Worth the Wait?:
The Equity of the Voluntary Interdistrict Desegregation Program in St. Louis, 2005-2009
Ain Akilah Grooms
University of Georgia
March 2013
Paper presented at the
38th Annual Conference of The Association for Education Finance and Policy
New Orleans, LA
Draft: Do Not Cite Without Permission
Worth the Wait? 2
Abstract
Voluntary interdistrict desegregation programs, born out of the Civil Rights Movement,
were designed specifically to address racial segregation in public schools by providing minority
students from urban areas with free transportation to schools in suburban districts. This paper
will use the voluntary interdistrict desegregation program currently in operation in St. Louis as
the foundation of the study, focusing on the five-year span between 2004 and 2009, following
the lifting of the court order in 1999. Parents of past program participants have openly
admitted to enrolling because of access to a “better education.” Using the equity framework
outlined by Berne and Stiefel (1984), this study serves to examine the variation in the
distribution of resources across the 15 suburban districts that participate in the program.
Draft: Do Not Cite Without Permission
Worth the Wait? 3
Worth the Wait?: The Equity of the Voluntary Interdistrict Desegregation Program in St.
Louis, 2005-2009
In 1954, the United States Supreme Court unanimously ruled, in the Brown v. Board of
Education of Topeka case, that racial segregation was unconstitutional, and that separation of
the races denied Black children the equal protection guaranteed by the Fourteenth
Amendment. This case remains one of the most influential lawsuits of the past hundred years,
and has helped in shaping not only the landscape of public education, but of society as a whole.
The Supreme Court’s decision initiated an ongoing nationwide discussion about equal
educational opportunity. Horace Mann (1848) once declared that education was to be “‘the
great equalizer’ of the conditions of men—the balance wheel of the social machinery” (para. 9),
yet, despite the Brown decision, our country’s history of racial, residential, and economic
segregation continues to pose a tremendous obstacle to the creation of equal schools and an
equal society.
Schools were, and still are, a reflection of our communities, and as neighborhoods
changed, so too did the schools. Public schools became more racially and socioeconomically
segregated, as Black, Hispanic/Latino, and poor White children attended city schools, and
affluent White students attended suburban schools as a result of White Flight. When White
families left the cities, they took with them financial and political support for the schools
(Morris, 2004; Schneider, 1992). As the divide between rich and poor (and between suburban
and urban) grew, questions arose regarding the educational opportunities that were being
provided in public schools across the country.
Draft: Do Not Cite Without Permission
Worth the Wait? 4
Voluntary interdistrict desegregation programs,1 born out of the Civil Rights Movement,
were designed specifically to address racial segregation in public schools by providing minority
students from urban areas with free transportation to schools in suburban districts. There are
currently seven programs in operation across the country: Boston, MA; Hartford, CT;
Milwaukee, WI; Minneapolis, MN; Rochester, NY; Palo Alto, CA; and St. Louis, MO. These
programs were either developed through state law as a result of local grassroots movements,
federal court rulings, or state rulings (Wells et al., 2009). Parents of participants in the
voluntary interdistrict desegregation programs have openly admitted to participating in the
program because of access to the “better education” that is being provided in the suburbs
(Armor, 1972; Eaton, 2001; Orfield et al., 1998; Wells & Crain, 1997).
This paper will use the transfer program currently in operation in St. Louis as the
foundation of the study, focusing on the five-year span between 2004 and 2009. The intent of
this paper is not to compare the resources available in St. Louis to those in the neighboring
communities, but rather, to bring attention to the differences between the participating
suburban areas. The word “suburb” generally brings to mind a White, middle- to upper-middle
class community, and using the equity framework outlined by Berne and Stiefel (1984), this
study serves to examine the variation in the distribution of resources across the 15 suburban
districts that participate in the voluntary interdistrict desegregation program in St. Louis.
The Voluntary Interdistrict Desegregation Program in St. Louis
The plaintiffs in the 1972 Liddell v. Board of Education of the City of St. Louis case argued
that the School Board had operated in a discriminatory manner following the 1954 Brown
1
The terms, voluntary interdistrict desegregation programs and transfer programs, are used interchangeably
throughout this research.
Draft: Do Not Cite Without Permission
Worth the Wait? 5
ruling, and as a result, deprived equal educational opportunities to the Black students in St.
Louis. A settlement was finally reached in 1983, which included the creation of a dual transfer
program, where Black students from St. Louis were provided with free transportation to
suburban schools at all grade levels, and non-Black suburban students were eligible to enroll in
city magnet schools. Through this transfer program, the suburbs agreed to increase the
percentage of Black students by at least 15% of their current enrollment, though not to exceed
25% of total enrollment (Heaney & Uchitelle, 2004).
By 1999, following a lengthy process, a bill was passed that ended court-ordered
desegregation of the city’s public schools, but would keep both the transfer program and the
magnet schools. The bill was passed through the combined support of the suburban
Republicans, who realized the program’s financial benefit to their districts, and the urban
Democrats, who supported desegregation. The settlement agreement proposed continued
financial support for the transfer program through a sales tax increase for city residents in order
to raise $40 million to replace the $60 million that had formerly been provided by the courts
(Freivogel, 2002). The suburban districts agreed to keep transfer enrollment at 85 percent of
the 1998-99 transfer enrollment for the first three years of the settlement, with a reduction to
70 percent of the 1998-99 transfer enrollment in years four through six. Beginning in the
seventh year following the settlement, there would be no minimum transfer enrollment
required for the participating districts (Voluntary Interdistrict Choice Corporation [VICC], 2012).
