Urban School Choice and Integration

Abstract

We examine the impact of charter schools on school integration in the Little Rock, Arkansas metropolitan area. We find that charters are less likely to be hyper-segregated than traditional public schools (TPS), but TPS have compositions more closely reflecting the region. However, differences in each case are slight. Using student-level data to follow students who left TPS for charters, we find that most transfers improve integration levels at the schools they left. This finding is attributed to the fact that most transfers involve minority students leaving predominately minority schools or White students leaving predominantly White schools.

Keywords

urban education charter schools school choice educational policy

Introduction

While the academic impact of charter schools continues to be a source of debate, some researchers have begun to investigate the impact of enhanced school choice on racial diversity. Critics of school choice have suggested that when families have the freedom to select the schools their children attend, they will sort themselves into schools that are segregated along racial and ethnic lines (e.g., Berliner, Farrell, Huerta, & Mickelson, 2000; Cobb & Glass, 1999; Frankenberg, Siegel-Hawley, & Wang, 2010). Such segregation, they argue, would exacerbate racial achievement gaps by limiting the opportunities for students to be educated in ethnically and culturally diverse environments (Frankenberg et al., 2010).

However, there is a plausible argument to be made that increased choices in schooling options could have the opposite effect. In most public school systems, a student’s schooling options are almost entirely a function of where families live. Household income largely constrains housing choices, and the net effect is that neighborhoods are largely segregated along racial and class lines. As such, students end up attending schools that are as racially and economically segregated as the neighborhoods in which they reside. School choice, however, has the potential to detach the schooling choice from the housing choice; students can choose to travel across racially segregated neighborhood and district boundaries (Greene, 2005). As a result, it is possible that choice systems could lead to lower levels of segregation than systems based on residential assignment.

While theoretical arguments on both sides of this question are reasonable, this is an empirical question for which there are real data available to guide policymakers. In this article, we present the results of our analysis on the impacts of the increasing presence of charter schools on the racial composition of schools in the urban area of Little Rock, Arkansas. The setting for this research is particularly noteworthy. In 1957, after state and local officials refused to comply with the Brown v. Board of Education decision mandating desegregation, President Eisenhower ordered the 101st Airborne Division of the United States Army into Little Rock to enforce the court’s order. As a result of this historic showdown between the federal government and state and local segregationists, Little Rock has long served as a historical example of the nation’s strong resistance to racial integration during the civil rights movement.

Segregation in Little Rock schools continues to be a concern, and the recent expansion of charter schools has brought up new fears with regard to the district’s desegregation efforts. Some representatives of the Little Rock School District (LRSD) have recently charged that increases in charter school attendance will hamper desegregation efforts in traditional public schools (TPS; DeMillo, 2012). It is this claim that we investigate empirically.

Our analysis benefits from student-level data over a 6-year time period (2004-2005 to 2009-2010) when the number of charter schools in Little Rock increased from 3 to 15, and the total charter enrollment increased from 477 to 3,179 students. Student-level data enable us to track individual student movement from TPS to charter schools, which allows us to examine how the growth of the charter sector in the Little Rock area has influenced racial integration. Put simply, we are able to assess how student transfers to charter schools have affected the racial balance of schools in Little Rock.

Review of Literature and Methodologies

The evidentiary record concerning charter schools and their effect on racial integration is mixed with respect to the methods employed, the quality of available studies, and the tone of the findings. Here, we discuss past research findings, but first we present our thoughts on methodological approaches.

To assess which school sector has the better track record with respect to racial segregation or integration, one must first decide how to quantify the levels of segregation or integration in a given school. Past studies have defined segregation by labeling schools as “racially homogeneous” or “hyper-segregated” if they meet a certain threshold, such as having 90% or more of their students represented by a single race, or 90% or more from underrepresented racial backgrounds (e.g., Frankenberg et al., 2010). Of course, one problem with this measure of segregation is the arbitrary cutoff employed. A second problem is that this measure ignores the demographic makeup of the broader community, which may in fact be just as racially homogeneous. Nevertheless, many observers would argue that a school in which at least 90% of the students are from underrepresented racial groups, regardless of surroundings, is racially segregated.

The concept of racial integration is perhaps harder to define because it requires some benchmark for comparison. Integration in schools may be best understood as the extent to which the students in a larger geographic area are distributed evenly among the schools in that area. However, there is no consensus among researchers as to the appropriate geographic area that should be considered in an assessment of integration (Egalite & Mills, 2014; Greene, 2005; Greene & Mellow, 2000; Ritter, Jensen, Kisida, & McGee, 2010). Is a school integrated if half of the students are White and half are non-White? Is a school integrated if it reflects the racial composition of the entire nation, state, county, city, or school district? How close do the racial compositions need to be for observers to say that the composition of the school adequately reflects that of the broader community? The fundamental question here involves determining the appropriate geographic unit of comparison against which we judge levels of integration (Echenique & Fryer, 2005).

