School Finance: Per-Pupil Spending Differences between Selected
Inner City and Suburban Schools Varied by Metropolitan Area
(09-DEC-02, GAO-03-234).
The No Child Left Behind Act of 2001 has focused national
attention on the importance of ensuring each child's access to
equal educational opportunity. The law seeks to improve the
performance of schools and the academic achievement of students,
including those who are economically disadvantaged. The Congress,
among others, has been concerned about the education of
economically disadvantaged students. This study focused on
per-pupil spending, factors influencing spending, and other
similarities and differences between selected high-poverty inner
city schools and selected suburban schools in seven metropolitan
areas: Boston, Chicago, Denver, Fort Worth, New York, Oakland,
and St. Louis.
-------------------------Indexing Terms-------------------------
REPORTNUM: GAO-03-234
ACCNO: A05669
TITLE: School Finance: Per-Pupil Spending Differences between
Selected Inner City and Suburban Schools Varied by Metropolitan
Area
DATE: 12/09/2002
SUBJECT: Academic achievement
Disadvantaged persons
Economically depressed areas
Educational research
Educational testing
Federal funds
Financial analysis
Funds management
Schools
Students
Comparative analysis
Department of Education National
Assessment of Educational Progress
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO Product. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
******************************************************************
GAO-03-234
* Results in Brief
* Background
* Spending Differences between Selected Inner City and Suburban Schools
Varied by Metropolitan Area
* Differences in Per-Pupil Spending between Selected Inner City
Schools and Suburban Schools Varied by Metropolitan Area
* Average Teacher Salaries, Student-Teacher Ratios, and Ratios of
Students to Student Support Staff Accounted for Most of the
Differences in School Spending in Selected Schools
* Inner City Schools at a Disadvantage When Spending Adjusted for
Student Needs
* Federal Funds Played Important Role in Helping Inner City Schools
Meet Expenses
* Inner City Schools Generally Faced Greater Challenges That May Have
Affected Student Achievement
* Inner City Students' Achievement Scores Were Generally Lower than
Suburban Students' Achievement Scores
* Inner City Schools We Reviewed Had More First-Year, Thus Less
Experienced, Teachers than Suburban Schools
* Enrollment Was Higher in Inner City Schools than in Suburban
Schools, and Buildings Were Older
* Inner City Schools Had Less Library and Technological Support
than Suburban Schools
* In-School Parental Involvement Differed between Selected Inner
City and Suburban Schools
* Conclusions
* Agency Comments
* Appendix I: Objectives, Scope, and Methodology
* Scope
* Metropolitan Area Selection
* Defining Inner City and Suburbs
* School Selection
* Data Collected
* Methodology to Analyze Differences in Spending and Factors
Accounting for Spending Differences
* Appendix II: School Profiles
* Appendix III: GAO Contacts and Staff Acknowledgments
* GAO Contacts
* Acknowledgments
* GAO's Mission
* Obtaining Copies of GAO Reports and Testimony
* Order by Mail or Phone
* To Report Fraud, Waste, and Abuse in Federal Programs
* Public Affairs
Report to the Ranking Minority
GAO
Member, Committee on Ways and Means, House of Representatives
December 2002
SCHOOL FINANCE
Per-Pupil Spending Differences between Selected Inner City and Suburban Schools
Varied by Metropolitan Area
GAO-03-234
SCHOOL FINANCE
Per-Pupil Spending Differences between
Highlights of GAO-03-234, a report to the
Selected Inner City and Suburban
Ranking Minority Member, Committee on Ways and Means, House of
Representatives
Schools Varied by Metropolitan Area
The No Child Left Behind Act of 2001 has focused national attention on the
importance of ensuring each child's access to equal educational
opportunity. The law seeks to improve the performance of schools and the
academic achievement of students, including those who are economically
disadvantaged. The Congress, among others, has been concerned about the
education of economically disadvantaged students. This study focused on
per-pupil spending, factors influencing spending, and other similarities
and differences between selected high-poverty inner city schools and
selected suburban schools in seven metropolitan areas: Boston, Chicago,
Denver, Fort Worth, New York, Oakland, and St. Louis.
Among the schools GAO reviewed, differences in per-pupil spending between
inner city and suburban schools varied across metropolitan areas, with
inner city schools spending more in some metropolitan areas and suburban
schools spending more in other areas. The inner city schools that GAO
examined generally spent more per pupil than suburban schools in Boston,
Chicago, and St. Louis, while in Fort Worth and New York the suburban
schools in GAO's study almost always spent more per pupil than the inner
city schools. In Denver and Oakland, spending differences between the
selected inner city and suburban schools were mixed. In general, higher
per-pupil expenditures at any given school were explained primarily by
higher staff salaries regardless of whether the school was an inner city
or suburban school. Two other explanatory factors were student-teacher
ratios and ratios of students to student support staff, such as guidance
counselors, nurses, and librarians. Federal funds are generally targeted
to low-income areas to compensate for additional challenges faced by
schools in those areas. In some cases, the infusion of federal funds
balanced differences in per-pupil expenditures between the selected inner
city and suburban schools.