The transfer program utilizes an extensive recruiting process by automatically sending
applications to all Black students currently enrolled in St. Louis Public Schools. Those families
are able to rank the participating suburban school districts paired with the attendance zone in
Draft: Do Not Cite Without Permission
Worth the Wait? 6
which they live, and are assigned to their preferred districts based on available space. Priority is
given to students already enrolled in the transfer program but attending a suburban school
outside of their assigned district, then to new applicants who have a sibling already enrolled in
the program. Finally, the remaining applications are processed on a first-come, first-served
basis. Students may be denied a space in a suburban school if there are no seats available or if
there are significant disciplinary problems in the student’s academic record (VICC, 2012).
Since the Settlement Agreement, the voluntary transfer program has had two five-year
extensions. The Voluntary Interdistrict Choice Corporation’s Board of Directors unanimously
approved the first program extension in 2007, extending the program and new student
enrollment through the 2013-2014 school year. The participating suburban districts could
decide individually whether to accept new students during the extension period, with the
understanding that all currently enrolled students can continue in the program. At the time of
the first extension, two districts decided not to admit any additional students. The second
extension was approved in November 2012, extending the program through the 2018-2019
school year. There are now a total of four districts no longer accepting new transfer students
into their schools (VICC, 2012). This study focuses on the second five years (2004-2009) after
the settlement agreement, which also includes the time before and after the first program
extension.
Based on the original settlement agreement, suburban school districts received a
reimbursement amount equal to the suburban tuition amount (VICC, 2012). Due to persistently
declining enrollment, the reimbursement rate is reviewed annually, as the majority of funding
for the program is provided through the state foundation formula and sales tax revenue.
Draft: Do Not Cite Without Permission
Worth the Wait? 7
Suburban reimbursement was eventually reduced to a maximum of $9,100 per transfer
students, and as of 2012, the reimbursement rate for the suburban districts was further
reduced to $7,000 per transfer student, or the actual per-pupil cost of the district, whichever is
less. The average cost for a transfer student, including tuition (per-pupil amount) and
transportation, is $11,928 (Glaser, 2012), greater than the reimbursement rate, but significantly
less than the per-pupil expenditure amount of $15,861 in the St. Louis Public School System
(Missouri Department of Education, 2013).
The 1999 settlement agreement did not require the participating suburban districts to
enroll the same percentage of transfer students into their schools each year, and as a result,
districts began phasing out a small percentage of available seats each year, approximately five
to six percent annually. Enrollment was at its peak of 14,227 total participating students,
including 1,249 suburban students attending city magnets, during the 1999-2000 school year
(VICC, 2012), the first year following the settlement agreement. Initially, the goal was to have
2,500 suburban students attending the city’s magnet schools; however, the long bus rides and
the lesser-quality academic programs in magnet schools were cited as reasons for the
withdrawal of suburban students from the transfer program (Freivogel, 2002; Heaney &
Uchitelle, 2004). By the 2009-2010 school year, ten years following the settlement agreement,
enrollment had dropped to 6,314 total participating students, including 167 suburban students
attending magnet schools. Enrollment again fell during the 2012-2013 school year to 5,130
total students, with 86 suburban students attending the city’s magnets (VICC, 2012).
The transfer program utilizes an extensive recruiting process by automatically sending
applications for the transfer program to all Black students currently enrolled in St. Louis Public
Draft: Do Not Cite Without Permission
Worth the Wait? 8
Schools. Black families are able to choose from selected suburban school districts paired with
the attendance zone in which they live. Priority is given to students already enrolled in the
transfer program but attending a suburban school outside of their assigned district, then to
new applicants who have a sibling already enrolled in the program. Finally, the remaining
applications are processed on a first-come, first-served basis. Students may be denied a space
in a suburban school if there are no seats available or if there are significant disciplinary
problems in the student’s academic record. Non-Black families living in selected suburbs are
eligible to apply to attend magnet schools in the city and can receive applications directly from
the VICC website or office (VICC, 2012).
Data
Although the voluntary interdistrict desegregation program in St. Louis is a dual transfer
program, this study will specifically examine the variation in the distribution of resources across
the 15 participating suburban districts, as families have chosen to participate because of the
“better” educational opportunities available there. The majority of the prior research on
voluntary desegregation busing programs has been qualitative in nature, where participants
and their families were asked questions about their reasons for enrolling in the program and
whether they would participate again (Armor, 1972; Eaton, 2001, 2006; Orfield et. al, 1998;
Wells & Crain, 1997). Past quantitative studies about these busing programs have looked
expressly at standardized test scores and/or high school graduation rates (Angrist & Lang, 2004;
Eaton & Chirichigno, 2011). While it is critical to understand the individual reasoning for
program participation as well as the resulting academic outcomes, it is equally important to
analyze the educational inputs (resources) available in the participating suburban districts,
Draft: Do Not Cite Without Permission
Worth the Wait? 9
which may ultimately impact achievement. This study will not include the tuition
reimbursement received by the participating suburban districts, but instead will focus on the
resources available to “typical” suburban students in each of their home districts.
The district-level financial data used to examine the resources available in the 16
participating districts (including St. Louis) were obtained from the Common Core of Data’s Local
Education Agency Finance Survey Data (F-33 file) gathered by the National Center for Education
Statistics (NCES), the Missouri Department of Education, the St. Louis County Department of
Revenue, and the U.S. Census Bureau. Berne and Stiefel (1994) suggested that district-level
data has less variability than data collected at the school level, and can be less complex.
Conversely, Card and Krueger (1996) argued that district-level aggregated data presents the
opportunity for omitted variables. While there are important differences in district size and in
the spending of elementary, middle, and high schools, the use of district-level data removes
some of the variation found in school-level data, especially as education dollars are often
allocated to specific categories (Berne & Stiefel, 1994). Four resource categories are included in
this study: spending, teacher, school, and community.
