Jump to ContentJump to Main Navigation
Integrating Schools in a Changing SocietyNew Policies and Legal Options for a Multiracial Generation$

Erica Frankenberg and Elizabeth DeBray

Print publication date: 2011

Print ISBN-13: 9780807835128

Published to North Carolina Scholarship Online: July 2014

DOI: 10.5149/9780807869208_frankenberg

Show Summary Details
Page of

PRINTED FROM UNIVERSITY OF NORTH CAROLINA PRESS SCHOLARSHIP ONLINE (www.northcarolina.universitypressscholarship.com). (c) Copyright University of North Carolina Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in NCSO for personal use. Subscriber: null; date: 23 September 2021

The Effects of Socioeconomic School Integration Policies on Racial School Desegregation

The Effects of Socioeconomic School Integration Policies on Racial School Desegregation

Chapter:
(p.187) The Effects of Socioeconomic School Integration Policies on Racial School Desegregation
Source:
Integrating Schools in a Changing Society
Author(s):

Sean F. Reardon

Lori Rhodes

Publisher:
University of North Carolina Press
DOI:10.5149/9780807869208_frankenberg.14

Abstract and Keywords

This chapter the effects socioeconomic school integration policies on racial school desegregation. It begins by presenting a brief historical overview of socioeconomic status-based student assignment (SBSA) plans. It then discusses it prevalence of SBSA plans in school districts. Before discussing the effects of SBSA plans on segregation patterns, it considers how different features of SBSA plans may affect patterns of socioeconomic segregation and racial segregation. The chapter ends by examining the estimated effect of SBSA plans on segregation levels for school districts.

Keywords:   socioeconomic school integration, racial school desegregation, socioeconomic status-based student assignment, SBSA plans, segregation patterns, socioeconomic segregation, racial segregation, school districts

Given the Supreme Court's decision in Parents Involved in Community Schools v. Seattle School District No. 1 (PICS, 2007), the use of individual student race in voluntarily adopted school assignment plans (as opposed to court-ordered plans) is no longer legally permissible in most cases.1 However, because socioeconomic status does not create a protected class under the 14th Amendment, the use of individual socioeconomic status in school assignment plans is legally permissible. Richard Kahlenberg, among others, has argued that socioeconomic integration will produce racial desegregation as a by-product, given the strong correlation between race and socioeconomic status in the United States.2 Sean Reardon, John Yun, and Michal Kurlaender demonstrate, however, that socioeconomic integration need not necessarily lead to increased racial desegregation.3 Their analysis suggests that socioeconomic integration is most likely to lead to reduced racial segregation in districts where racial residential segregation is relatively low and where socioeconomic integration is based on a measure such as family income or parental education, rather than only on student free or reduced-price lunch eligibility.

As yet, however, there is no systematic empirical study of the claims made by Kahlenberg or of the analysis of Reardon, Yun, and Kurlaender regarding the effects of socioeconomic status (SES)-based student assignment plans on racial desegregation.4 In fact, there is not even any systematic empirical evidence regarding the effects of SES-based student assignment plans on socio-economic desegregation. Our goal in this essay is to provide evidence on these two issues.

A Brief History of Socioeconomic Status-Based Student Assignment Plans

As the name implies, socioeconomic status-based student assignment (SBSA) plans use socioeconomic characteristics of students' families or neighborhoods (p.188) as factors in determining which schools students attend. Typically, this means that districts classify students by one or more socioeconomic characteristics (free or reduced-price lunch eligibility, parental education, neighborhood poverty rate) and seek to assign students to schools in ways that roughly balance these characteristics among schools so that poor students are not concentrated in a subset of the district's schools. The defining characteristic of an SBSA plan, then, is the aim of constructing school enrollments so that each school is relatively socioeconomically representative of the district enrollment as a whole, at least with regard to the set of socioeconomic factors used in the plan.

Socioeconomic status-based school assignment plans are relatively new. A few districts began using SES-based student assignment plans more than a decade ago (Montgomery County, Maryland, in 1986; Topeka, Kansas, in 1991; La Crosse, Wisconsin, in 1992). Most of the districts using some form of SBSA plan, however, began doing so in the last decade. Thirty-two districts implemented some form of SBSA plan between fall 1998 and fall 2008; three more have plans to begin SBSA in the next two years (we give details on our data collection procedures below). In the fall of 1997, roughly 200,000 students were enrolled in districts with SBSA plans; in the fall of 2008, there were 1.6 million students in such districts, an eightfold increase in a decade.

In part, the impetus for such plans likely comes from the courts' relatively recent retreat from racial desegregation policies, which in some cases has resulted in a rapid—and troubling—resegregation of schools. As Reardon and Yun describe, three Supreme Court decisions in the early 1990s, Board of Education of Oklahoma City Public Schools v. Dowell, Freeman v. Pitts, and Missouri v. Jenkins,5 made it easier for districts to be released from court-ordered desegregation:

In Dowell, the court emphasized that desegregation orders were intended to be temporary and that a return to local control was preferable, once a district had “complied in good faith with the desegregation decree since it was entered, and … the vestiges of past discrimination had been eliminated to the extent practicable” (Dowell, 498 U.S. at 249–50). In Freeman, the Court ruled that districts could be released from desegregation orders piecemeal—district courts might end their judicial over-sight in areas where sufficient progress had been shown, for example, in student or faculty assignments, while retaining oversight in other areas where progress was still needed. See Freeman, 503 U.S. at 489–90. More-over, the Court emphasized the need for a district's “good faith commitment” (p.189) to end segregation, see Freeman at 491, rather than the stronger requirement that desegregation efforts “work, and … work now,” an approach the Court had previously emphasized. Green v. County Sch. Bd., 391 U.S. 430, 439 (1968). Most recently, in Missouri v. Jenkins, 515 U.S. 70 (1995), the Court appeared to shift the burden of proof from school districts (who, since Green, had been required to explain racial disparities) to plaintiffs, who, it said, must identify “the incremental effect” that prior de jure segregation had upon any continuing racial disparities if they are to be considered by federal courts. See id. at 101.6

As a result of these decisions, a large number of school districts have been released from court order in the past decade. In some cases, school districts have responded by implementing SES-based student assignment plans, in the hope that such plans will ameliorate some of the potential resegregation of schools.