There is a broad consensus that poverty itself adversely affects academic
achievement, and inner city students in the schools reviewed performed
less well academically than students in the suburban schools. The
disparity in achievement may also be related to several other differences
identified in the characteristics of inner city and suburban schools. At
the schools GAO visited, inner city schools generally had higher
percentages of first-year teachers, higher enrollments, fewer library
resources, and less in-school parental involvement--characteristics that
some research has shown are related to school achievement.
9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0
Dollars per pupil
High Middle Low High Middle Low
Chicago New York
Inner city school Suburban school
www.gao.gov/cgi-bin/getrpt?GAO-03-234
To view the full report, including the scope and methodology, click on the
link above. For more information, contact Marnie Shaul at (202) 512-7215
or [email protected]
Contents
Letter Results in Brief Background Spending Differences between 1 3
Selected Inner City and Suburban Schools Varied by 4 8
Metropolitan Area Inner City Schools Generally Faced 17
Greater Challenges That May Have Affected Student 27
Achievement Conclusions Agency Comments 28
Appendix I Objectives, Scope, and Methodology 29
Appendix II School Profiles 40
Appendix III GAO Contacts and Staff Acknowledgments GAO Contacts 47
Acknowledgments 47
47
Tables Table 1: Total Enrollment and Percentages of Children in
Poverty, Students with Disabilities, and Students with
Limited English Proficiency for Selected Schools in the
Seven Metropolitan Areas Reviewed Table 2: Spending Per
Pupil, Average Teacher Salary, Student-Teacher Ratio,
and Student-Support Staff Ratio at the Median Spending
School in Each Metropolitan Area Table 3: Per-Pupil
Spending with and without Federal Dollars for Selected
Inner City and Suburban Schools in Seven Metropolitan
Areas Table 4: Metropolitan Areas Selected for Study
Table 5: Selected Inner City Census Tracts and Child 6 13
Poverty Rates Table 6: Selected Suburban School 17
Districts' Child Poverty Rates Table 7: School 30
Characteristics, Assessment Measure, and Measurement 33
Description Table 8: Regression Results for Factors 34
Explaining Differences in Per-Pupil Spending at Selected 36
Schools 39
Table 9: School-Level Student Characteristics for Selected Schools in
Seven Metropolitan Areas 41
Table 10: Spending Per Pupil and Spending Per Pupil at Low, Medium, and
High Weights for Selected Schools in Seven Metropolitan Areas 43
Table 11: Percent of First-Year Teachers, Federal Dollars Per Pupil, and
Federal Dollars as a Percent of Total Spending at Selected Schools in
Seven Metropolitan Areas 45
Figures
Figure 1: Paired Comparison (High to High, Middle to Middle, and
Low to Low) of Per-Pupil Spending at Selected Inner City
and Suburban Schools in Metropolitan Areas Where Inner
City Schools Spent More than Suburban Schools 9
Figure 2: Paired Comparison (High to High, Middle to Middle,
and
Low to Low) of Per-Pupil Spending at Selected Inner City
and Suburban Schools in Metropolitan Areas Where
Suburban Schools Spent More than Inner City Schools 10
Figure 3: Paired Comparison (High to High, Middle to Middle,
and
Low to Low) of Per-Pupil Spending at Selected Inner City
and Suburban Schools in the Denver and Oakland
Metropolitan Areas 11
Figure 4: Spending Per Pupil by the Median Inner City and
Suburban School in Four Metropolitan Areas for Different
Weight Adjustments for Students' Needs 15
Figure 5: Average Student Achievement Scores for Selected
Schools in Fort Worth, New York, Oakland, and St. Louis 19
Figure 6: Percentage of First-Year Teachers by School and
Metropolitan Area 21
Figure 7: Student Enrollments at Selected Schools 23
Figure 8: Playgrounds of an Inner City School in St. Louis and
a
Neighboring Suburban School 24
Figure 9: Number of Library Books per 100 Students at Selected
Schools 26
Abbreviations
EdTrust Education Trust
NAEP National Assessment of Educational Progress
SMSA metropolitan statistical area
United States General Accounting Office Washington, DC 20548
December 9, 2002
The Honorable Charles Rangel Ranking Minority Member Committee on Ways and
Means House of Representatives
Dear Mr. Rangel:
The No Child Left Behind Act of 2001 has focused national attention on the
importance of ensuring each child's access to equal educational
opportunity. The law seeks to improve the performance of schools and the
academic achievement of students, including those who are economically
disadvantaged. The heightened challenge of meeting the act's new
accountability requirements underscores the necessity of ensuring that all
schools have the support they need to provide students with a quality
public education. The challenge is particularly great for inner city
schools serving low-income neighborhoods, where students on average
continue to perform below students in suburban areas. The Congress and
other policymakers have been concerned that this achievement gap may be
related to possible differences in the amount of funding and resources
available to low-income schools and school districts and affluent schools
and school districts. Research has shown that such funding gaps are common
at the district level; for example, a recent study by The Education Trust
found that in 30 of the 47 states studied, school districts with the
greatest numbers of poor children had less money to spend per student than
districts with the fewest poor children.1 However, little research has
been done at the school level.