Four spending variables are included in this study: per pupil expenditures, per pupil
revenue received from property taxes, average teacher salary, and local tax effort. Using the F33 file, the number of students, the average expenditures, and the average revenue received
from property taxes were obtained. Based on this information, the average per-pupil
expenditure and the average per-pupil revenue received from property taxes were calculated.
The average teacher salary was obtained from the Missouri Department of Education, and the
school district tax effort was obtained from the Missouri Department of Revenue.
Draft: Do Not Cite Without Permission
Worth the Wait? 10
Two teacher variables were included, and were obtained from Missouri Department of
Education. This information included the percentage of teachers in the district with at least a
Master’s degree and the average years of experience. These teacher variables are often used
to illustrate the differences in teacher characteristics (see Iatarola & Stiefel, 2003; Greenwald,
Hedges, & Laine, 1996; Hanushek, 1994, 1996; Rubenstein et al., 2006). Five district-level
school variables were also obtained from the Missouri Department of Education, including the
percentage of Black students, the percentage of White students, the percentage of students
that qualify for free and reduced priced lunch, the pupil-teacher ratio, and the percentage of
Black graduates.
Finally, in an effort to provide a snapshot of the larger demographic differences
between St. Louis and the participating suburban school districts, this study will also include
data collected from the U.S. Census Bureau. All of the school districts are located within the
larger St. Louis County, which includes unincorporated areas. As a result, census data included
here are based on county subdivisions, which are defined by the Census Bureau as “the primary
divisions of counties and statistically equivalent entities for the reporting of decennial census
data” (U.S. Census Bureau, 2013b, para. 1). The six community variables included are: median
home price, median family income, and demographical information pertaining to race, family
poverty and educational attainment.
Physical distance (number of miles) between St. Louis and the suburban districts is
included to provide an estimate of students’ travel time. Integrationists that have advocated
for busing in the past argued that, “the greater the distance the student travels to get to the
school, relative to options available to him, the more the school should offer him when he
Draft: Do Not Cite Without Permission
Worth the Wait? 11
arrives” (Campbell, 1973, p. 482). The number of miles between the sending and receiving
districts is used to illustrate the divide between urban and suburban districts, and to determine
whether there is a relationship between the resources available in a suburban district relative
to its proximity to the sending district.
Methods
This study relies upon both horizontal and vertical equity analyses to examine the
variation in the distribution of educational resources among the suburban receiving districts
only. These particular 15 suburban districts are unique in that they chose to continue
partnering with the St. Louis Public School system after the end of court ordered desegregation
(even if they are no longer enrolling new transfer students), and therefore, it is necessary to
document the educational resources to which the transfer students will have continued access.
These equity analyses cover the five years between the 2004-05 school year and the 2008-09
school year.
Horizontal equity was developed to assess the equality of the distribution of a particular
resource across districts, arguing that equal students should be treated equally (Berne & Stiefel,
1984). The range, the coefficient of variation, and the McLoone Index will be utilized to
examine horizontal equity across each of the five years. The appropriate equity thresholds
used in this study were developed by Odden and Picus (2008).
Vertical equity relies upon multivariate linear regression and was developed to assess
the variation in access to resources across districts, based on the assumption that different
students require differing amounts of resources to improve achievement (Berne & Stiefel,
1984). In this study, the vertical equity of the percentage of Black students graduating from
Draft: Do Not Cite Without Permission
Worth the Wait? 12
high school in the participating districts was analyzed using selected variables from each of the
four resource categories. The vertical equity analysis was conducted via a weighted least
squares regression analysis with district enrollment as the weighted variable to adjust for
district size.
As noted, the sample size for this study was only 15 suburban districts. Researchers
have suggested a ratio of 10 cases per each variable (Maxwell, 2000). However, Steinberg
(2007) notes that policy research “often deliberately focuses on a small number of cases that
show a unique departure from the norm—whether these are exemplary accomplishments or
cautionary tales—and which therefore contain important lessons for the larger universe of
policy practice” (p. 185). These 15 districts represent a “departure from the norm” in that they
have volunteered to remain policy actors in addressing the larger metropolitan issues of
desegregation and educational equity. In an effort to maximize the statistical significance of
the regression model, a composite database was created in which each suburban district in
each of the five years was assigned an individual case, totaling a sample size of 75 cases.
Findings
Descriptive Statistics
Prior to conducting the horizontal and vertical input equity analyses, it was necessary to
conduct an analysis using descriptive statistics (Tables 1-4) in order to note the differences
between the sending district, St. Louis, and the receiving suburban districts between 2005 and
2009. As evidenced in Table 1, across the four spending variables (per pupil revenue from
property tax, local tax effort, per pupil expenditure, and average teacher salary), St. Louis had
lower teacher salaries and less per pupil revenue from property taxes than the suburban
Draft: Do Not Cite Without Permission
Worth the Wait? 13
average across all five years. However, St. Louis had higher per pupil expenditures than the
suburban average in four out of the five years.