The courts retreat from race-based desegregation strategies is not the only factor leading to the adoption of SBSA plans, however. The impetus for SBSA plans likely also stems from a growing societal deemphasis on race and race-based remedies for racial and social inequality, as evident in the PICS decision. Although most of the plans in effect today predate the Supreme Court's decision in PICS, they may stem from the same societal move away from race- based remedies. Moreover, research suggests that the concentration of low-income students in schools leads to negative educational outcomes,7 a finding that many districts adopting SBSA plans cite as part of their rationale for such assignment plans.

Prevalence of Socioeconomic Status-Based Student Assignment Plans

There is no comprehensive national data source identifying all public school districts that use some form of SBSA plan. As a result, we conducted a comprehensive search to identify districts using SBSA plans.8 We identified 40 school districts using, or planning to use, socioeconomic status factors in determining student assignments (details below).9 These districts are widely distributed across the United States; few states have more than one or two such districts. Although the 40 districts using SBSA plans comprise only roughly one-quarter of one percent of all districts in the United States, most of them are relatively large districts (averaging roughly 42,000 students enrolled in 2006–7); collectively (p.190)

Figure 10.1. Number of Students in All Districts with SES Integration Plan by Year

The Effects of Socioeconomic School Integration Policies on Racial School Desegregation

Note: Data include district years with imputed enrollments.

they enroll more than 1.6 million students (roughly 3.5 percent of U.S. public school enrollment) (see figure 10.1).

Not only are the districts that have implemented SBSA plans larger than the typical U.S. school district, they are also much more racially and ethnically diverse than the population of U.S. students as a whole: in 2005–6, 42 percent of students in these districts were non-Hispanic white; 28 percent were non- Hispanic black; 24 percent were Hispanic; and 6 percent were Asian.10 It appears, therefore, that large, racially and ethnically diverse districts are much more likely to have adopted SBSA plans in the past decade than smaller and more racially homogenous districts. Almost half a million black and a quarter of a million Hispanic students currently attend school in districts with SBSA plans (close to 6 percent of all black and 3 percent of all Hispanic students in the United States).

The Effects of Socioeconomic Status-Based Student Assignment Plans

Before turning to our analysis of the effects of SBSA plans on segregation patterns, it is useful to consider how different features of SBSA plans may affect patterns of socioeconomic segregation, racial segregation, or both. There are (p.191) three broad mechanisms used by SBSA plans to reduce socioeconomic segregation: (1) explicit socioeconomic balancing of school composition (SES-balancing plans), (2) the use of attendance zones to produce socioeconomic diversity (attendance-zone plans), and (3) SES-based preferences in school choice or transfer practices (transfer priority plans), SBSA plans that rely on the first of these mechanisms (SES-balancing plans) are potentially the strongest type of plan, because they explicitly attend to how evenly students are distributed among schools. Even among SBSA plans that rely on SES balancing, however, the extent to which such balancing affects socioeconomic segregation levels will likely depend on what socioeconomic factors are used, and the level of socioeconomic balance required by a plan. For example, SBSA plans that rely on information about parental education levels and family income (measured with some reasonable detail) are likely to distinguish poor and middle-class students more effectively (and so will likely produce greater socioeconomic integration, all else being equal) than plans that rely solely on students' eligibility for free or reduced-price lunch. Such eligibility is a weak measure of poverty, both because it dichotomizes continuous family income and because it is error-prone. Likewise, SBSA plans that require relatively strict balance among schools in student socioeconomic characteristics (such as requiring that no school have socioeconomic characteristics that differ by more than 5–10 percent from the district average) are likely to produce greater integration than those requiring only a much cruder level of balance.11

SBSA plans that rely on attendance zones also have the potential to have strong effects on socioeconomic segregation levels, but this will depend, again, largely on the particulars and context of the plan. In districts where large numbers of poor students live near large numbers of nonpoor students, it will be relatively easy to draw school attendance-zone boundaries in such a way that each school draws its student body from both high- and low-poverty neighborhoods, resulting in a situation where most students attend schools that are both near their homes and socioeconomically diverse. If poor and nonpoor students live, on average, very far from one another, it will be much harder, or impossible, to delineate compact attendance zones that are socioeconomically diverse. Moreover, drawing attendance zones to include both high- and low- poverty neighborhoods does not guarantee that each school will have equal proportions of poor and nonpoor students. Salvatore Saporito and Deeneesh Sohoni, for example, show that school compositions often do not mirror the racial and socioeconomic compositions of the schoolage population living in their attendance zones, because of differential patterns of private-school enrollment and use of school-transfer options.12

(p.192) In contrast to SES-balancing plans and attendance-zone plans, SBSA plans that rely on SES-based preferences in transfer or school choice decisions are likely to have little substantial impact on segregation patterns. Transfer priority plans typically give low-income students priority in transferring into low-poverty schools (and give higher-income students priority in transferring into high-poverty schools), but such transfers are generally rare. Relatively few students seek to transfer schools in most school districts (often because transportation to a school out of one's neighborhood poses a substantial barrier). Moreover, lower-poverty schools are less likely to be underenrolled than are high-poverty schools, so there are generally few available seats for poor students to apply for in low-poverty schools. Conversely, few nonpoor students apply to transfer to high-poverty schools. As a result, transfer priority SBSA plans are likely to have little effect on overall segregation levels, because the number of students affected by a SES-based transfer priority system is likely to be very small.