To provide you with information about inner city school spending and other
school characteristics, we determined similarities and differences between
selected inner city and suburban schools in (1) per-pupil spending and (2)
other factors that may relate to student achievement, such as teacher
experience, school enrollment, educational facilities and materials, and
types of parental involvement.
1
The Education Trust, The Funding Gap: Low-Income and Minority Students
Receive Fewer Dollars, August 2002.
Page 1 GAO-03-234 Inner City Schools
This study focuses on differences between inner city and suburban schools,
and as such is distinct from a study of differences between urban and
suburban schools; inner city schools, as a subset of urban schools, are in
the central core of the city and have higher poverty rates. We selected 42
schools, 21 inner city and 21 suburban public elementary schools, to
gather information on (1) school level, per-pupil spending, and federal
revenues and (2) school, teacher, other staff, and student characteristics
for the 2000-01 school year. We analyzed data from three inner city and
three suburban schools from each of seven different metropolitan areas of
medium, large, and very large population sizes: Oakland and St. Louis
(medium); Boston, Denver, and Fort Worth (large); and New York and Chicago
(very large).2 In analyzing these data, we applied weights-a technique
that allowed us to make adjustments to account for varying compositions of
student need. We applied three different levels of weights.
To obtain a selection of "typical" schools, we chose the inner city
schools in each metropolitan area that were at the median for poverty
among the inner city schools; similarly we chose the school districts at
the median for poverty among the suburban school districts. We attempted
also to include one high-performing inner city school in each metropolitan
area we visited, but were able to identify only two high-performing inner
city schools-1 in St. Louis and 1 in Oakland. For this selection, we used
The Education Trust database, which includes high-performing schools in
low-income areas. We did not include high-performing schools that were
special schools (e.g., magnet schools, science academies, etc.)3
In addition, we visited 24 of the 42 selected schools in the New York, St.
Louis, Fort Worth, and Oakland areas. We visited these schools to obtain
supplementary information on student achievement, the condition of the
buildings and facilities, educational materials, and parental involvement.
We analyzed similarities and differences separately for each geographic
area and for all seven sites collectively. Our results are not
generalizable beyond the schools in these seven sites. We conducted our
2
Inner city schools and suburban schools in Miami were part of the original
selection process but were dropped from the study because the district did
not provide the necessary data.
3
The criteria for including a school from the Education Trust database
included the following: (1) The school was located in a selected inner
city area. (2) The census child poverty rate for the school exceeded 40
percent. (3) The school placed in the top 50th percentile among all
schools on the state's most recent reading assessment test. (4) The school
was not a special school, for example, magnet school, science academy,
etc.
Results in Brief
work from January to November 2002 in accordance with generally accepted
government auditing standards. (A detailed explanation of our methodology
is found in app. I.)
Among the schools we reviewed, differences in per-pupil spending between
inner city and suburban schools varied by metropolitan area, with inner
city schools spending more in some areas and suburban schools spending
more in others. In Boston, Chicago, and St. Louis, the selected inner city
schools generally outspent suburban schools on a per-pupil basis. In Fort
Worth and New York, the suburban schools in our study generally spent more
per pupil than the selected inner city schools. In Denver and Oakland,
spending differences between inner city and suburban schools were mixed.
In general, higher per-pupil expenditures at any given school were
explained primarily by higher staff salaries regardless of whether the
school was an inner city or suburban schools. Two other important factors
included lower student-teacher ratios and lower ratios of students to
student support staff, such as guidance counselors, nurses, and
librarians. While the selected inner city schools in Boston, Chicago, and
St. Louis generally spent more per pupil than neighboring suburban
schools, when we made adjustments using the highest weights the suburban
schools generally spent more in every metropolitan area reviewed, because
inner city schools had higher percentages of low-income students. Some
research has shown that children from low-income families may require
extra resources to perform at the same levels as their nonpoor peers. To
address the additional needs of some children in low-income areas, federal
education programs target funds to schools in these areas. In some cases,
the infusion of federal funds has balanced differences in per-pupil
expenditures between selected inner city and suburban schools.