In our view, it is clearly not appropriate to measure integration with respect to the demographic makeup of children across geographic units as large as the nation or an entire state. These areas are simply too large and are not representative of the many diverse metropolitan areas and school districts across the nation. For example, we would not want to judge the racial integration of Miami’s schools, which serve large numbers of Hispanic students, based on the extent to which they mirror the population of the United States or even just the State of Florida. High levels of variation within most states make the statewide population an inappropriate benchmark for school integration.

In our view, a more appropriate benchmark for the question of racial integration is the larger metropolitan area because smaller geographic units such as cities and school districts are often segregated themselves. That is, we do not want to judge the integration of a school based on the extent to which its racial composition reflects that of a heavily segregated city or school district. For example, in an urban area with one predominantly White city (95% White) sitting next to a predominantly Black city (95% Black), we would not want to identify schools that are nearly entirely White or entirely Black as integrated simply because their student composition was similar to that of the immediate cities, which themselves are racially segregated. Therefore, we maintain that the most reasonable unit of analysis is the larger metropolitan area, which is large enough to capture some racial and economic diversity but small enough to allow reasonable mobility of students within the area.

As is already apparent, designing a study of racial integration in TPS and charter schools is not clear-cut and requires researchers to make design decisions that will likely influence the outcome of the study (Echenique & Fryer, 2007), and thus far we have only discussed the definitions of segregation and integration. To engage in the comparative study of relative racial integration, we must make one further set of decisions. Because nearly all schools will deviate from perfect integration, the relevant question is indeed one of comparison: How does racial integration in charter schools compare with TPS? Moreover, we must choose which set of TPS will serve as the counterfactual for the charter schools in question. That is, do we want to compare the level of integration of all the charter schools in a given state or in the nation with all of the TPS in the state or in the nation overall? Indeed, we would if each type of school were randomly sprinkled across the nation or at least across areas within states. However, as we know from past research, charter schools are located disproportionately in economically disadvantaged areas and in areas with high numbers of racial minorities (Henig & MacDonald, 2002).

Consequently, we argue that the best way to assess the relative racial integration of charter schools is to compare them with the TPS that students would likely have attended in the absence of charter schools. This means that comparison schools should not be so far from where charter schools locate that they, in effect, are not part of the choice-set that parents are considering. Thus, studies which employ different levels of analysis for each sector (individual charter schools vs. TPS districts) are not at all helpful for addressing the question of the relative racial segregation in charter schools.

In our view, there are various problems with the methods employed in studies on this question that call into question the conclusions drawn. The first major problem is that some evaluations have used average demographics (for school districts, for example) to discuss the issues of racial integration and segregation. However, the overall racial composition within a given school district tells us nothing about the distribution of students at the schools within the district and thus tells us nothing about racial integration. An example of a recent study of this type compares the racial compositions of individual charter schools with the racial composition of school districts as a whole (Miron, Urschel, Mathis, & Tornquist, 2010). Miron et al. conclude from these comparisons that charter schools are plagued by problematic levels of racial segregation because the racial composition within these charter schools did not perfectly reflect that of the nearby districts as a whole. However, the point that Miron et al. (2010) fail to make is that the racial compositions of the individual TPS within these districts also did not reflect that of the overall district. The question the authors fail to ask here was which set of individual schools within each sector experienced greater levels of segregation.

The authors of a high-profile study by the Civil Rights Project (Frankenberg et al., 2010) avoid the “apples to oranges” comparison problem described above, and they use a reasonable definition of segregation for each of the individual schools. In this study, Frankenberg et al. conclude that charter schools have a negative impact on integration. However, the study’s primary findings are unreliable because segregation in charter schools was compared with an inappropriate counterfactual—segregation in all TPS in the nation or within a given state. Highlighted findings, for example, include the following: “At the national level, 70% of black students attending charter schools attend intensely segregated minority charter schools (those with more than 90% of students coming from under-represented minority backgrounds) in contrast to 34% of black students attending a TPS” (Frankenberg et al., 2010, p. 4).

As noted earlier, such an analysis compares charter schools with TPS that are not near the geographic areas where charter schools locate. Similar studies by Frankenberg and Lee (2003) and Rapp and Eckes (2007) claim similar findings while incorporating similar flaws. A reanalysis of the same data used in Frankenberg et al. (2010) demonstrates that their results are highly sensitive to the geographic boundaries that define the comparison schools (Author, 2010). When more appropriate comparison regions are specified, charter schools and TPS in the same urban areas are found to have much more comparable levels of segregation. Given the flaws in the prior studies, the available research tells us little about the current state of racial integration or segregation in charter schools relative to that of TPS in the United States.