Table 1. Descriptive Statistics, Spending Variables, 2005-2009
Mean (SD)
Spending Variables
Per Pupil Revenue from
Property Tax
St. Louis
Suburban Average (n=15)
Tax Effort
St. Louis
Suburban Average
Per Pupil Expenditure
St. Louis
Suburban Average (n=15)
Average Teacher Salary
St. Louis
Suburban Average (n=15)
2005
2006
2007
2008
2009
3,340
5,532 (2,245)
3,325
5.747 (2,568)
3,779
6.103 (2,352)
5,362
6,636 (2,659)
5,318
6,616 (2,550)
*
3.95 (.68)
*
3.92 (.68)
*
3.58 (.55)
*
3.60 (.57)
*
3.81 (.63)
11,885
10,208 (2,714)
11,729
10,397(2,881)
10,979
11,351 (2,822)
15,576
12,365 (3,432)
16,675
13,382 (4,162)
40,657
50,960 (5,117)
43,250
51,831 (5,867)
45,454
53,093 (5,623)
47,260
54,870 (5,767)
45,840
56,150 (5,918)
*unavailable at the time of this writing
Source: Missouri Department of Education, 2013; National Center for Education Statistics,
2013; St. Louis County Department of Revenue, 2013
For the two teacher variables (Table 2), St. Louis Public Schools employed teachers with
fewer years of experience on average than did the suburbs, and employed substantially fewer
teachers with Masters degrees. The descriptive analysis of the five school variables (Table 3)
shows that St. Louis is a majority Black school district (although the city is only 49.2% Black)
while the suburban communities and schools are majority White. Both the Black population in
the suburban schools and the White population in the city schools have declined over the five
years. The percentage of students who qualify for free and/or reduced lunch is also greater in
the St. Louis school district, but is also steadily declining, while that same population is rising in
suburban schools. The pupil-teacher ratio in St. Louis and in the suburbs have both declined
over the past five years, but it still remains larger in St. Louis. Finally, the percentage of Black
Draft: Do Not Cite Without Permission
Worth the Wait? 14
graduates in St. Louis, which experienced a dip in 2007, is substantially smaller than the
percentage in the suburbs, which also experienced a dip in 2006.
Table 2. Descriptive Statistics, Teacher Variables, 2005-2009
Teacher Variables
Average Years of Experience
St. Louis
Suburban Average (n=15)
Percent Teachers w/ Masters
degrees
St. Louis
Suburban Average (n=15)
2005
2006
Mean (SD)
2007
2008
2009
12.7
12.95 (1.68)
12.2
12.61 (2.04)
12.5
12.61 (1.91)
11.6
12.67 (1.78)
11.5
12.55 (1.77)
42.5
68.35 (10.59)
40.9
68.71 (11.62)
42.4
70.7 (11.36)
43.9
71.27 (9.75)
44.3
70.81 (10.18)
Source: Missouri Department of Education, 2013
Table 3. Descriptive Statistics, School Variables, 2005-2009
School Variables
Percent Black Students
St. Louis
Suburban Average (n=15)
Percent White Students
St. Louis
Suburban Average (n=15)
Percent FRL Students
St. Louis
Suburban Average (n=15)
Student Teacher Ratio
St. Louis
Suburban Average (n=15)
Percent of Black Graduates
St. Louis
Suburban Average (n=15)
2005
2006
Mean (SD)
2007
2008
2009
80.9
20.35 (7.31)
81.8
19.54 (7.48)
81.7
19.67 (8.13)
81.4
19.58 (8.33)
81.0
19.13 (8.65)
15.2
73.69 (8.69)
14.0
73.73 (9.06)
13.6
73.1 (9.80)
13.6
72.53 (10.31)
13.7
72.44 (10.63)
86.1
29.67 (18.54)
81.0
29.97 (19.30)
80.1
29.14 (19.39)
72.3
29.2 (19.70)
68.7
30.82 (20.67)
20
16.93 (2.6)
19
16.80 (2.54)
19
16.86 (2.64)
18
16.53 (2.56)
18
15.8 (2.62)
59.6
89.65 (8.3)
55.1
86.1 (7.60)
54.2
89.85 (5.46)
56.2
90.48 (5.29)
57.5
91.76 (7.29)
Source: Missouri Department of Education, 2013
Data on the community variables (Table 4) obtained from the U.S. Census Bureau finds
that St. Louis residents are, on average, poorer, less White, and have less educational
achievement than the residents of the participating suburban districts. Median family income
is approximately $43,000 less in St. Louis, home values are approximately $110,000 less in St.
Draft: Do Not Cite Without Permission
Worth the Wait? 15
Louis, and there are approximately 17% more families living in poverty in St. Louis than in the
participating suburbs. While the percentage of White families in St. Louis is 46.6% as compared
to the suburban average of 88.13%, the percentage of Black families in the suburbs is almost
negligible at 5.7%, compared to the St. Louis average of 48.2%. Interestingly, however, is that
Black students account for an average of approximately 20% of the student population in the
suburbs, while Black families only account for an average of 6% of suburban families. According
to these data, the transfer program is responsible for the majority of the Black students
attending suburban schools, traveling an average of almost 30 miles round trip each day.
Table 4. Descriptive Statistics, Community Variables, 2009
Variable
Median Family Income (in dollars)
Mean (SD)
2009
St. Louis
Suburban Average (n=15)
41,349
84,329 (22,568)
St. Louis
Suburban Average (n=15)
District Percent Families in Poverty
St. Louis
Suburban Average (n=15)
District Percent Residents Over 25 with Bachelors Degrees
St. Louis
Suburban Average (n=15)
District Percent White
St. Louis
Suburban Average (n=15)
District Percent Black
St. Louis
Suburban Average (n=15)
Distance from St. Louis (in miles)
St. Louis
Suburban Average (n=15)
119,900
220,972 (84,550)
Median Home Value (in dollars)
Source: U.S. Census Bureau, 2013a
21
3.77 (2.33)
14.9
26.2 (9.11)
46.6
88.13 (6.74)
48.2
5.7 (6.14)
14.79 (5.20)
Draft: Do Not Cite Without Permission
Worth the Wait? 16
Horizontal Equity Analysis
To calculate the horizontal equity of resources among the 15 participating suburban
districts, three equity measures were used: the range, the coefficient of variation, and the
McLoone Index (see Table 5). The range is used in this study, rather than the restricted range
or range ratio, due to the small sample size.