Inner city students in the schools we reviewed generally performed poorly
in comparison to students in suburban schools, a disparity that may be
related to several differences we identified in the characteristics of
inner city and suburban schools. Although research results are
inconclusive on the importance of various factors, some studies have shown
that greater teacher experience, smaller class size, more library and
computer resources, and higher levels of parental involvement are
positively related to student achievement. The inner city schools we
visited generally had higher percentages of first-year teachers, higher
enrollments, fewer library and computer resources, and less in-school
parental involvement. For example, first-year teachers comprised more than
10 percent of the teaching staff in 8 of the 12 inner city schools
visited, but the same was
Background
true in just 4 of 12 suburban schools. In New York City, the selected
inner city schools had fewer than 1,000 library books per 100 students,
whereas the selected suburban schools had more than 2,000 library books
per 100 students.
The Congress, among others, has been concerned about the academic
achievement gap between economically disadvantaged students and their more
advantaged peers. The disparity between poor students' performance on
standardized tests and the performance of their nonpoor peers is well
documented, and there is broad consensus that poverty itself adversely
affects academic achievement. For example, on the National Assessment of
Educational Progress (NAEP) reading assessment, 14 percent of fourth grade
students who qualified for the free and reduced lunch program (a measure
of poverty)4 performed at or above the proficient level in comparison to
41 percent of those students who did not qualify for the program.5
Furthermore, research has indicated the importance of socioeconomic status
as a predictor of student achievement.6 Research has shown that the
achievement gap falls along urban and nonurban lines as well: students
living in high-poverty, urban areas are even more likely than other poor
students to fall below basic performance levels.7
In addition to the achievement gap between poor and nonpoor students,
concerns exist that this gap may be related to differences between
per-pupil spending among schools that serve poor and nonpoor communities.
School district spending is generally related to wealth and tax levels,
and differences in school district spending can have an impact on spending
at
4
Eligibility for free lunches is set at 130 percent of the official poverty
line ($22,165 for a family of four during the 2000-01 school year), and
eligibility for reduced-price lunches extends up to 185 percent of the
poverty line ($31,543 for a family of four during the 2000-01 school
year).
5
U.S. Department of Education, National Center for Education Statistics,
NAEP, The Nation's Report Card: Fourth-Grade Reading 2000, April 2001.
6
See for example, U.S. General Accounting Office, School Finance: State and
Federal Efforts to Target Poor Students, GAO/HEHS-98-36 (Washington, D.C.:
Jan. 28, 1998).
7
Educational Testing Service, unpublished tabulations from 1994 NAEP
reading test. Cited in Education Week "Quality Counts," 1998.
http://www.edweek.org/sreports/qc98/challenges/achieve/ac-c1.htm.
the school level.8 Recently, efforts have been made to achieve greater
spending equity. Using a variety of approaches, a number of states have
targeted some additional funding to poor students to amend the unequal
abilities of local districts to raise revenues for public schools.9
Comparing spending between schools in simple dollar terms provides one way
to check for differences; however, this type of straightforward comparison
may be insufficient to explain spending differences because it does not
capture the higher cost of educating students with special needs. Schools
with similar spending per pupil may actually be at a comparative
disadvantage when adjustments are made to account for differing
compositions of student needs. Though not definitive, some research shows
that children with special needs-low-income students, students with
disabilities, and students with limited English proficiency-may require
additional educational resources to succeed at the level of their
nondisadvantaged peers. Because these additional resources require higher
spending, some researchers have adjusted per-pupil expenditures by
"weighting" these students to account for the additional spending they may
be required.10
Weighting counts each student with special needs as more than one student,
so that the denominator in the expenditures to students ratio is
increased, causing the weighted per-pupil expenditure figure to decrease
accordingly. For example, a school with an enrollment of 100 students may
have 20 low-income students, 20 students with disabilities, and 10
students with limited English proficiency. Weighting these three groups of
special needs students twice as heavily as other students causes weighted
enrollment to rise to 150 students. If spending per-pupil is $4,000
without weighting, it drops to $2,667 when weights are applied. The actual
size of the weights assigned to low-income students, students
8
For recent statistics of finance equity among states, see American
Education Finance Association, Equitable School Finance Systems: Grading
The States, American Education Finance Association meeting, Austin, TX,
Mar. 9-11, 2000.
9
Spending per pupil reported in this study reflects nominal dollars after
such adjustments have been made by the state to account for student needs.