A number of relatively recent studies have the added benefit of employing student-level data that incorporate actual student transfers. Garcia (2008) performs an analysis of attendance patterns of individual students, Grades 2 to 9, over a 4-year period. Garcia finds that, at both the state and school levels, students who enroll in charter schools are entering more racially segregated environments than they experienced at their prior TPS. Bifulco, Ladd, and Ross (2009) find similar results in their analysis of charter school transfers in North Carolina. Black students, on average, transferred to schools that were proportionally more Black, whereas White students, on average, transferred to schools that were more White. While these studies show that students transferring to charter schools enter into more segregated environments, Zimmer et al. (2009) discover no differences between the sectors. Based on longitudinal data of student transfers to charter schools from eight different states, Zimmer et al. conclude that, in most cases, students who transferred to charter schools actually moved to schools with racial compositions not significantly different from the ones they previously attended. The authors summarize,

Across 21 comparisons (seven sites with three racial groups each), we find only two cases in which the average difference between the sending TPS and the receiving charter schools is greater than 10 percentage points in the concentration of the transferring student’s race. (p. 18)

A strength of the Zimmer et al. (2009) study is its use of student-level data across several states; however, they do not explicitly address whether student transfers positively or negatively influence racial integration in the TPS sector. This issue poses a slightly different question, which requires individual student transfer data to address. That is, if a charter school enters an area and draws some Black students away from TPS that serve an above-average number of Black students, then this would be favorable for the racial integration of the TPS system. However, if charters enter an area and drew a large set of White students from TPS that were previously very well integrated, this change would have an unfavorable effect on the racial integration within the TPS system.

Bifulco et al. (2009) define “integrative moves” as student transfers where students “move from schools that have a higher percentage of students from their own group to schools with lower percentages of students with their own group” (p. 10). However, this definition may be problematic in that the net effect of these transfers is not assessed against the backdrop of the larger demographic makeup of the metropolitan area. For example, in a region with a student population that is 50% Black, what if Black students were transferring from TPS that were around 40% Black to a charter school that was around 30% Black? According to the Bifulco et al. (2009) definition, these moves are considered integrative. In fact, these moves would be both positive and negative with respect to integration, as it would move the demographic composition of the charter schools closer to the demographics of the broader community while moving the demographics of the TPS further away.

In the analyses that follow, we attempt to avoid the problematic methodological approaches that have cast doubt upon existing studies. First, we assess the current state of racial segregation at charter schools using appropriate measures of segregation and integration along with an appropriate comparison group of TPS. We also follow the lead of studies such as those conducted by Bifulco et al. (2009) by employing student-level data to assess the effect of charter transfers on the racial mix of sending schools. Finally, we hope to improve upon the Bifulco et al. strategy by assessing the integrative impact of such transfers against a common benchmark that represents the racial mix of the broader community.

Research Questions and Method

To examine the extent to which public charter schools affect the racial balance of schools in Little Rock, we ask two broad questions—one static and the other dynamic.

In its current state, which school sector, charters or TPS, is characterized by less racial segregation and better racial integration?

What is the impact of student transfers from TPS to charter schools on the racial composition of all schools in the metropolitan area?

Question 1: Which school sector is currently characterized by less racial segregation and better racial integration?

In this section, we ask which school sector, charter or TPS, is currently characterized by less racial segregation and better racial integration. We consider both racial segregation and racial integration by addressing these sub-questions:

What percentage of students in each school sector attend school in a hyper-segregated (90% or more White or 90% or more minority) environment?

Which school sector boasts a student population that is better integrated? That is, which school type is more likely to have student populations that are representative of the larger metropolitan area?

In our initial question on the current status of segregation/integration in the two sectors, we first identify schools in each sector that were hyper-segregated. For these purposes, we label a school as being hyper-segregated if either White students or minority students represented 90% or more of the entire student population. Then, we simply compute what percentage of students in each sector attended racially hyper-segregated schools. In this way, we can determine the extent to which students were exposed to any type of racial diversity. Finally, we simply compare the percentage of charter students who attended school in racially hyper-segregated environments with that of students in TPS.

This initial question of hyper-segregation is relatively simple in that we do not need to take into consideration the racial composition of the surrounding area; that is, for this set of analyses, we are willing to assume that a school with 90% or more White students is segregated, regardless of the community in which the school is situated. However, the question of racial integration is context-specific. As we have argued in the previous section, the appropriate context for the question of racial integration is the larger metropolitan area. Because charter schools are not restricted by city or school boundaries, the charter schools in the Little Rock area draw the majority of their students from three different urban school districts. As such, identifying a school as integrated if it reflects the population of the city of Little Rock alone does not seem reasonable. Moreover, because municipal and school district boundaries are often heavily segregated (racially and economically), it does not make sense to classify a school as racially integrated on the basis that its racial composition is reflective of the racially segregated city in which the school resides.

Question 2: How have charter school transfers influenced racial segregation?

After asking which sector is “better” with respect to current racial composition, we investigate the impact that transfers from TPS to charter schools had on the racial composition of all schools in the metropolitan area. First, by tracking individual student transfers, we examine the differences in racial balance of the TPS that students left (imitating the Zimmer et al., 2009, analysis), and ask whether students transferred into schools that were substantially different. Then, we assess whether the student transfers had a beneficial or detrimental effect on the racial integration of TPS that students left. We are able to draw inferences of this nature by examining whether student transfers moved the racial composition of each TPS closer to or further away from the metropolitan area average.