The horizontal analysis was conducted on a total of 15 variables from all four resource
categories. The community variables were included in this analysis with the understanding that
although transfer students do not have direct access to those particular variables, it is inferred
that the transfer students experience an increase in social and cultural capital through
interactions with the suburban students who do have direct access to those variables by living
in these communities. The enrollment variable was only included in the range calculations.
The range is the difference between the highest and lowest resource amounts in the
sample—the greater the range, the greater the inequity. Over the five years, the ranges of
most of the variables exhibited a positive average annual change; however, the range in district
enrollments (between 21,000 and 22,000 students) and the percentage of students who
qualified for free and reduced lunch (hovering at 62-63%) remained relatively stable over the
five years, with average changes of 0.40% and 0.19%, respectively. The range in the percentage
of White students in the school experienced the greatest average annual increase, at 9.47%,
growing from a difference of 28% in 2005 to a difference of almost 41% in 2009. Similarly, the
percentage of Black students experienced an average annual increase of 8.29%, growing from a
difference of 23% in 2005 to almost 32% in 2009. The range of per pupil expenditures also
increased at an average annual change of 9.17%, from $9,900 in 2005 to $13,400 in 2009. This
Draft: Do Not Cite Without Permission
Worth the Wait? 17
data shows that schools are becoming more racially and socioeconomically isolated as the gap
widens between the districts with fewer and greater numbers of Black and White students and
between wealthier and less wealthy suburban districts.
Two variables exhibited a substantial negative average annual change in their ranges
over the five-year period: tax effort (-3.90%) and the percentage of Black graduates (-6.15%).
The range in tax effort decreased to 1.88 in 2009, down from 2.25 in 2005. Interestingly,
although the gap is decreasing between property-rich and property-poor districts as evidenced
by the decrease in the range in tax effort, the average change in the range between per pupil
expenditures is increasing almost twice as much over the same time period. Also of note is the
decrease in the gap between districts with fewer and greater numbers of Black graduates (20%
in 2009, down from 28% in 2005), yet, as evidenced by Table 3, there has been an overall
decrease in the average percentage of Black students enrolled in suburban schools over the five
years. The decrease in the range simply signifies that of the few Black students enrolled, more
are graduating.
The coefficient of variation, calculated by dividing the standard deviation by the mean,
is used to determine the variation about the mean, or the percentages of the observations that
fall within one standard deviation of the mean (Picus et al., 2001). A coefficient of variation at
or below 0.05 represents equity (Berne & Stiefel, 1984; Guthrie et al., 2007; Odden & Picus,
2008). None of the variables experienced appropriate levels of equity over the five-year span.
Draft: Do Not Cite Without Permission
Worth the Wait? 18
Table 5: Horizontal Equity for Select Variables, Suburban Districts (n=15)
Measure
Variable
2005
Per Pupil Revenue from Property Tax
8,205
Tax Effort
2.2561
Per Pupil Expenditure
9,861
Average Teacher Salary
18,560
Average Years of Experience
6.1
Percent Teachers w/ Masters degrees
37.8
School Percent Black
23.00
School Percent White
28.40
Range
School Percent FRL
63.20
Student Teacher Ratio
8.0
Percent of Black Graduates
28.2
Enrollment
21,389
Percent Residents with Bachelors*
Percent Families in Poverty*
Median Family Income*
Median Home Value*
Per Pupil Revenue from Property Tax
.41
Tax Effort
.17
Per Pupil Expenditure
.27
Average Teacher Salary
.10
Average Years of Experience
.13
Percent Teachers w/ Masters degrees
.15
School Percent Black
.36
Coefficient of
School Percent White
.12
Variation
School Percent FRL
.62
Student Teacher Ratio
.15
Percent of Black Graduates
.09
Percent Residents with Bachelors*
Percent Families in Poverty*
Median Family Income*
Median Home Value*
Per Pupil Revenue from Property Tax
.76
Tax Effort
.88
Per Pupil Expenditure
.82
Average Teacher Salary
.91
Average Years of Experience
.86
Percent Teachers w/ Masters degrees
.88
School Percent Black
.62
McLoone
School Percent White
.89
Index
School Percent FRL
.79
Student Teacher Ratio
.82
Percent of Black Graduates
.91
Percent Residents with Bachelors*
Percent Families in Poverty*
Median Family Income*
Median Home Value*
*five year average as reported by the U.S. Census Bureau (2013a)
2006
8,940
2.2860
10,343
23,582
7.4
42.32
24.40
32.10
60.60
8.0
28.9
21,841
2007
8,691
1.8245
9,873
21,922
6.6
42.9
27.30
36.20
62.60
8.0
19.1
21,756
2008
9,401
1.8584
10,483
22,719
7.2
39.7
29.30
39.00
62.80
8.0
20.5
21,929
.47
.18
.28
.11
.16
.17
.38
.12
.64
.15
.09
.39
.15
.25
.11
.15
.16
.41
.13
.67
.16
.06
.40
.15
.27
.10
.14
.13
.43
.14
.67
.15
.06
.59
.88
.81
.89
.83
.84
.58
.90
.72
.81
.91
.65
.90
.82
.89
.87
.86
.56
.90
.73
.81
.93
.68
.90
.78
.91
.90
.90
.56
.90
.73
.82
.94
2009
9,298
1.8843
13,644
21,565
7.2
34.8
31.60
40.70
63.60
11.0
20.4
21,731
23.6
8.6
67,296
0.75
.38
.16
.31
.11
.14
.14
.45
.15
.67
.17
.08
0.35
0.62
0.27
0.38
.67
.89
.85
.91
.88
.88
.55
.90
.68
.85
.93
0.53
0.68
0.75
0.73
%Change
3.31%
-3.90%
9.17%
4.64%
4.90%
-1.62%
8.29%
9.47%
0.19%
9.38%
-6.15%
0.40%
-1.21%
-1.03%
3.95%
2.73%
2.54%
-0.90%
5.74%
5.79%
1.98%
3.44%
0%
-2.26%
0.29%
1.03%
0.01%
0.64%
0.07%
-2.92%
0.28%
-3.58%
0.92%
0.55%
Draft: Do Not Cite Without Permission
Worth the Wait? 19
The two most inequitable variables were the percentage of students who qualify for
free and reduced lunch and the per pupil revenue from property taxes. The coefficient of
variation for the percentage of students qualifying for free and reduced lunch grew from 0.62 in
2005 to 0.67 in 2009, meaning that the distribution of poorer students was inequitable across
the 15 participating suburban districts. The average change in over the five years was also
relatively small, just under 2%. Although the coefficient of variation in per pupil revenue from
property tax decreased from 0.41 in 2005 to 0.38 in 2009, the distribution remained inequitable
across the five years.