10
See: S. Chaikind, et al., "What Do We Know About the Costs of Special
Education? A Selected Review," The Journal of Special Education, 26, no. 4
(1993): 344-370; American Institutes for Research, What Are We Spending on
Special Education Services in the United States, 1999-2000?, Advance
Report No. 1.( Special Education Expenditure Project, Mar. 2002.);
GAO/HEHS-98-36; T. Parrish, "A Cost Analysis of Alternative Instructional
Models for Limited English Proficient Students in California," Journal of
Education Finance (Winter 1994): 256-278.
with disabilities, and students with limited English proficiency is
subject to debate and generally ranges between 1.2 and 2.0 for low-income
students, between 1.9 and 2.3 for students with disabilities, and between
1.10 and 1.9 for students with limited English proficiency.11
The inner city schools selected for our study had high proportions of
children in poverty in comparison to the selected suburban schools. The
elected inner city schools also generally had more students with limited
English proficiency than their suburban counterparts. However, the
proportions of students with disabilities in our selected inner city and
suburban schools differed within and among metropolitan areas. In Denver,
the selected inner city schools consistently had a higher proportion of
students with disabilities than the selected suburban schools while in
Fort Worth, the suburban schools had a higher proportion of students with
disabilities. (See table 1 for total enrollment and percentages of
children in poverty, students with disabilities, and students with limited
English proficiency for selected schools in the seven metropolitan areas
reviewed in this study.)
Table 1: Total Enrollment and Percentages of Children in Poverty, Students
with Disabilities, and Students with Limited English Proficiency for
Selected Schools in the Seven Metropolitan Areas Reviewed
Metropolitan area Inner city/suburb Enrollment Percent poor Percent disabled
Percent LEP
Boston Inner city 1 712 51 21 0
Inner city 2 193 50 9 0
Inner city 3 250 49 17 12
Suburban 1 386 7 12 0
Suburban 2 979 7 15 0
Suburban 3 335 7 8 3
Chicago Inner city 1 466 59 9 0
Inner city 2 900 59 14 5
Inner city 3 692 59 12 0
Suburban 1 503 5 17 1
Suburban 2 401 5 8 2
Suburban 3 280 5 6 5
Denver Inner city 1 562 52 12 52
Inner city 2 372 52 13 19
Inner city 3 468 51 12 32
Suburban 1 407 9 13 0
11Ibid.
Metropolitan Inner city / Enrollment Percent Percent Percent
area suburb poor disabled LEP
Suburban 2 292 10 8 11
Suburban 3 623 11 6 10
Fort Worth Inner city 1 760 52 6 17
Inner city 2 555 51 3 10
Inner city 3 937 51 3 15
Suburban 1 413 12 18 2
Suburban 2 392 12 6 5
Suburban 3 373 14 17 13
New York Inner city 1 484 56 9 22
Inner city 2 645 52 11 18
Inner city 3 630 43 6 3
Suburban 1 457 5 16 9
Suburban 2 553 5 9 3
Suburban 3 536 5 9 0
Oakland Inner city 1 745 45 5 64
Inner city 2 312 50 9 73
Inner city 3 1,238 47 6 41
Suburban 1 402 8 8 15
Suburban 2 877 8 0 4
Suburban 3 460 8 8 3
St. Louis Inner city 1 163 85 12 0
Inner city 2 292 55 13 0
Inner city 3 499 55 8 0
Suburban 1 602 11 18 3
Suburban 2 391 11 5 0
Suburban 3 459 9 11 1
Source: GAO's data analysis.
Differences in school spending can affect characteristics that may be
related to student achievement. There is a large body of research on
factors that may directly or indirectly contribute to student achievement.
Spending has been the factor most studied for its effect on student
achievement. Differences in student outcomes have also been related to
factors such as teacher quality, class size, quality of educational
materials, and parental involvement. Our study describes how some of these
factors may differ across selected inner city and suburban schools.
Spending Differences between Selected Inner City and Suburban Schools Varied
by Metropolitan Area
Differences in per-pupil spending between selected inner city and suburban
schools varied by metropolitan areas in our study.12 Inner city schools in
Boston, Chicago, and St. Louis generally spent more per pupil than
neighboring suburban schools, whereas selected suburban schools in Fort
Worth and New York almost always spent more per pupil than the inner city
schools. In Denver and Oakland, no clear pattern of spending emerged.
Three factors generally explained spending differences between inner city
and suburban schools: (1) average teacher salaries; (2) student-teacher
ratios; and (3) ratios of students to student support staff, such as
guidance counselors, librarians, and nurses. When we adjusted per-pupil
expenditures to account for the extra resources students facing poverty,
disabilities, and limited English proficiency might need, inner city
schools almost always spent less per pupil than suburban schools. To
compensate for additional challenges faced by schools in these areas,
federal education dollars are generally targeted to low-income areas. As a
result, federal funds have played an important role in increasing funding
to inner city schools.