Data

Data were obtained from the Arkansas Department of Education and span the 6-year period from the 2004-2005 school year, the first year that the Little Rock metropolitan area experienced substantial activity in the charter sector, to the 2009-2010 school year. For students in grades K-12 in both public charter schools and TPS in the Little Rock metropolitan area, we obtained individual-level data that included identifiers for school and district, grade, gender, race, and free and reduced lunch (FRL) eligibility. Our database also includes a unique identification number for each student, which allows us to match student records over the 6-year period. As a result, we can identify when students switched schools during that time period and determine how these student transfers affected the racial balance of the TPS that students exited.

Sample

In Table 1, we present racial and FRL data from 2004-2005 to 2009-2010 for three different school sectors: charter schools, Little Rock TPS, and the TPS in the larger metropolitan area. Overall, charter schools tend to be more White than TPS in the LRSD and the metropolitan area. There are also fewer FRL-eligible students in charter schools than in the surrounding TPS. Since 2005-2006, however, charter schools have changed in composition to become comprised of a lower percentage of White and FRL-eligible students and a greater percentage of minority students. Racial trends for TPS in both Little Rock and the broader metropolitan area have remained relatively stable during that same time frame, with both sectors experiencing increases in the percentage of FRL-eligible students. Total charter enrollment over the 6 years from 2004-2005 to 2009-2010 increased from 474 to 3,179 students, with the total number of charter schools increasing from 3 to 15 during that same time period. TPS enrollment in Little Rock and in the wider metropolitan area remained relatively stable, though the number of TPS in operation decreased slightly.

Table 1.

Demographics of Charter Schools, LRSD, and Pulaski County, 2004-2005 to 2009-2010.

	2004-2005	2005-2006	2006-2007	2007-2008	2008-2009	2009-2010
All metro area charter schools
n of students	474	500	705	1,058	2,458	3,179
% White	63.5	65.2	62.2	46.8	41.4	43.3
% minority	34.4	28.2	28.1	44.0	51.9	50.4
% FRL	41.3	16.2	10.8	30.0	39.2	38.0
Little Rock TPS
n of students	24,424	25,095	25,500	25,738	24,660	24,380
% White	24.4	24.4	23.6	22.4	21.7	22.0
% minority	73.8	74.1	74.7	75.9	76.5	76.2
% FRL	56.8	59.0	57.8	61.6	62.6	68.1
Metro area TPS
n of students	51,495	52,406	52,590	52,107	51,040	50,625
% White	37.5	36.6	35.4	34.2	33.7	33.5
% minority	61.2	62.1	63.2	64.3	64.7	64.8
% FRL	53.7	56.1	55.2	56.0	60.9	63.3

Note. Asians students are excluded from these percentages and as a result, categories may not sum to 100. LRSD = Little Rock School District; %FRL = percentage of students eligible for free or reduced price lunch; TPS = traditional public schools.

While the percentage of students attending charter schools in the Little Rock metropolitan area has increased since 2004-2005, the total number of charter students still represents a small fraction of the public school student population—just more than 7% in Little Rock and just fewer than 6% in the metropolitan area as a whole. Moreover, an even smaller fraction of students transferred out of TPS to charter schools each year. For example, in the year when the greatest number of charter seats opened in the area (2008-2009), only 2.4% of the students from the Little Rock Public School district transferred to a charter school.

Of course, these simple comparisons of aggregate racial composition tell us nothing of the level of integration or segregation within individual schools. For example, if we knew the enrollment from 10 charter schools in the area was 50% White and 50% Black, we would still be unable to draw any conclusions about racial integration within charter schools. In this scenario, there could be five segregated schools that were 100% White alongside of five other segregated schools that were 100% Black, or all 10 schools could be perfectly integrated at 50% and 50% Black. Thus, aggregate figures tell us nothing about the question of integration. In the following sections, we present the results of our analysis on the extent of racial segregation and relative racial integration in both school sectors.

Results

Research Question 1: Which school sector is currently characterized by less racial segregation and better racial integration?

First, we present the results of our most straightforward question, in which we examine what percentage of students in each sector attended school in a hyper-segregated environment. As mentioned earlier, there is no agreed upon definition of hyper-segregation. We borrow from the definition of Frankenberg et al. (2010), in which the authors defined hyper-segregated as any school with a composition of 90% or more of its students from underrepresented racial backgrounds. We also include schools with 90% or more White students in the hyper-segregated category to fully determine the extent to which students have little to no exposure to any type of racial diversity. In Table 2, we present the results of these comparisons between charter schools and TPS.

Table 2.

Little Rock Metropolitan Area: Percentage of Charter and Traditional Public School Students in Hyper-Segregated Schools, by Racial Group, 2009-2010.

	All students in hyper-segregated schools (%)	White students in hyper-segregated White schools (%)	Minority students in hyper-segregated minority schools (%)
Metro area charter schools	17.3	0.0	32.2
Metro area traditional public schools	22.1	1.6	31.2
Little Rock charter schools	27.6	0.0	43.2
Little Rock traditional public schools	39.9	0.0	49.5

Note. Charters in the Metro area are located geographically within the boundaries of the LRSD, the North LRSD, and the Pulaski County Special School District. We do not include Asian students as minority students because the question of the segregation of minority students generally refers to traditionally disadvantaged racial groups. In the Little Rock region, Asian students are less likely than other minority groups to be economically disadvantaged. We conducted these same analyses with Asian students included in the minority category, and the results were nearly identical. LRSD = Little Rock School District.