The two variables with the lowest coefficients of variations were average teacher salary
and the percentage of Black graduates. Despite the low coefficient of variation in average
teacher salary, it still increased from 0.10 in 2005 to 0.11 in 2009. The coefficient of variation in
the percentage of Black graduates decreased from 0.09 in 2005 to 0.08 in 2009, which may be
again attributed to the decline in the number of Black students attending suburban schools.
While there was an unequal distribution across all of the community variables, the least
inequitable was the median family income, with a coefficient of variation of 0.27. The
percentage of families in poverty was the most inequitable, with a coefficient of variation of
0.62, almost mirroring the coefficient of variation of the percentage of students who qualify for
free and reduced priced lunch. This data reflects the within-race socioeconomic disparities
present in the suburbs, as gaps persist between rich and poor in the increasingly White
participating suburbs.
Overall, the distributions of the selected school, teacher, and spending variables not
only remained inequitable over the five years, but for most variables, the inequality increased
Draft: Do Not Cite Without Permission
Worth the Wait? 20
during that same time period. The two variables with the greatest average increase over time
in their coefficients of variation were the percentages of Black and White students. The
distribution of the percentage of Black students was more inequitable (growing from 0.36 to
0.45) than the distribution of the percentage of White students (growing from 0.12 to 0.15), but
both increased at approximately 6% over the five years. This reinforces the notion of the
widening gap between racially isolated and diverse schools, despite the efforts of the voluntary
interdistrict desegregation program.
Unlike the previously mentioned calculations, the McLoone Index focuses only on those
districts below the median level of resources. The McLoone measures the ratio of the actual
resources for the districts below the median to what the resources would be if the resources in
those districts were raised to the median level. Another difference between the McLoone
Index and the other calculations is that, for the McLoone, a ratio of 1.0 represents equality and
zero represents inequality (Berne & Stiefel, 1984; Guthrie et al., 2007; Odden & Picus, 2008).
The results from Table 5 show that the distribution in resources using the McLoone
Index were, in general, more equitably distributed as compared to the calculations relying on
the coefficients of variation; however, the only equitable distributed variable was the
percentage of Black graduates, consistently reaching the appropriate equity threshold of 0.90
each year. Other more equitably distributed resources included tax effort, average teachers
salary, and the percentage of White students. Most variables experienced negligible or very
small average annual increases over time, denoting their stability. As the McLoone Index is
reported in the reverse of the coefficient of variation, a negative annual percent change
indicates an increase in inequity.
Draft: Do Not Cite Without Permission
Worth the Wait? 21
The three least equitably distributed variables also experienced the greatest negative
average annual changes over time, exhibiting an increase in inequality. The percentage of Black
students remained the most inequitably distributed, with a McLoone Index declining from 0.62
in 2005 to 0.55 in 2009, demonstrating increasing inequality. The McLoone Index of the
percentage of students that qualify for free and reduced priced lunch declined from 0.79 in
2005 to 0.68 in 2009. Finally, the McLoone of the per pupil revenue from property taxes
declined from 0.76 in 2005 to 0.67 in 2009, experiencing a significant dip in 2006 with a
McLoone Index of 0.59.
As evidenced by the horizontal equity analyses, the percentage of Black students in the
suburban school districts is the most inequitably distributed across all five years and across all
three equity measures. This is important to note because although the 15 participating suburbs
have continuously committed to the transfer program’s mission of school integration, as the
program policies and district enrollment became left up to the discretion of the individual
suburbs, levels of inequality increased. Despite the equal distribution among Black graduates,
those levels are still reflective of the decline in enrollment.
The distribution of local tax effort, or the amounts that residents are taxed for
education, has become more equitably distributed across all years and all equity measures
while per pupil expenditures, the amount that districts are spending, have become less
equitably distributed. The distributions of the two spending variables demonstrate the
variations in spending and wealth among the suburban districts.
Draft: Do Not Cite Without Permission
Worth the Wait? 22
Vertical Equity Analysis
Parents of past program participants stated that they chose to enroll their children in
the program because of the access to a “better” education. For the purposes of this study, a
“better” education is defined as the percentage of Black students that graduate. Table 3 shows
a stark contrast in the graduation rates of the St. Louis Public Schools (where program
participants would have been assigned) to that in the suburban schools. In 2009, Black
students in St. Louis had an average graduation rate of 57.5%, compared to 91.76% in the
suburbs, evidence that Black students in suburban schools are more likely to graduate from
high school, and thus more likely to have access to improved postsecondary opportunities,
including college. Table 3 also indicates that the transfer students comprise the majority of the
Black students enrolled in the suburban schools, therefore, the percentage of Black graduates
will be used as the dependent variable in the model. District enrollment was included as a
weighted variable.