Differences in Per-Pupil Spending between Selected Inner City Schools and
Suburban Schools Varied by Metropolitan Area
Differences between inner city and suburban school per-pupil spending were
related to the particular metropolitan area studied and generally seemed
to be most influenced by teacher salaries. The selected inner city schools
tended to outspend the suburban schools in the Boston, Chicago, and St.
Louis metropolitan areas.13 For example, in the Boston metropolitan area,
the lowest spending inner city school spent more per pupil than the
highest spending suburban school. (See fig. 1 for a comparison of
per-pupil spending at selected inner city and suburban schools in these
areas.)
12
We gathered operational school-level spending on personnel salaries,
building maintenance and repair, and educational materials and supplies.
Other operational expenditures, for example transportation and capital
expenditures, are not considered spending for this report. Total spending,
as used herein, refers to the total amount spent on salaries, building
maintenance and repair, and educational materials and supplies.
13
See appendix I for technical details.
Page 8 GAO-03-234 Inner City Schools
Figure 1: Paired Comparison (High to High, Middle to Middle, and Low to
Low) of Per-Pupil Spending at Selected Inner City and Suburban Schools in
Metropolitan Areas Where Inner City Schools Spent More than Suburban
Schools
9,000 Dollars per pupil
8,000
7,000 6,000 5,000 4,000 3,000 2,000 1,000 High Middle Low High Middle Low
High Middle Low Boston Chicago St. Louis
Inner city school Suburban school High-performing inner city school
Source: GAO's data analysis.
In contrast, in the Fort Worth and New York metropolitan areas, suburban
schools generally outspent inner city schools. For example, among the
selected schools in the Fort Worth metropolitan area, the lowest spending
suburban school had per-pupil expenditures 21 percent higher than the
highest spending inner city school. (See fig. 2 for a comparison of
per-pupil spending at selected inner city and suburban schools in these
areas.)
Figure 2: Paired Comparison (High to High, Middle to Middle, and Low to
Low) of Per-Pupil Spending at Selected Inner City and Suburban Schools in
Metropolitan Areas Where Suburban Schools Spent More than Inner City
Schools
9,000 Dollars per pupil
8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 High Middle Low High
Middle Low Fort Worth New York
Inner city school Suburban school
Source: GAO's data analysis.
In Denver and Oakland, an examination of spending differences among the
selected suburban and inner city schools revealed mixed results. That is,
analysis of spending differences showed no general pattern of spending
that favored either inner city or suburban schools. (See fig. 3 for a
comparison of per-pupil spending at selected inner city and suburban
schools in the Denver and Oakland metropolitan areas.)
Figure 3: Paired Comparison (High to High, Middle to Middle, and Low to
Low) of Per-Pupil Spending at Selected Inner City and Suburban Schools in
the Denver and Oakland Metropolitan Areas
9,000 Dollars per pupil8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000
High Middle Low High Middle Low Denver Qakland
High-performing inner city school Source: GAO's data analysis.
Inner city school Suburban school
Average Teacher Salaries, Student-Teacher Ratios, and Ratios of Students
to Student Support Staff Accounted for Most of the Differences in School
Spending in Selected Schools
Among the schools in our study, three factors influenced per-pupil
spending: average teacher salaries, student-teacher ratios, and the ratio
of students to student support staff.14 Average teacher salaries appeared
to have the greatest impact on per-pupil spending, followed by lower
student-teacher ratios and lower ratios of students to student support
staff.
Average teacher salaries influenced per-pupil spending in areas where
inner city schools spent more per pupil (Boston and Chicago), where
suburban schools spent more per pupil (New York), and where spending was
mixed (Oakland). For example, in Chicago, where inner city schools
generally outspent suburban schools, the median inner city school average
teacher salary was $47,851, compared with $39,852 in the suburbs. In
Oakland, where spending between suburban schools and inner city schools
was mixed, the average teacher salary at the median spending school was
$60,395 and per-pupil spending was $4,849, compared with $52,440 and
$4,022 at the median spending inner city school.
Student-teacher ratios and ratios of students to student support staff
were factors that could offset the influence of teacher salaries in
explaining per-pupil spending.15 For example, in Fort Worth, where the
three suburban schools typically spent more per student than inner city
schools, inner city teacher salaries were generally higher than suburban
teacher salaries. However, ratios of students to both teachers and student
support staff were lower in our selected suburban schools. For example,
the median spending inner city school in Fort Worth had 21 students per
teacher, compared with 17 students per teacher in the suburbs.