One additional feature of student assignment plans may affect their ability to influence socioeconomic segregation levels. Student assignment plans that focus solely on socioeconomic integration may be more effective than those that attend to many factors and use other nonsocioeconomic factors in determining school assignments (such as prior achievement, language, proximity to schools, etc.) because the latter have to balance many competing demands. Attempting to balance schools on a number of factors, for example, makes it less likely that optimal balance will be attained on any one factor.

The effects of SBSA plans on racial segregation also may depend on a variety of factors. Reardon, Yun, and Kurlaender argue that SBSA plans are most likely to produce racial integration as a by-product if the correlation between race and socioeconomic status is strong within a district's population (so that socioeconomic status serves as a good proxy for race) and if residential patterns are characterized by high levels of socioeconomic segregation within racial groups.13 If there is high racial segregation but little socioeconomic segregation within racial groups (i.e., if poor and nonpoor blacks live near one another but far from poor and nonpoor whites), many students could attend neighborhoods schools and the system could have relatively high levels of socioeconomic integration without achieving much racial desegregation. Reardon, Yun, and Kurlaender show that such residential patterns are typical of most large urban school districts.14

The effects of SBSA plans on racial segregation may also depend on whether the SBSA plan replaces a preexisting race-based student assignment (RBSA) (p.193) plan, operates in conjunction with it, or is introduced in a district with no prior plan. In the case where SBSA plans are implemented when race-based student assignment plans end (as has happened in some cases when desegregation orders have ended or when districts have moved away from voluntary race-based plans to socioeconomic plans), for example, it is quite possible that the implementation of the SBSA plan may—seemingly paradoxically—lead to an increase in racial segregation among schools. We might expect that SBSA plans produce less racial integration than do RBSAS but more racial integration than no plan at all. As a result, in our analyses, we investigate whether the effects of SBSA plans differ depending on whether there was a prior RBSA plan and whether the SBSA plan supplanted or supplemented the RBSA.

Data and Methods

As noted above, we attempted to identify every school district currently using some form of SES-based student assignment plan. Table 10.1 lists the 40 districts we identified that have some form of SBSA plan. We contacted district officials and interviewed them regarding the characteristics of the SBSA plans. Based on these interviews and other information (district websites, school board meeting minutes, other published articles and reports), we coded SBSA plans along several key dimensions. Most important, we coded districts as using SES balancing, attendance zones, transfer priority strategies, or some combination of these. These are not mutually exclusive: districts can, and often do, employ more than one feature in their SBSA plans. Specifically, we defined these categories as follows.

SES-balancing plan: Districts were coded as using an SES-balancing plan if their SBSA policy explicitly seeks to ensure that each school has socioeconomic student characteristics that fall within some specified range (e.g., districts where each school must have proportions of students eligible for free or reduced-price lunch within some specified range of the district proportion; or districts where all schools must have fewer than 40 percent of students from high-poverty neighborhoods).

Attendance-zone plan: Districts were coded as using an attendance-zone plan if their student assignment plan relies primarily on neighborhood schools but draws school attendance-zone boundaries in order that each school enrolls students from a socioeconomically diverse set of neighborhoods.

Transfer priority plan: Districts were coded as using a transfer priority plan (p.194)

Table 10.1. Student Assignment-Plan Characteristics in School Districts with Socioeconomic-Based Student Assignment Plans