The percentage of Black students and the percentage of students that qualify for free
and reduced lunch was not found to be highly correlated (r=.481, p<.000), making an
interaction term unnecessary for the model. The percentage of students that qualify for free
and reduced priced lunch was found to be highly correlated with each of the community
variables corresponding to wealth and/or educational attainment (median family income,
median home value, percentage of residents over the age of 25 with Bachelors degrees, and
the percentage of families living in poverty), and as such, the community variables were not
included in the model. A dichotomous variable, “continue”, was created to specify whether the
district chose to continue enrolling new students in that particular year. Although each district
Draft: Do Not Cite Without Permission
Worth the Wait? 23
participated throughout each of the five years of the study, by 2009, four districts had decided
to no longer enroll new transfer students with the understanding that currently participating
students would be able to graduate from their suburban high schools.
The results of the vertical equity analysis can be found in Table 6. Initially, five separate
vertical equity analyses were conducted in order to determine the variation among predictor
variables by year. Although several variables were statistically significant across years, the
models (designed the same as the one in Table 6) were not, and instead, the analysis was
conducted on the composite database.
Table 6. Vertical Equity Regression Analysis, Graduation Rate of Black Students by Weighted
District Enrollment, 2005-2009
Independent Variables
School Percent Black
School Percent FRL
Per Pupil Expenditure
Percent Teachers with Masters Degrees
Tax Effort
Distance
Continue
F-score
Adjusted R2
Standardized Coefficient
-.490*
.631*
.589*
.242**
3.488*
.191
N=75; *p<.05; **p<.10
The model explained 19% of the variation in the graduation rates of Black students in
these participating suburban districts. The results indicate that the strongest predictor of the
graduation rate of Black students is the percentage of students that qualify for free and
reduced price lunch. The standardized coefficient, .631, was statistically significant. The
second strongest predictor, with a statistically significant standardized coefficient of .589, was
the per pupil expenditure. The weakest predictor of the graduation rate of Black students was
the distance of the suburb from the city of St. Louis. The standardized coefficient of .242 was
Draft: Do Not Cite Without Permission
Worth the Wait? 24
statistically significant. While this supports Coleman’s (1973) claim that the farther a student
travels to school, the more resources should be available, this model demonstrates that
distance is not the primary predictor of achievement.
Interestingly, the percentage of Black students in the district was the only negative
predictor of the graduation rate of Black students, with a statistically significant standardized
coefficient of -.490. These results imply that as the percentage of Black students enrolled in a
district increases (with the descriptive analyses showing that the transfer program is
responsible for the majority of the Black students enrolled), the graduate rate of the Black
students decrease. There was no statistically significant relationship between the graduation
rate of Black students and either the percentage of teachers in the district with Masters
degrees, the district’s tax effort, or whether districts continued to enroll new transfer students.
Conclusion
The use of the horizontal and vertical equity analyses has proven useful in determining
the equity of school finance funding formulas across the nation. Over the years, these analyses
have helped bring attention to the differing provision, whether legitimately or illegitimately, of
educational opportunities to particular groups of students. The results of the horizontal and
vertical equity analyses presented here bring attention to issues of integration in the
metropolitan St. Louis area. The settlement agreement passed in 1999 was designed to
continue the program for an additional ten years, and in 2007, the districts voted to continue
the program for another five years, through the 2012-13 school year. Following the 2007 vote,
districts were allowed to individually determine the number of transfer students they would
Draft: Do Not Cite Without Permission
Worth the Wait? 25
accept into their schools. While it is true that most of the participating suburban districts
continue to accept new transfer students from St. Louis, that number is steadily declining.
Two conclusions can be drawn from the horizontal equity analyses. First, all suburbs are
not the same, as evidenced by the levels of inequity present in per pupil revenue from property
tax, per pupil expenditures, the percentage of residents age 25 and over with Bachelor’s
degrees, the percentage of students that qualify for free and reduced priced lunch, and median
home values and family incomes. Just because a transfer student chooses and is assigned to a
particular suburban district, assumptions cannot automatically be made about the type of
suburban community in which the schools are located. Second, as districts began to make
increasingly individualized program and policy decisions regarding enrolling additional transfer
students, levels of inequity increased, as evidenced by the percentage of Black students in the
schools.
The vertical equity analysis here supports previous research claims that school
composition (the percentages of Black and White students and students that qualify for free
and reduced lunch) and community wealth (per pupil expenditures and per pupil revenue from
property taxes) affect graduation rates. However, because these districts choose to
intentionally increase their percentage of Black students, despite the decreasing rate over time,
it is critically important to conduct further regression analyses to uncover additional predictors
of the graduation rates of Black students, including peer effects pertaining to social and cultural
capital and funding specifically designated for support services and/or low-income students.
Throughout this research, questions persist about policy implementation, resource
allocation, and equality of educational opportunity in the metropolitan St. Louis area. The
Draft: Do Not Cite Without Permission
Worth the Wait? 26
opposite side of this dual transfer program, (White, suburban students attending the city’s
magnet schools), is also experiencing declining enrollment, but based on individual choice
rather than district policy. Diversity and achievement are two components explicit to the
mission and success of the transfer program, yet the declining enrollments of White students in
city magnets by choice suggests, first, their belief that city schools will not provide high-quality
education; second, that Black students and families will continue to shoulder the burden of
integration more often than White students and families; and third, diversity and integration is
not as important as academic achievement.
At the height of the program, almost 15,000 transfer students participated—the vast
majority coming from the city—and despite smaller enrollment numbers due to individual
enrollment policies adopted by the participating suburban districts, this program is still a
popular school choice option among Black families in St. Louis. Attention is often rightly
focused on the disparities between urban and suburban districts, yet as Black families continue
to participate in this voluntary transfer program with the belief that their children will be
attending school in better districts, this research rightfully expands the discussion to include
disparities between the participating suburban districts.