Additionally, the median spending inner city school had 1 student support
staff professional for every 162 students, whereas in the suburbs the
ratio was 1 to 68. (Table 2 lists factors contributing to higher per-pupil
spending-average teacher salaries, student-teacher ratios, and ratios of
students to support staff-for the median spending school in each reviewed
metropolitan area.)
14
Regression analysis was employed to identify factors influencing per-pupil
spending. The t-scores of average teacher salary, student-teacher ratio,
and the ratio of students-to-student support staff were found to be
significant at the 0.05 level. Maintenance and repair spending was found
to be positively related to per-pupil spending, but not at the 0.05 level.
See appendix I for technical details.
15
Student support staff was defined as including guidance counselors, social
workers, psychologists, librarians, nurses, speech therapists, principals,
and assistant principals.
Inner City Schools at a Disadvantage When Spending Adjusted for Student
Needs
Table 2: Spending Per Pupil, Average Teacher Salary, Student-Teacher
Ratio, and Student-Support Staff Ratio at the Median Spending School in
Each Metropolitan Area
Average Student-
Spending teacher teacher Students-student
per pupil salary ratio support staff ratio
Boston
Inner city $5,770 $61,079 16:1 119:1
Suburb $4,433 $38,180 17:1 61:1
Chicago
Inner city $4,482 $46,661 23:1 58:1
Suburb $3,216 $39,852 21:1 100:1
Denver
Inner city $3,852 $38,044 20:1 171:1
Suburb $3,313 $32,753 17:1 86:1
Fort Worth
Inner city $3,058 $41,402 21:1 162:1
Suburb $4,246 $33,316 17:1 68:1
New York
Inner city $6,057 $42,285 a a
Suburb $7,218 $72,591 18:1 73:1
Oakland
Inner city $4,022 $52,440 30:1 233:1
Suburb $4,849 $60,395 20:1 155:1
St. Louis
Inner city $5,337 $33,223 25:1 28:1
Suburb $3,467 $34,304 13:1 87:1
Note: School districts in New York City did not provide us with
information on student-teacher ratios and the ratio of students to student
support staff.
a
Not applicable. Source: GAO's data analysis.
Despite higher per-pupil spending by about half of the inner city schools
in our study, inner city schools generally spent less compared with
neighboring suburban schools when spending was weighted to account for
differing compositions of student needs. To account for the greater costs
that may be associated with educating low-income students, students with
disabilities, and students with limited English proficiency, some
researchers have used formulas that weight these students more heavily
than other students. In a similar fashion, we applied weights to our
per-pupil expenditure data.
The use of the lowest and medium weights had little impact on spending
differences between inner city and suburban schools.16 Inner city schools
in Boston, Chicago, and St. Louis continued to outspend neighboring
suburban schools in most cases. For example, in Chicago, when students
were weighted with the lowest weight, the median per-pupil spending for
inner city school was $3,743 per pupil compared with $2,996 for the
suburban school. Similarly, the use of medium weights generally did not
result in higher per-pupil spending at suburban schools. For example,
using medium weights, the median inner city school in Chicago still spent
more than the median suburban school, although the difference was
smaller-$3,089 compared with $2,858.
However, when the highest weight was applied, inner city per-pupil
spending fell below suburban school spending in almost all cases.17 For
example, in Chicago when the highest weight was applied, per-pupil
spending at the median inner city school was less than that of the
suburban school, $2,629 as compared with $2,734. Similarly, in the New
York metropolitan area, where suburban schools we reviewed outspent inner
city schools, the use of the highest weights to adjust for student needs
caused the differences between inner city and suburban school spending to
be substantially enlarged. (See fig. 4 for examples of how spending
changes as different weights are applied for per-pupil spending at the
median inner city and suburban schools in four metropolitan areas.)
16
The actual size of the weights assigned to low-income, special education,
and limited English proficiency students is subject to debate and
generally ranges from a 1.2 to 2.0 for low-income students, from 1.9 to
2.3 for special education students, and from 1.1 to 1.9 for students with
limited English proficiency. Consequently, low-weights were 1.2 for
low-income students, 1.9 for special education students, and 1.1 for
students with limited English proficiency. The medium weights were 1.6 for
low-income students, 2.1 for special education students, and 1.5 for
students with limited English proficiency. The high weights were 2.0 for
low-income students, 2.3 for special education students, and 1.9 for
students with limited English proficiency.
17
The differences between inner city and suburban weighted per-pupil
spending was most affected by differences in the proportion of low-income
students in inner city and suburban schools. The inner city schools in our
study served populations with very high proportions of low-income
students.