Plan Mechanisms

District

State

SES Balancing

Attendance Zones

Transfer Priority

Strong Plan

Free Lunch Eligibility

Neighborhood Poverty

Other SES Measures

Non-SES Measures

Never Used Race

Prior RSBA Plan

Currently Uses Race

Beaumont

TX

0

0

1

0

1

0

0

1

0

1

0

Berkeley

CA

1

0

0

1

0

1

1

1

0

0

1

Boulder Valley RE 2

CO

0

0

1

0

1

1

0

0

1

0

0

Brandywine

DE

1

1

0

1

1

0

0

0

1

0

0

Burlington

IA

0

1

0

0

1

0

0

0

0

0

1

Burlington

VT

0

0

1

0

1

0

0

1

1

0

0

Cambridge

MA

1

0

0

1

1

0

0

1

0

0

1

Chapel Hill—Carrboro City

NC

0

1

0

1

1

1

0

1

0

1

0

CHRISTINA

DE

0

0

1

0

1

0

0

0

0

1

0

DAVENPORT

IA

0

0

1

0

1

0

1

1

0

1

0

DES MOINES

IA

0

0

1

0

1

0

0

1

0

1

0

DUVAL COUNTY

FL

0

0

1

0

1

0

0

0

0

1

0

EUGENE 4J

OR

0

0

1

0

1

0

0

0

1

0

0

GREENVILLE COUNTY

SC

0

1

0

0

1

0

0

1

0

1

0

GUILFORD COUNTY

NC

0

0

1

0

1

0

0

0

1

0

0

HAMILTON COUNTY

TN

0

0

1

0

1

0

0

0

1

0

0

JEFFERSON COUNTY

KY

1

0

0

1

0

0

1

0

0

0

1

LA CROSSE

WI

0

1

1

0

1

0

0

0

1

0

0

Lafayette Parish

LA

0

0

1

0

1

0

0

1

0

1

0

Manatee County

FL

0

1

1

1

1

0

0

1

0

1

0

McKinney

TX

0

1

0

1

1

0

0

1

1

0

0

Minneapolis

MN

0

0

1

0

1

1

0

0

1

0

0

Montgomery County

MD

0

1

1

0

1

0

0

1

0

0

1

Moorpark

CA

0

0

1

0

1

0

0

1

0

0

1

Napa Valley

CA

0

0

1

0

1

0

0

0

1

0

0

Omaha

NE

0

0

1

0

1

0

0

1

0

1

0

Palm Beach County

FL

0

1

1

0

1

0

0

1

1

0

0

Portland 1J

OR

0

0

1

0

0

0

0

0

1

0

0

Rock Hill Three

SC

1

0

0

1

1

0

0

1

1

0

0

Rosemount-Apple Valley-Eagan

MN

0

0

1

0

1

0

0

0

1

0

0

San Diego

CA

0

0

1

0

1

0

0

1

1

0

0

San Francisco

CA

1

0

0

1

1

0

1

1

0

1

0

San Jose

CA

0

0

1

0

1

0

0

1

0

1

0

Seminole County

FL

0

0

1

0

1

0

0

1

0

1

0

St. Lucie County

FL

1

0

0

1

1

0

0

1

0

1

0

Stamford

CT

0

1

0

1

1

1

0

1

0

1

0

Topeka

KS

0

0

1

0

1

0

0

1

0

0

1

Troup County

GA

0

1

1

1

1

0

0

1

0

1

0

Wake County

NC

1

0

0

1

1

1

0

1

0

1

0

Waterloo

IA

0

0

1

0

1

0

0

0

0

1

0

All (Proportions)

0.20

0.28

0.68

0.33

0.93

0.15

0.10

0.63

0.38

0.45

0.18

(p.195) (p.196) if they use SES factors in determining whether a given student can voluntarily transfer from his or her neighborhood or assigned school to another school in the district (e.g., districts that give low-income students preference in transferring to schools with low proportions of low-income students). Some districts with transfer priority SBSA plans use SES factors in all voluntary transfers; others use SES factors only in voluntary transfers to some specific subset of schools in the district, such as magnet schools.

For descriptive purposes, we characterize SBSA plans that rely solely on SES-based transfer priority as weak SBSA plans. Those that use SES balancing we characterize as strong plans. Those that use SES-based attendance zones but not SES balancing we characterize as either weak or strong SBSA plans depending on how we read the plans' particulars. The distinction between weak and strong plans is intended to capture our hypotheses regarding the potential of the SBSA plan to reduce segregation levels.

In addition to coding the SBSA plans of the 40 districts according to the features of the plans, we also collected information from each district on when the SBSA plan was first implemented (coded as the fall of the school year in which the plan was first in effect). We also gathered information on what SES factors the SBSA plan uses (eligibility for free or reduced-price lunch, neighborhood poverty rate, parental education) and whether student assignments are also based on other, non-SES factors (such as race, sibling priority, prior achievement, language spoken at home, and gender). Finally, we also collected information on whether each district formerly had or currently has a race-based student assignment plan. Table 10.1 lists the characteristics of the SBSA plans in the 40 districts in our sample.

In order to investigate the effect of SES-based assignment plans on segregation levels, we compute segregation levels for each district in each year for which data are available. We use data from the Common Core of Data (CCD), which covers the school years 1987–88 to 2006–7, to compute racial composition, enrollment, and segregation levels. For our longitudinal analyses, we include only districts for which we observe at least one year's data prior to the implementation of the SBSA policy and at least one year following it. This excludes data from 17 districts: one whose plan began in 1986, prior to the start of the CCD availability (Montgomery County, Maryland); one whose plan began in 1991 but for whom the CCD contains no data prior to 1992 (Topeka, Kansas); and 15 whose plans began, or are scheduled to begin, after the 2006–7 school year, the latest available year of CCD data available at the time of writing). Because some schools are missing enrollment counts by race, eligibility for free (p.197) or reduced-price lunch, or both in some years of the CCD (particularly in the early years of our data), we include in our analysis only district-by-year observations in which the CCD includes enrollment counts by race and free or reduced-price lunch for at least 90 percent of the students in the district.15 This excludes 70 district-by-year observations (57 of which are from the school years 1987–88 to 1996–97), leaving us with a final analytic sample of 390 valid observations in 23 school districts.

We measure segregation using the information theory index (H), developed by Henri Theil.16 The information theory index measures the extent to which students of two groups are evenly distributed among schools. The index ranges from a minimum of o, indicating no segregation (each school has identical proportions of each group) to a maximum of 1, indicating complete segregation (no student of one group attends school with any member of another group). Using this index, we compute segregation between white and black students, white and Hispanic students, white and nonwhite students, and between students who are eligible for free or reduced-price lunch and those who are not.