Overall, the percentage of charter school students in hyper-segregated schools is lower than the percentage of TPS students in hyper-segregated schools, both within the larger metropolitan area (17.3% compared with 22.1%) and in Little Rock (27.6% compared with 39.9%). We find that the percentage of minority students enrolled in hyper-segregated minority schools is slightly higher in the traditional sector than in the charter sector. In the city of Little Rock, 49.5% of minority students in TPS attended hyper-segregated minority schools compared with 43.2% of minority students in the Little Rock charter sector. In the entire metropolitan area, the difference is less pronounced. In the TPS, 31.2% of minority students attended hyper-segregated minority schools, whereas 32.2% of minority charter students attended similarly segregated schools.

As noted above, this question of hyper-segregation is not relative to the context or the racial composition of the surrounding community. A reasonable consideration of racial integration requires a measure of comparison. To determine which school sector boasts a student population that is better integrated, we compare the racial composition of the individual schools with that of the larger metropolitan area. More specifically, we compute the average absolute “distance” between the percentage of minority students for each school and the total percentage of minority students in all public schools (traditional and charter) in the metropolitan area (64.8% in 2009-2010).¹

We find that charter students in the metropolitan area, on average, attended schools that were 25 percentage points different than the overall composition of minority students in the broader metropolitan area, whereas students in TPS across the area attended schools that were 19 percentage points different (see Table 3). Thus, in the wider metropolitan region, TPS were more representative of the region as a whole than were the region’s charter schools. In the city of Little Rock, however, the charter sector appears to be similarly integrated by this measure. Charter students in Little Rock, on average, attended schools with percentages of minority students that were 21 percentage points away from the regional average, while students in TPS across the area attended schools that were roughly 22 percentage points away from the regional average. As far as the directionality of these differences, charter schools were more likely to serve below-average numbers of minority students, while TPS served above-average numbers of minority students.

Table 3.

Average Distance From the Percent Minority of the Metropolitan Area Average, 2009-2010.

	Metro area charter sector (%)	Metro area TPS sector (%)	Little Rock charter sector (%)	Little Rock TPS sector (%)
Absolute distance from Metropolitan area minority average	±25.2	±19.0	±20.7	±21.8
Average distance for students above the Metropolitan area average	+31.9	+23.6	+31.9	+25.5
Average distance for students below the Metropolitan area average	−23.8	−15.8	−16.4	−14.9

Note. In this analysis, we again exclude Asian students from the racial minority category. We also ran the analysis with Asian students included in the minority category and the overall results did not change. TPS = traditional public schools.

The benefit of this particular “distance” measure is that it does not rely on an arbitrary or artificial benchmark, such as the 90% measure of hyper-segregation. When we measure integration as a distance from a regional average, we avoid this problem. However, this measure is perhaps not as easy to interpret. What does it mean, for example, if a school’s percentage of minority students is 20 percentage points different from the average? As a way to aid in the interpretation of our integration measure, we classify all of the schools in the region as either “integrated” or not. For these purposes, we define a school as being “integrated” if the percentage of minority students in the school falls within a certain range relative to the regional average. Establishing such cutoffs, however, is somewhat arbitrary in nature. Therefore, we establish two different thresholds: within ±10% of the metropolitan area average and within ±15% of the metropolitan area average. Recall that the percentage of minority students in the Little Rock metropolitan area at the start of the 2009-2010 school year was 64.8%.

The results of this analysis are presented in Table 4 and show that students in both sectors are not very likely to attend school in an integrated environment. Using the ±10% definition, we find only 1 charter school and 21 TPS that we would classify as integrated in the metropolitan area, 7 of which are in the city of Little Rock. If we expand the “integration window” to classify schools within 15 percentage points of the regional fraction of minority students, we find that nearly half of Little Rock’s charter students (47%) and nearly one third of the city’s students in TPS (29%) attend integrated schools. In the larger metropolitan area, however, the 15% “integration window” shows nearly identical conditions.

Table 4.

Percentage of Charter, Little Rock, and Metro Area Students in Racially Integrated Schools, 2009-2010.

	Metro area charter students	Metro area students in TPS	Little Rock charter students	Little Rock students in TPS
Percentage of students within ±10 % schools	13.6	33.0	21.8	26.8
Number of schools 54.8%-74.8% minority (±10 %)	1	21	1	7
Percentage of students in ±15% schools	39.2	39.3	47.4	28.6
Number of schools 49.8%-79.8% minority (±15 %)	4	28	3	8

Note. TPS = traditional public schools.

Here again, we find no clear advantage for either sector in terms of racial integration. While the ±10% integration definition suggests that Little Rock’s TPS are better integrated, the ±15% definition favors Little Rock’s charter schools. In the larger metropolitan area, TPS appear to be better integrated when using the ±10% integration definition, while the results using the ±15% definition are roughly the same.

