Anderson New York University Jacob P. K. Gross University of Louisville A. G. Rud Washington State University Michael W. Apple University of Wisconsin, Madison Eric M. Haas WestEd Patricia Sánchez University of University of Texas, San Antonio Jeff Bale OISE, University of Toronto, Canada Julian Vasquez Heilig California State University, Sacramento Janelle Scott University of California, Berkeley Aaron Bevanot SUNY Albany Kimberly Kappler Hewitt University of North Carolina Greensboro Jack Schneider College of the Holy Cross David C. Berliner Arizona State University Aimee Howley Ohio University Noah Sobe Loyola University Henry Braun Boston College Steve Klees University of Maryland Nelly P. Stromquist University of Maryland Casey Cobb University of Connecticut Jaekyung Lee SUNY Buffalo Benjamin Superfine University of Illinois, Chicago Arnold Danzig San Jose State University Jessica Nina Lester Indiana University Adai Tefera Virginia Commonwealth University Linda Darling-Hammond Stanford University Amanda E. Lewis University of Illinois, Chicago Tina Trujillo University of California, Berkeley Elizabeth H. DeBray University of Georgia Chad R. Lochmiller Indiana University Federico R. Waitoller University of Illinois, Chicago Chad d'Entremont Rennie Center for Education Research & Policy Christopher Lubienski University of Illinois, Urbana-Champaign Larisa Warhol University of Connecticut John Diamond University of Wisconsin, Madison Sarah Lubienski University of Illinois, Urbana-Champaign John Weathers University of Colorado, Colorado Springs Matthew Di Carlo Albert Shanker Institute William J. Mathis University of Colorado, Boulder Kevin Welner University of Colorado, Boulder Michael J. Dumas University of California, Berkeley Michele S. Moses University of Colorado, Boulder Terrence G. Wiley Center for Applied Linguistics Kathy Escamilla University of Colorado, Boulder Julianne Moss Deakin University, Australia John Willinsky Stanford University Melissa Lynn Freeman Adams State College Sharon Nichols University of Texas, San Antonio Jennifer R. Wolgemuth University of South Florida Rachael Gabriel University of Connecticut Eric Parsons University of Missouri-Columbia Kyo Yamashiro Claremont Graduate University Amy Garrett Dikkers University of North Carolina, Wilmington Susan L. Robertson Bristol University, UK Gene V Glass Arizona State University Gloria M. Rodriguez University of California, Davis School choice and racial segregation in Pennsylvania 33 archivos analíticos de políticas educativas consejo editorial Editor Consultor: Gustavo E. Fischman (Arizona State University) Editores Asociados: Armando Alcántara Santuario (Universidad Nacional Autónoma de México), Jason Beech, (Universidad de San Andrés), Angelica Buendia, (Metropolitan Autonomous University), Ezequiel Gomez Caride, (Pontificia Universidad Católica Argentina), Antonio Luzon, (Universidad de Granada), José Luis Ramírez, Universidad de Sonora) Claudio Almonacid Universidad Metropolitana de Ciencias de la Educación, Chile Juan Carlos González Faraco Universidad de Huelva, España Miriam Rodríguez Vargas Universidad Autónoma de Tamaulipas, México Miguel Ángel Arias Ortega Universidad Autónoma de la Ciudad de México María Clemente Linuesa Universidad de Salamanca, España José Gregorio Rodríguez Universidad Nacional de Colombia, Colombia Xavier Besalú Costa Universitat de Girona, España Jaume Martínez Bonafé Universitat de València, España Mario Rueda Beltrán Instituto de Investigaciones sobre la Universidad y la Educación, UNAM, México Xavier Bonal Sarro Universidad Autónoma de Barcelona, España Alejandro Márquez Jiménez Instituto de Investigaciones sobre la Universidad y la Educación, UNAM, México José Luis San Fabián Maroto Universidad de Oviedo, España Antonio Bolívar Boitia Universidad de Granada, España María Guadalupe Olivier Tellez, Universidad Pedagógica Nacional, México Jurjo Torres Santomé, Universidad de la Coruña, España José Joaquín Brunner Universidad Diego Portales, Chile Miguel Pereyra Universidad de Granada, España