Models

Following some brief description of the characteristics of existing SBSA plans, we conduct several analyses to estimate the effects of such plans on segregation levels. We begin by estimating the effect of the SBSA policy in each district separately.17 We then fit a series of models to test whether the size of the SBSA effect within each district is related to characteristics of the districts and their plans.18 Because we have only 23 districts in them, these models are necessarily parsimonious. In particular, we fit models including the following SBSA characteristics: (1) the mechanisms the SBSA plan uses to create socioeconomic desegregation (SES balancing, attendance zones, or both, with transfer priority-only plans as the omitted category) and (2) the district's use of an RBSA plan (we include an indicator for districts that never used an RBSA plan and an indicator for districts that used an RBSA plan both prior to and after the implementation of the SBSA plan, with districts that substituted SBSA plans for RBSA plans as the omitted category). We fit several versions of this model for each of the four outcomes of interest (segregation levels between poor and nonpoor, white and black, white and Hispanic, and white and non- white students).

(p.198) Results

Before turning to the estimated effects of SBSA plans, it is useful to examine some of the characteristics of these plans (table 10.2). Districts with “strong” SBSA plans are, on average, somewhat larger than those with “weak” plans, but otherwise the racial and socioeconomic characteristics of districts with each type of plan are relatively similar. The characteristics of strong and weak SBSA plans differ substantially, however, as we expect. Strong plans are much more likely to use SES balancing or attendance-zone boundaries than are weak plans (largely because these are the features that determine whether a plan is coded as strong or weak). Moreover, strong plans typically rely on a broader range of socioeconomic factors than weak plans—only 3 of 27 weak plans use neighborhood poverty or any family socioeconomic factor other than eligibility for free or reduced-price lunch, compared to 6 of 13 strong plans.

In 18 of the 40 districts using SBSA plans, the SBSA plan replaced a prior race-based school assignment plan. Fifteen of the districts had no RBSA plan in the years immediately prior to implementing their SBSA plan, SBSA plans in such districts are generally weak plans that rely only on transfer priority provisions to reduce segregation. Finally, in 7 of the 40 districts, SBSA plans were implemented to supplement existing (and continuing) RBSA plans. Most of these districts have had voluntary RBSA plans to which socioeconomic factors were added (e.g., Berkeley, California; Cambridge, Massachusetts; Burlington, Iowa; Montgomery County, Maryland; and Jefferson County, Kentucky). In general, table 10.2 suggests that strong SBSA plans are somewhat more likely than weak plans to have been implemented in districts that previously used an RBSA; weak SBSA plans are more common in districts that have not used an RBSA in recent years.

In order to investigate the effects of SBSA plans on segregation levels, we begin by plotting trends in segregation levels for each of the school districts in our analytic sample, in order to get a sense of whether segregation levels appear to change sharply following the implementation of SBSA plans. Figure 10.2 shows the trends in socioeconomic and racial segregation for each of the nine districts with SBSA plans coded as “strong” and for which we have segregation data before and after the start of the plan. For the most part, the trend lines suggest little effect of SBSA plans on segregation levels, with a few exceptions. For example, in San Francisco, racial segregation sharply reversed its upward trend at the same time as the SBSA plan was implemented. In Cambridge, Massachusetts, socioeconomic segregation levels declined sharply following (p.199)

Table 10.2. Descriptive Characteristics ofsesA Plan Districts, by Plan Strength

Strong

Weak

Total

Number of districts

13

27

40

District Characteristics

Average enrollment

34,164

45,551

41,851

Average percentage white

50

48

48

Average percentage black

28

26

27

Average percentage Hispanic

12

18

17

Average percentage poor

31

35

34

Student Assignment-Plan Features

SES balancing

8 (62%)

0 (0%)

8 (20%)

Attendance zones

6 (46%)

5 (19%)

11 (28%)

Transfer priority

2 (15%)

25 (93%)

27 (68%)

Uses free/reduced-price lunch eligibility

11 (85%)

26 (96%)

37 (93%)

Uses neighborhood poverty

4 (31%)

2 (7%)

6 (15%)

Uses other SES measures

3 (23%)

1 (4%)

4 (10%)

Uses NON-SES measures

11 (85%)

14 (52%)

25 (63%)

Never used RSBA plana

3 (23%)

12 (44%)

15 (38%)

Currently uses RSBA plan

3 (23%)

4 (15%)

7 (18%)

RSBA plan before SBSA plan

7 (54%)

11 (41%)

18 (45%)

(a) “Never used RBSA plan” applies to districts that did not use an RBSA plan in the two years prior to the implementation of the SBSA plan.

the implementation of the SBSA plan, though segregation levels were declining before the plan started as well. Although we do not show segregation trends for “weak plans,” they are similar to those shown in that there is no evident average effect of the SBSA plans on segregation levels.

After estimating the effects in each district separately (as described in note 18), we investigate whether the magnitude and direction of the effects are associated with several characteristics of the SBSA plan itself. Recall that we hypothesize the effects will be larger (a greater reduction in segregation levels) in districts that had no race-based plan prior to the start of the SBSA or that had a race-based plan that remained in place when the SBSA plan was implemented. In districts where an SBSA plan replaced an RBSA plan, we expect that segregation will not change or may even increase. In addition, we expect SBSA plans to have larger effects in districts that use stronger mechanisms for producing socioeconomic integration, such as SES balancing and attendance zones, than in districts that use only transfer priority mechanisms. (p.200)

The Effects of Socioeconomic School Integration Policies on Racial School Desegregation

Figure 10.2. Trends in Racial, Ethnic, and Socioeconomic Segregation: Nine Districts with “Strong” Socioeconomic Status–Based Student Assignment Plans, 1990–2006

(p.201) The results of the models testing these hypotheses are shown in table 10.3. For each of the three segregation outcomes (between poor and nonpoor, white and black, white and Hispanic, and white and nonwhite), we fit three models: one including only the SBSA plan mechanism variables, a second including only the district RBSA status variables, and a third containing both sets of variables.