In the city of Little Rock, we find that students in charter schools are less likely to attend hyper-segregated minority schools. In the overall region, there is essentially no difference between the traditional and charter sectors. When we consider measures of integration, we find that both charter schools and TPS serve student populations with minority percentages that differ from the regional average by approximately 20 percentage points. Finally, when using various definitions for racially integrated schools, we find no clear winner with respect to providing a racially integrated school environment. Given that nearly half of the minority students in Little Rock attend hyper-segregated minority schools in both the charter (43.2%) and traditional sectors (49.5%), each sector should be concerned about racial segregation.

Research Question 2: What impact do transfers from TPS to charter schools have on the racial balances of both the TPS they have left and the charter schools that they enter?

In the section above, we have attempted to address the question which sector is “better” with respect to current racial composition. We now take the next step by investigating a more dynamic issue: The impact that transfers from TPS to charter schools have had on the racial composition of schools in the metropolitan area. Here, we look at the specific impact charter schools in the Little Rock area had on the TPS that students left. As noted earlier, some have expressed concern that charters in Little Rock may be negatively affecting the racial and economic balance of Little Rock TPS. In the Arkansas context, the concerns about the harmful effects on TPS are made specifically about Little Rock TPS and not about the other two districts in the metropolitan area. Thus, we focus in this section only on those students who transfer from the LRSD to area charters.

We first identify all students in our dataset who transferred to charter schools from Little Rock TPS from the 2005-2006 to the 2009-2010 academic years. Using these individual-level student transfer data, we are able to provide detailed descriptions of the schools that students left and the schools they entered. Before delving into this question by categorizing the transfers as harmful or helpful with respect to racial integration, we first present descriptive statistics on the racial compositions of the charter schools as well as the TPS of origin (see Table 5). Here, we illustrate the racial composition of current and previous schools separately for White and minority students to see how a student’s new school environment compares with his or her previous school.

Table 5.

Charter and LRSD Peer Environments for Charter Movers, by Racial and Ethnic Background of Student.

	2008-2009		2009-2010
	White students (%)	Minority students (%)	White students (%)	Minority students (%)
LRSD schools that White students left	35.0	62.1	37.4	58.8
Charter schools that received them	40.6	53.1	40.3	48.2
Change	+5.6	−9.0	+3.0	−10.6
LRSD schools that minority students left	20.1	78.0	18.2	79.9
Charter schools that received them	28.5	66.4	21.0	72.5
Change	+8.4	−11.6	+2.8	−7.4

Note. In this table, we are only looking at student transfers. Thus, the charter school demographics are for the charter school to which students transferred, and the LRSD school demographics are from the school in which the student was enrolled during the previous year (before transferring to the charter school). LRSD = Little Rock School District.

Overall, in the current and previous school years, minority students transferring from a Little Rock TPS to a charter school entered into school environments that had a lower percentage of minority students (and consequently, more White students) than their previous school. In other words, minority students transferred into charter schools with a more equal balance of White and minority peers than in their previous schools.

For White students transferring to charter schools, we observe the opposite occurring: White students tended to transfer into charter schools with a lower percentage of minority students and a greater percentage of White students. However, this is not necessarily indicative of a move into a racially isolated all-White environment. In fact, the distribution of White and minority students in the charter schools into which White students move was actually more even than that of the Little Rock TPS they previously attended. For example, in 2008-2009 White students left Little Rock TPS that were, on average, 35.0% White and 62.1% minority, and entered into charter schools that were 40.6% White and 53.1% minority.

This descriptive analysis suggests that the effect of charter transfers on students, with respect to exposure to racial diversity, was minimal. Indeed, our findings here are consistent with the findings reported by Zimmer et al. (2009); that is, “Transfers to charter schools do not involve dramatic shifts in the sorting of students by race in any of the sites included in the study” (p. 84).

Finally, we sought to determine the effect that these transfers had on the racial integration of the sending Little Rock TPS out of which these students transferred. To do so, we coded every student transfer as having a positive or negative effect on integration. Of course, the difficulty in this type of analysis is determining what constitutes a positive or negative effect. Our strategy in establishing these definitions was to categorize transfers as having a positive effect if they moved the racial composition of the Little Rock TPS closer to the racial composition of the metropolitan area and a negative effect if the transfers resulted in the TPS becoming less like the demographic profile of the metropolitan area.

In Table 6, we present the types of transfers to charter schools from Little Rock TPS that occurred from the 2006-2007 to the 2009-2010 school year. In the 2009-2010 columns, we describe the transfers during the most recent school year. However, because it is possible that the results of this particular year were anomalous in some way, we also present information on all such student transfers over the past 4 academic years. In fact, the results of the 2009-2010 transfers look very much like the results of the transfers over the past 4 years.

Table 6.

Impact on the LRSD TPS of Student Transfers to Charter Schools.

	2009-2010	2006-2007 to 2009-2010
Type of transfers	All transfers (%)	All transfers (%)
White student transfers
Good result (left above-average White schools)	20.1	25.0
Bad result (left below-average White schools)	9.6	12.2
Minority student transfers
Good result (left above-average minority schools)	53.6	47.8
Bad result (left below-average minority schools)	16.7	15.0
Total transfers (n)	239	1,100

Note. LRSD = Little Rock School District; TPS = traditional public schools.