Yengny Marisol Silva Laya Universidad Iberoamericana, México Damián Canales Sánchez Instituto Nacional para la Evaluación de la Educación, México Mónica Pini Universidad Nacional de San Martín, Argentina Juan Carlos Tedesco Universidad Nacional de San Martín, Argentina Gabriela de la Cruz Flores Universidad Nacional Autónoma de México Omar Orlando Pulido Chaves Instituto para la Investigación Educativa y el Desarrollo Pedagógico (IDEP) Ernesto Treviño Ronzón Universidad Veracruzana, México Marco Antonio Delgado Fuentes Universidad Iberoamericana, México José Luis Ramírez Romero Universidad Autónoma de Sonora, México Ernesto Treviño Villarreal Universidad Diego Portales Santiago, Chile Inés Dussel, DIE-CINVESTAV, México Paula Razquin Universidad de San Andrés, Argentina Antoni Verger Planells Universidad Autónoma de Barcelona, España Pedro Flores Crespo Universidad Iberoamericana, México José Ignacio Rivas Flores Universidad de Málaga, España Catalina Wainerman Universidad de San Andrés, Argentina Ana María García de Fanelli Centro de Estudios de Estado y Sociedad (CEDES) CONICET, Argentina Juan Carlos Yáñez Velazco Universidad de Colima, México javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/819') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/820') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/4276') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/1609') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/825') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/797') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/555') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/823') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/798') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/814') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/801') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/1617') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/2703') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/802') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/816') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/826') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/804') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/3264') Education Policy Analysis Archives Vol. 25 No. 22 34 arquivos analíticos de políticas educativas conselho editorial Editor Consultor: Gustavo E. Fischman (Arizona State University) Editoras Associadas: Geovana Mendonça Lunardi Mendes (Universidade do Estado de Santa Catarina), Marcia Pletsch, Sandra Regina Sales (Universidade Federal Rural do Rio de Janeiro) Almerindo Afonso Universidade do Minho Portugal Alexandre Fernandez Vaz Universidade Federal de Santa Catarina, Brasil José Augusto Pacheco Universidade do Minho, Portugal Rosanna Maria Barros Sá Universidade do Algarve Portugal Regina Célia Linhares Hostins Universidade do Vale do Itajaí, Brasil Jane Paiva Universidade do Estado do Rio de Janeiro, Brasil Maria Helena Bonilla Universidade Federal da Bahia Brasil Alfredo Macedo Gomes Universidade Federal de Pernambuco Brasil Paulo Alberto Santos Vieira Universidade do Estado de Mato Grosso, Brasil Rosa Maria Bueno Fischer Universidade Federal do Rio Grande do Sul, Brasil Jefferson Mainardes Universidade Estadual de Ponta Grossa, Brasil Fabiany de Cássia Tavares Silva Universidade Federal do Mato Grosso do Sul, Brasil Alice Casimiro Lopes Universidade do Estado do Rio de Janeiro, Brasil Jader Janer Moreira Lopes Universidade Federal Fluminense e Universidade Federal de Juiz de Fora, Brasil António Teodoro Universidade Lusófona Portugal Suzana Feldens Schwertner Centro Universitário Univates Brasil Debora Nunes Universidade Federal do Rio Grande do Norte, Brasil Lílian do Valle Universidade do Estado do Rio de Janeiro, Brasil Flávia Miller Naethe Motta Universidade Federal Rural do Rio de Janeiro, Brasil Alda Junqueira Marin Pontifícia Universidade Católica de São Paulo, Brasil Alfredo Veiga-Neto Universidade Federal do Rio Grande do Sul, Brasil Dalila Andrade Oliveira Universidade Federal de Minas Gerais, Brasil