In the models predicting the magnitude of segregation between poor and nonpoor students, the plan features are not predictive of the effect, but the RBSA status variables are. In districts that substituted an SBSA plan for an RBSA plan, segregation by poverty status increased slightly (by 0.011 in model 2), on average, while in districts that kept an RBSA plan in place, segregation by poverty status decreased by 0.036 (0.011 + -0.047; see model 2), on average. This suggests that SBSA plans may be effective at reducing socioeconomic segregation when they augment race-based plans but not when they supplant them.

A somewhat similar pattern is evident in the models predicting the effect of SBSA plans on levels of segregation of black and white children. First, note that the intercept in model 3 is positive, indicating that, on average, districts that replace race-based student assignment plans with SES-based plans that rely solely on transfer priority provisions (the weakest type of SBSA plans) tend to experience significant increases in segregation of whites and blacks (+0.40 on average, a moderate increase).19 The effects of SBSA plans on segregation of whites and blacks, however, are stronger (meaning they increase segregation less, or even reduce it) when SBSA plans use SES balancing or attendance-zone mechanisms or when SBSA plans are implemented in districts with no prior RBSA plan (rather than supplanting prior RBSA plans). For example, model 3 predicts that an SBSA plan using only transfer priority mechanisms that is implemented in a district with no prior RBSA plan will have no impact on segregation of whites and blacks (0.040 − 0.042 = −0.002). Likewise, model 3 predicts that an SBSA plan that uses either SES balancing or attendance zones and that replaces a prior RBSA plan will result in no significant change in levels of segregation of whites and blacks. This suggests that strong SBSA plans may be as effective as the race-based plans they replace (though whether that is because the RBSA plans they replace are relatively weak or because the SBSA plans are particularly effective is unclear). We caution, however, against over-interpreting these results, as they are based on a very small sample of districts and there are few districts in the sample with any particular set of characteristics. (p.202)

Table 10.3. Estimated Effects of Socioeconomic-Based Student Assignment-Plan Characteristics on Segregation Levels (H)

Free Lunch-Non-Free Lunch

White-Black

White-Hispanic

White-Nonwhite

Model 1

Model 2

Model 3

Model 1

Model 2

Model 3

Model 1

Model 2

Model 3

Model 1

Model 2

Model 3

Plan Features

Attendance zones

0.006

−0.006

−0.005

−0.036+

0.020

0.013

−0.005

−0.023

(0.017)

(0.017)

(0.017)

(0.019)

(0.010)

(0.013)

(0.012)

(0.015)

SES balancing

0.004

−0.006

−0.011

−0.042*

−0.012

−0.016

−0.014

−0.030*

(0.010)

(0.010)

(0.013)

(0.016)

(0.011)

(0.012)

(0.008)

(0.011)

RBSA Status

Never used RBSA plan

−0.016*

−0.019+

−0.011

−0.042*

0.010

0.010

−0.001

−0.022+

(0.007)

(0.009)

(0.012)

(0.015)

(0.010)

(0.010)

(0.008)

(0.011)

Uses RBSA plan

−0.047*

−0.047+

−0.004

−0.008

0.023*

0.019

0.003

0.000

(0.022)

(0.023)

(0.027)

(0.025)

(0.011)

(0.012)

(0.020)

(0.019)

Constant

0.000

0.011*

0.015+

0.007

0.007

0.040*

0.011+

0.006

0.006

0.007

0.003

0.025*

(0.005)

(0.005)

(0.008)

(0.008)

(0.008)

(0.014)

(0.005)

(0.006)

(0.007)

(0.005)

(0.006)

(0.010)

N

23

23

23

23

23

23

23

23

23

23

23

23

(+) p 〈 .10;

(*) p 〈 .05

Note: All models are weighted by the inverse of the sampling variance of the district-specific SBSA effects. For plan features, omitted category is districts that use only transfer priority provisions. For race status variables, omitted category is districts that used an RBSA plan prior to the SBSA plan but have not used an RBSA plan since implementing the SBSA plan (those that substituted an SBSA plan for an RBSA plan).

Relatively little pattern of effects is evident in the models predicting the magnitude of the effect of SBSA plans on segregation of whites and Hispanics, likely because the correlation between income and Hispanicity is less strong than the correlation between race and income, and because most RBSA plans do not target Hispanic students. Finally, the last set of models, predicting the effect of SBSA plans on white/nonwhite segregation show patterns similar to the white/black models, albeit weaker, as we would expect given the lack of pattern in the white/Hispanic models.

Socioeconomic status-based student assignment plans hold the promise of producing two beneficial outcomes: they may be effective in reducing racial school segregation in the absence of race-based assignment plans; and they may produce socioeconomic integration, a valuable end in its own right. Our analyses here, however, suggest that the SES-based plans that have been implemented to date are few and weak. We are able to identify only 40 school districts using some form of explicit SBSA plan; two-thirds of these use plans (p.203) that employ weak mechanisms and so are likely to have little or no impact on racial or socioeconomic segregation patterns. Of the 13 districts we identified that use mechanisms that appear, on their face, strong enough to have some effect on segregation levels, only 9 have been in place long enough to evaluate. Our graphical and district-specific analyses of segregation trends in these districts suggest little impact on either racial or socioeconomic segregation levels.