There were 44 Little Rock TPS in our dataset, and we labeled each of these as above-average White or above-average minority. Across the region in 2009-2010, 64.8% of students in the metropolitan area were minority (excluding Asian students) and 33.5% of the students were White. Of the 44 Little Rock TPS in 2009-2010, 33 had above-average minority student enrollments and the remaining 11 schools had above-average White student enrollments. With these figures as a background, we can begin to assess the extent to which these charter transfers were favorable or not with respect to racial integration.

Across all years, the majority of charter transfers for White students involved White students leaving schools with an above-average percentage of White students. As a result, these student transfers resulted in the TPS looking more like the metropolitan region as a whole.² Furthermore, across all years, examples of White students transferring out of schools that were predominately non-White (i.e., “white flight”) comprised of a very small percentage (12.2%) of the total transfers to charter schools.

The percentage of minority students leaving above-average minority TPS (transfers that have a positive effect on the integration of the district) also exceeded the percentage of such transfers from below-average minority schools (transfers that have a negative effect on the integration of Little Rock TPS) for each of the past 4 school years. Of the transfers in which minority students exited Little Rock TPS and entered charter schools, 47.8% of these were favorable in terms of racial integration. In these cases, minority students left Little Rock TPS with disproportionate numbers of minority students, thus leaving these TPS more reflective of the racial composition of the entire region. Overall, of all the transfers from Little Rock TPS to charters, 72.8% (25.0 + 47.8) would be classified as having a positive effect on the racial integration of Little Rock’s TPS sector.

Conclusion

In this article, we ask two questions related to charter schools using the city of Little Rock, Arkansas, as a backdrop. First, we ask a static question about the current levels of racial segregation in charter schools as compared with Little Rock TPS. This question is important to many who are concerned that the increasing prevalence of charter schools in the United States might lead to increased segregation of various sorts in our schools, as critics of school choice fear that families will choose to further segregate themselves if given the opportunity. Proponents of school choice, however, contend that by detaching the choice of school from the choice of neighborhoods (most of which are segregated), charter schools may actually decrease levels of segregation. A review of the literature in this area reveals numerous claims and some empirical evidence; however, as we argue above, most of these analyses are built on flawed methods and, unfortunately, we still know very little about whether choosing charter schools is likely to result in increased segregation.

With the benefit of student-level data, we focus our analysis on the Little Rock metropolitan area and find that charter schools in the region are less likely to be hyper-segregated than TPS, but TPS have racial compositions that more closely reflect the regional averages. In each of these cases, however, the differences are slight. Thus, in Little Rock at least, the concerns of charter critics—that charter schools are far more likely to be racially segregated—are not supported by the data.

Nevertheless, choice critics continue to make the claim that charter schools are more segregated (e.g., Orfield, 2007) and further claim that charters lead to greater segregation in TPS. The first piece of data germane to this question is simple but nonetheless important—very few students actually leave Little Rock TPS each year for charter schools. For example, in 2004-2005, 0.4% of the students in Little Rock TPS transferred to charter schools; this figure grew to only 1.2% of the Little Rock TPS student population in the 2009-2010 school year. Furthermore, and even more importantly, the students who transferred from Little Rock TPS to charter schools were more likely to be minority students than White students. It is difficult to imagine that this small number of diverse students leaving Little Rock TPS is having the negative impact on the desegregation efforts of the entire district.

Our second research question examines the impact of student transfers from TPS to charter schools on the racial composition of schools in the metropolitan area. When we look only at students who left Little Rock TPS for charters, we find that the majority of these transfers actually improve the levels of racial integration at the TPS from which they transferred. This finding is attributed to the fact that the majority of transfers involve minority students leaving predominately above-average minority schools or White students leaving above-average White schools. In all of these cases, the student transfers help the exiting school because the Little Rock TPS is left less segregated.

We also do not find a disproportionate number of student transfers that would be of particular concern to critics of school choice, such as only White students exiting from high-minority schools (“white flight”). If we found that only these types of transfers occurred, there would certainly be cause for concern. However, these types of transfers were actually quite infrequent when compared with the majority of beneficial transfers that have occurred since 2006-2007. Thus, we can find little evidence that the charter schools had a negative impact on the racial balance of Little Rock TPS.

In general, it seems that proponents of increased racial integration are focusing on the wrong target when attacking charter schools. Sadly, most students living in inner cities attend intensely segregated minority schools, whether they attend charter schools or TPS. Yet, across the United States, only 2.5% of public school children roam the halls in charter schools each day; the remaining 97.5% attend TPS (U.S. Department of Education, National Center for Education Statistics, 2011). Those who claim to be truly concerned about limiting segregation should be focusing on the segregation in TPS to address this problem.