Draft: Do Not Cite Without Permission
Worth the Wait? 27
References
Angrist, J.D. & Lang, K. (2004). Does school integration generate peer effects? Evidence from
Boston’s METCO Program. The American Economic Review, 94(5), 1613-1634.
Armor, D. J. (1972). The Evidence on Busing. The Public Interest, 28, 90-126.
Berne, R. & Stiefel, L. (1984). The measurement of equity in school finance: Conceptual,
methodological, and empirical dimensions. Baltimore, MD: The Johns Hopkins
University Press.
Berne, R. & Stiefel, L. (1994). Measuring equity at the school level: The finance perspective.
Educational Evaluation and Policy Analysis, 16, 405-421.
Campbell, E. Q. (1973). Defining and attaining equal educational opportunity in a pluralistic
society. Vanderbilt Law Review, 26, 461-486.
Card, D. & Krueger, A. B. (1996). School resources and student outcomes: An overview of the
literature and new evidence from North and South Carolina. Journal of Economic
Perspectives, 10, 31-50.
Eaton, S. (2001). The other Boston busing story: What’s won and lost across the boundary line.
New Haven, CT: Yale University Press.
Eaton, S. & Chirichigno, G. (2011). METCO merits more: The history and status of METCO.
Boston, MA: Pioneer Institute.
Freivogel, W. H. (2002). St. Louis: Desegregation and school choice in the land of Dred Scott.
In The Century Foundation (Ed.) Divided we fail: Coming together through public school
choice (pp. 209-235). New York, NY: Century Foundation Press.
Draft: Do Not Cite Without Permission
Worth the Wait? 28
Glaser, D. (2012). Memorandum. Retrieved from
http://www.boarddocs.com/mo/sdclay/Board.nsf/files/8XGSGA6C70BE/$file/VST2012r
eport.pdf.
Greenwald, R., Hedges, L. V., & Laine, R. D. (1996). The effect of school resources on student
achievement. Review of Education Research, 66, 361-396.
Guthrie, J. W., Springer, M. G., Rolle, A. R., & Houck, E. A. (2007). Modern educational finance
and policy. Boston, MA: Pearson.
Hanushek, E. A. (1994). The impact of differential expenditures on school performance.
Educational Researcher, 18(4), 45-62.
Hanushek, E. A. (1996). School resources and student performance. In G. Burtless (Ed.) Does
money matter? The effect of school resources on student achievement and adult
success, (pp. 43-73). Washington, DC: Brookings Institution Press.
Heaney, G. W. & Uchitelle, S. (2004). Unending struggle: The long road to an equal education
in St. Louis. St. Louis, MO: Reedy Press.
Iatarola, P. & Stiefel, L (2003). Intradistrict equity of public education resources and
performance. Economics of Education Review, 22, 69-78.
Mann, H. (1848). Twelfth Annual Report to the Massachusetts State Board of Education.
Retrieved from usinfo.org/docs/democracy/16.htm.
Maxwell, S. E. (2000). Sample size and multiple regression analysis. Psychological Methods, 5,
434-458.
Missouri Department of Education. (2013). Missouri comprehensive data system. Retrieved
from http://mcds.dese.mo.gov/quickfacts/SitePages/DistrictInfo.aspx.
Draft: Do Not Cite Without Permission
Worth the Wait? 29
Morris, J. E. (2004). Can anything good come from Nazareth? Race, class, and African
American schooling and community in the urban south and midwest. American
Educational Research Journal, 41, 69-112.
National Center for Education Statistics. (2013). Common core of data - Local education agency
(school district) finance survey (F-33) data. Retrieved from
http://nces.ed.gov/ccd/f33agency.asp.
Odden, A. R. & Picus, L. O. (2008). School finance: A policy perspective (4th ed). New York, NY:
McGraw-Hill.
Orfield, G., Arenson, J., Jackson, T., Bohrer, C., Gavin, D., Kalejs, E. (1998). Summary of “CitySuburban Desegregation: Parent and student perspectives in Metropolitan Boston,” a
Report by the Harvard Civil Rights Project. Equity & Excellence in Education, 31, 6-12.
Rubenstein, R., Schwartz, A. E. Steifel, L, & Bel Hadj Amor, H. (2007). From districts to schools:
The distribution of resources across schools in big city school districts. Economics of
Education Review, 26, 532-545.
Schneider, W. (1992). The suburban century begins. The Atlantic Monthly. Retrieved from
http://www.theatlantic.com/past/politics/ecbig/schnsub.htm
St. Louis County Department of Revenue. (2013). Tax rates summary. Retrieved from
http://revenue.stlouisco.com/collection/TaxRates.aspx.
Steinberg, P. F. (2007). Causal assessment in small-n policy studies. The Policy Studies Journal,
35, 181-204.
U.S. Census Bureau. (2013a). American fact finder. Retreived from
http://factfinder2.census.gov/faces/nav/jsf/pages/index.xhtml.
Draft: Do Not Cite Without Permission
Worth the Wait? 30
U.S. Census Bureau. (2013b). Cartographic boundary files. Retrieved from
https://www.census.gov/geo/www/cob/cs_metadata.html.
Voluntary Interdistrict Choice Corporation. (2012). Voluntary interdistirict choice corporation.
Retrieved from http://www.choicecorp.org/default.htm.
Wells, A. S., Baldridge, B. J., Duran, J. D., Grzesikowski, C., Lofton, R., Roda, A., …White, T.
(2009). Boundary crossing for diversity, equity and achievement: Inter-district school
desegregation and educational opportunity. Cambridge, MA: Charles Hamilton Houston
Institute for Race & Justice.
Wells, A. S. & Crain, R. L. (1997). Stepping over the color line: African-American students in
white suburban schools. New Haven, CT: Yale University Press.