Our analyses suggest that one reason that SES-based student assignment plans have relatively little apparent effect on segregation levels is that they supplant prior race-based plans. Indeed, six of the nine districts we evaluate that have relatively strong SBSA plans implemented their SBSA plan to replace a prior race-based plan. On average, in these districts the change from RBSA to SBSA plans has yielded little change in either racial or socioeconomic segregation. The other reason that SES-based student assignment plans appear to have little effect on segregation levels is that most such plans use only very weak mechanisms, typically relying on SES-based priority (in conjunction (p.204) with non-SES factors) in school assignments for a small number of students requesting school transfers. Such plans have no effect on segregation levels when implemented in districts that had no prior RBSA plan. We do find evidence indicating that such plans are poor proxies for race-based plans, however. Both racial and socioeconomic segregation appear to increase, on average, when race-based student assignment plans are replaced by weak SES-based plans. In sum, then, when SBSA plans do not supplant RBSA plans, or when SBSA plans use SES-balancing or attendance-zone strategies to reduce socioeconomic segregation, they appear to lead to modest reductions in socio-economic segregation and to no change in racial segregation.

Our findings here should be considered tentative, given that they are based on only a small number of school districts. Nonetheless, they do suggest that districts interested in using SES-based school assignment plans to create more integrated schools should use SES-balancing and/or attendance-zone boundary mechanisms. Our data suggest that such plans may be able to substitute for race-based assignment plans, but it is unclear whether that is because the RBSA plans replaced by strong SBSA plans are weak or because the SBSA plans are sufficiently strong. With only six districts where RBSA plans are replaced by strong SBSA plans, we have little empirical evidence from which to generalize.

Finally, our evidence suggests that weaker mechanisms, such as SES-based transfer priority mechanisms, are poor substitutes for race-based assignment plans, leading to higher levels of both racial and socioeconomic segregation than the RBSA plans they replace. And when implemented in districts with no prior RBSA plan, these weak SBSA plans have no effect on segregation levels. SES-based transfer mechanisms appear to be ineffective methods of producing either racial or socioeconomic balance among school enrollments.

We thank Doug Lauen for his comments on an earlier draft. We are indebted to the staff members at dozens of schools districts who took time to answer our questions and explain their student assignment plans to us. In addition, we thank Richard Kahlenberg for sharing data on school district student assignment plans. Demetra Kalogrides provided excellent research assistance. Any errors in the essay remain our responsibility.

(p.205)

(p.206)

(p.207)

Notes:

(1) .Parents Involved in Community Schools v. Seattle School District No. 1, 551 U.S. 701 (2007).

(2) . R. D. Kahlenberg, All Together Now: Creating Middle-Class Schools through Public School Choice (Washington, DC: Brookings Institution Press, 2001); R. D. Kahlenberg, A New Way on School Integration (New York: Century Foundation Press, 2006).

(3) . S. F. Reardon, J. T. Yun, and M. Kurlaender, “Implications of Income-Based School Assignment Policies for Racial School Segregation,” Educational Evaluation and Policy Analysis 28 (2006): 49–75.

(4) . See R. D. Kahlenberg, Rescuing Brown v. Board of Education: Profiles of Twelve School Districts Pursuing Socioeconomic School Integration (New York: Century Foundation Press, 2007). Although Kahlenberg presents case studies of 12 districts that use socioeconomic integration, his analysis is more descriptive than causal and so does not provide clear estimates of the average effect of the socioeconomic integration plans on racial desegregation.

(5) . Board of Education of Oklahoma City Public Schools v. Dowell, 498 U.S. 237 (1991); Freeman v. Pitts, 503 U.S. 467 (1992); Missouri v. Jenkins, 515 U.S. 70 (1995).

(6) . S. F. Reardon and J. T. Yun, “Integrating Neighborhoods, Segregating Schools: The Retreat from School Desegregation in the South, 1990–2000,” North Carolina Law Review 81 (2003): 1566 (n. 7).

(7) . J. Anyon, Ghetto Schooling: A Political Economy of Urban Educational Reform (New York: Teachers College Press, 1997); J. S. Coleman et al., Equality of Educational Opportunity (Washington, DC: U.S. Department of Health, Education, and Welfare, Office of Education, 1966); D. R. Entwisle, K. L. Alexander, and L. S. Olson, Children, Schools, and Inequality (Boulder, CO: Westview, 1997); S. E. Mayer, “How Economic Segregation Affects Children's Educational Attainment,” Social Forces 81 (2002): 153–76; G. Natriello, E. L. McDill, and A. L. Pallas, Disadvantaged Children: Racing against Catastrophe (New York: Teachers College Press, 1990).