Finally, and perhaps more important, the fact that poor and minority students exit segregated TPS for, in some cases, similarly segregated charters, does not imply that charter school policy is imposing segregation upon these students. Rather, the racial patterns we observe in charter schools are the result of the active choices these students and families make to seek more attractive schooling options. Clearly, the student attendance patterns that emerge from these increased choices offered to minority families are quite different than attendance patterns that resulted from the forced segregation of our nation’s past. Indeed, it is likely that the parents who are now able to choose charter schools for their students view these options as ones that enhance, rather than undermine, their civil rights.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Gary W. Ritter is professor and holder of the Endowed Chair in Education Policy at the University of Arkansas. His research interests include teacher compensation, racial segregation in schools, and evaluating programs in economically disadvantaged schools. His research has been published in journals such as Phi Delta Kappan, Review of Educational Research, Education Finance & Policy, Educational Evaluation and Policy Analysis, and Education Next.

Nathan C. Jensen, is a research scientist at Northwest Evaluation Association (NWEA). Before joining NWEA, he was a senior research associate in the Office for Education Policy in the Department of Education Reform at the University of Arkansas. He holds a PhD in counselor education from the University of Arkansas.

Brian Kisida is a senior research associate at the University of Arkansas. His research interests include school choice, school integration, arts education and cultural institutions, and curriculum standards. His work has been cited in congressional testimony before the U.S. House and Senate and has been featured in numerous media outlets.

Daniel H. Bowen is a postdoctoral fellow at Rice University with the Houston Education Research Consortium. He holds a PhD in education policy from the University of Arkansas. His research interests include teacher quality policies, school choice, and nontraditional measures of educational outcomes.

References

Berliner

Farrell

Huerta

Mickelson

(2000). Will vouchers work for low-income students? Center for education research, analysis, and innovations (CERAI-00-31). Retrieved from http://nepc.colorado.edu/files/cerai-00-37.htm

Bifulco

Ladd

H. F.

Ross

(2009). Public school choice and integration evidence from Durham, North Carolina. Social Science Research, 38(1), 71-85. /

Cobb

C. D.

Glass

G. V.

(1999). Ethnic segregation in Arizona charter schools. Education Policy Analysis Archives, 7(1). Retrieved from http://epaa.asu.edu/ojs/article/download/536/659

DeMillo

(2012). Little Rock district charter school suit in court. Houston Chronicle. Retrieved from http://www.chron.com/news/article/Little-Rock-district-charter-school-suit-in-court-3443286.php

Echenique

Fryer

R. G.

(2005). On the measurement of segregation (NBER Working Paper No. 11258). Cambridge, MA: National Bureau of Economic Research.

Echenique

Fryer

R. G.

(2007). A measure of segregation based on social interactions. The Quarterly Journal of Economics, 122, 441-485.

Egalite

A. J.

Mills

J. N.

(2014). The Louisiana scholarship program: Contrary to justice department claims, student transfers improve racial integration. Education Next, 14, 66-69.

Frankenberg

Lee

(2003). Charter schools and race: A lost opportunity for integrated education. Education Policy Analysis Archives, 11(32). Retrieved from http://epaa.asu.edu/epaa/v11n32/

Frankenberg

Siegel-Hawley

Wang

(2010). Choice without equity: Charter school segregation and the need for civil rights standards. The Civil Rights Project, University of California Los Angeles. Retrieved from http://civilrightsproject.ucla.edu/news/press-releases/new-report-explains-that-charter-schools-political-success-is-a-civil-rights-failure

10.

Garcia

D. R.

(2008). The impact of school choice on racial segregation in charter schools. Educational Policy, 22, 805-829.

11.

Greene

J. P.

(2005). Choosing integration. In Scott

J. T.

(Ed.), School choice and diversity: What the evidence says (pp. 24-41). New York, NY: Teachers College Press.

12.

Greene

J. P.

Mellow

(2000). Integration where it counts: A study of racial integration in public and private school lunchrooms. Texas Education Review, 1(1), 15-26.

13.

Henig

J. R.

MacDonald

J. A.

(2002). Locational decisions of charter schools: Probing the market metaphor. Social Science Quarterly, 83, 962-980.

14.

Miron

Urschel

Mathis

W. J.

Tornquist

(2010). Schools without diversity: Education management organizations, charter schools, and the demographic stratification of the American school system. Boulder, CO: Education and the Public Interest Center. Retrieved from http://nepc.colorado.edu/publication/schools-without-diversity

15.

Orfield

(2007). Myron Orfield: Segregation is back, and our future is on the line. The Star Tribune. Retrieved from http://www.startribune.com/opinion/commentaries/12230281.html

16.

Rapp

K. E.

Eckes

S. E.

(2007). Dispelling the myth of “white flight”: An examination of minority enrollment in charter schools. Educational Policy, 21, 615-661.

17.

Ritter

Jensen

Kisida

McGee

J. B.

(2010). A closer look at charter schools and segregation. Education Next, 10(3), 69-73.

18.

U.S. Department of Education, National Center for Education Statistics. (2011). Digest of education statistics: Number and enrollment of public elementary and secondary schools, by school type, level, and charter and magnet status. Retrieved from http://nces.ed.gov/programs/digest/d12/tables/dt12_108.asp

19.

Zimmer

Gill

Booker

Lavertu

Sass

T. R.

Witte

(2009). Charter schools in eight states: Effects on achievement, attainment, integration, and competition. Santa Monica, CA: RAND.