(8) . In order to identify districts with SBSA plans, we used several search methods. First, we began with several lists of school districts identified by Kahlenberg as using some form of SBSA plan; Kahlenberg, Rescuing Brown v. Board of Education. Kahlenberg lists 40 districts using some form of SBSA. In the fall of 2008, Kahlenberg (personal communication) provided us with a list of an additional 20 districts that he had subsequently identified as using some form of SBSA. A more recent essay by Kahlenberg lists 65 districts identified as using an SBSA plan, including four districts neither he nor we had previously identified; R. D. Kahlenberg, “Socioeconomic School Integration: Preliminary Lessons from More than 60 Districts,” paper presented at the conference “Looking to the Future: Legal and Policy Options for Racially Integrated Education in the South and the Nation,” Chapel Hill, NC, 2 April 2009. Second, we searched the scholarly literature for articles about SBSA plans. Third, we conducted Internet searches of school-district websites and media sites for references to SBSA plans. These methods yielded an initial list of 116 districts (64 initially identified by Kahlenberg; 52 we identified through other sources) using some form of SBSA plan. We contacted each of these districts by phone or e-mail (we also examined their websites for information on their student assignment plan when possible) in order to verify that the district had some form of SBSA plan and to determine the particulars of the plan. This exercise revealed that many of the districts initially identified do not, in fact, use any form of socioeconomic status measure in their student assignment plan. Many districts that were initially identified as potentially using SBSA plans turned out not to use an SBSA plan in practice. Most commonly, this was because we (or Kahlenberg) identified a district as potentially using an SBSA plan on the basis of a newspaper article or minutes from a school board meeting that indicated some discussion of the issue of socioeconomic segregation, a proposal to adopt an sbsa plan, or both. When we contacted district officials, however, we often learned that no formal policy had been adopted, or that the use of socioeconomic factors in school assignment had been authorized by a school board but never implemented in practice in the district. In addition, we sometimes found that districts had considered socioeconomic factors in some small way when deciding where to locate a single new school or in determining student assignments for a single school. Because these were not formal policies and affected a very small number of students, we did not consider these formal sbsa plans.

(9) . Thirty-six of these were districts identified by Kahlenberg; four others (Burlington Community School District, IA; Stamford, CT; Topeka, KS; and Troup County, GA) were identified from other sources (and were later included in Kahlenberg, “Socioeconomic School Integration”). There were also 4 districts among the 116 for which we were not able to determine with certainty whether they used SBSA plans: University Place, WA; Lee County, NC; Rockford, IL; and Hillsborough County, FL. Given that we were not able to identify many additional districts beyond those identified by Kahlenberg, we are relatively confident that our list contains most of the districts in the United States currently using SBSA plans.

(10) . Throughout this essay, we use the terms “white,” “black,” and “Asian” to refer to non-Hispanic students of the designated race. We use the term “Hispanic” to refer to His-panic students of any race. Thus, the categories “white,” “black,” “Asian,” and “Hispanic” are mutually exclusive in our terminology.

(11) . Reardon, Yun, and Kurlaender, “Implications of Income-Based School Assignment Policies.”

(12) .S. Saporito and D. Sohoni, “Coloring outside the Lines: Racial Segregation in Public Schools and Their Attendance Boundaries,” Sociology of Education 79 (2006): 81–105; S. Saporito and D. Sohoni, “Mapping Educational Inequality: Concentrations of Poverty among Poor and Minority Students in Public Schools,” Social Forces 85 (2007): 1227–53.

(13) . Reardon, Yun, and Kurlaender, “Implications of Income-Based School Assignment Policies.”

(14) . Ibid.

(15) . Although the CCD contains imputed racial, ethnic, and free or reduced-price lunch enrollment counts for schools with missing data in the years 1987–88 to 1997–98, we count the imputed data as missing, because of concerns about the imputation methods used in the CCD (in many cases, data were imputed for every school in a state in a given year, for example).

(16) . D. R. James and K. E. Taeuber, “Measures of Segregation,” in Sociological Methodology, edited by Nancy Tuma (San Francisco: Jossey-Bass, 1985), 1–32; S. F. Reardon and G. Firebaugh, “Measures of Multi-group Segregation,” Sociological Methodology 32 (2002): 33–67; H. Theil, Statistical Decomposition Analysis (Amsterdam: North-Holland, 1972); H. Theil and A. J. Finezza, “A Note on the Measurement of Racial Integration of Schools by Means of Informational Concepts,” Journal of Mathematical Sociology 1 (1971): 187–94; B. S. Zoloth, “Alternative Measures of School Segregation,” Land Economics 52 (1976): 278–98.

(17) . Specifically, for each of the 23 districts for which we have segregation data before and after the implementation of the SBSA plan, we fit a model of the form Ht = βo + f(Yt) + δ(Pt) + ΧtB + εt, where Yt is a variable measuring time (years) centered on the implementation year of the SBSA policy in the district; f is a polynomial function (usually a cubic, though it depends how many years of data we have available for a given district) of Y; Pt is an indicator of the presence of an SBSA policy in the district in year t; and Xt is a vector of time-varying district covariates (racial and ethnic composition, proportion eligible for free or reduced-price lunch, logged district enrollment, and number of schools). For each district i, the model for that district yields The Effects of Socioeconomic School Integration Policies on Racial School Desegregation, an estimate of the change in segregation in the year of the implementation of the SBSA, net of the (cubic) temporal trend and changing district covariates.

(18) . These models have the form The Effects of Socioeconomic School Integration Policies on Racial School Desegregation = ao+ WiГ + ωi, where The Effects of Socioeconomic School Integration Policies on Racial School Desegregation is the estimated effect of the SBSA plan in district i and Wi is a vector of characteristics of the SBSA plan in district i. We fit these models using weighted least squares, using the inverse of the variance of the estimated The Effects of Socioeconomic School Integration Policies on Racial School Desegregation's as weights.

(19) . S. F. Reardon and J. T. Yun, “Suburban Racial Change and Suburban School Segregation, 1987–1995,” Sociology of Education 74, no. 2 (2001): 79–101.

(20) . For example, although one maybe tempted to conclude from model 3 that an SBSA plan that used both SES balancing and attendance zones in a district that did not have a prior RBSA plan would decrease segregation of whites and blacks by −0.080 (a large decrease), this conclusion is not warranted because it relies on extrapolation of the model: there are no districts among the 23 that use both SES balancing and attendance zones and only one (McKinney, TX) that uses either of these mechanisms and did not have a prior RBSA plan.