Figure 4: Spending Per Pupil by the Median Inner City and Suburban School in
Four Metropolitan Areas for Different Weight
Adjustments for Students' Needs
Chicago St. Louis
Spending per pupil (median spending 8,000 Spending per pupil (median
school)8,000 spending school)
7,000 7,000
6,000 6,000
1,000 1,000
0 0
No Low Medium High No Low Medium High
Weighting Weighting
Inner city weight
Suburban weight
Source: GAO's data
analysis.
Federal Funds Played Important Role in Helping Inner City Schools Meet
Expenses
Because federal programs, such as Title I, specifically target funds to
schools in low-income areas, these federal funds generally helped reduce
or eliminate the gap between selected inner city and suburban schools in
terms of per-pupil expenditures.18 In the Denver and St. Louis
metropolitan areas, federal funds generally eliminated the gap between
inner city and suburban schools' per-pupil spending. In Fort Worth,
without federal funds per-pupil spending at the selected inner city
schools would have been about 63 percent of selected suburban schools, and
in Oakland, per-pupil spending would have been about 78 percent of
suburban schools. However, selected inner city schools in Boston and
Chicago would have still spent more than suburban schools without federal
funds. (See table 3 for a comparison of inner city and suburban per child
spending with and without federal dollars.)
18
See U.S. General Accounting Office, Title I Funding: Poor Children Benefit
Though Funding Per Poor Child Differs, GAO-02-242 (Washington, D.C.: Jan.
31, 2002).
Page 16 GAO-03-234 Inner City Schools
Inner City Schools Generally Faced Greater Challenges That May Have Affected
Student Achievement
Table 3: Per-Pupil Spending with and without Federal Dollars for Selected
Inner City and Suburban Schools in Seven Metropolitan Areas
Per-pupil spending with Per-pupil spending federal dollars without federal
dollars
Inner city Suburb Metropolitan area Inner city Suburb
$6,731 $4,737 Boston $6,283 $4,620
$5,770 $4,433 $5,191 $4,264
$4,998 $4,277 $4,669 $4,147
$4,529 $3,975 Chicago $4,049 $3,747
$4,482 $3,216 $3,859 $3,145
$3,595 $3,053 $2,936 $2,790
$3,865 $5,404 Denver $3,437 $5,275
$3,852 $3,313 $3,136 $3,203
$3,240 $2,810 $2,567 $2,717
$2,984 $4,391 Fort Worth $2,474 $4,202
$3,058 $4,040 $2,472 $3,727
$3,282 $4,246 $2,316 $3,655
$7,263 $8,637 New Yorka $6,318 $8,457
$5,904 $7,218 $5,236 $7,204
$6,566 $6,298
$4,022 $4,849 Oakland $3,567 $4,840
$4,100 $4,975 $3,355 $4,796
$2,810 $2,470 $2,517 $2,453
$5,337 $4,395 St. Louis $3,844 $4,202
$6,311 $3,467 $3,730 $3,361
$3,162 $3,106 $2,287 $3,000
aSchool level federal dollars received for only two inner city schools in
New York. Source: GAO's data analysis.
Factors that may relate to student achievement differed between inner city
and suburban schools in our study. Research has shown a positive
relationship between student achievement and factors such as teacher
experience, lower enrollment, more library books and computer resources,
and higher levels of parental involvement. Among the 24 schools we
visited, the average student achievement scores were generally lower in
inner city than in suburban schools. Along with lower achievement scores,
these inner city schools were more likely to have a higher percentage of
first-year teachers, whose lack of experience can be an indicator of lower
teacher quality. In addition, in comparison to the suburban schools, inner
city schools generally were older, had higher student enrollments, and had
fewer library books per pupil and less technological support. Finally, the
type of in-school parental involvement in the inner city and suburban
schools differed.
Inner City Students' Achievement Scores Were Generally Lower than Suburban
Students' Achievement Scores
In general, at the schools we visited in the metropolitan areas of Fort
Worth, New York, Oakland, and St. Louis, inner city students' average
achievement scores on state reading assessment tests were lower than
scores at the neighboring suburban schools. Two schools were exceptions to
this pattern. In St. Louis, we specially selected one high-performing
inner city school; students at this school performed higher than students
at the three suburban schools we visited. In the Fort Worth metropolitan
area, one inner city school performed slightly higher than two of the
three suburban schools we visited. (See fig. 5 for average student
achievement scores for selected schools in the four metropolitan areas.)
Figure 5: Average Student Achievement Scores for Selected Schools in Fort Worth,
New York, Oakland, and St. Louis
Fort Worth New York
90 Average student achievement scores 750 Average student achievement
scores
State average: 82.6
700
State average: 653
80 650
600 70
550
60
500 Inner city Suburban Inner city Suburban
Oakland St. Louis
700 Average student achievement scores 800 Average student achievement
scores
750
State average 635.9
650
700
State average: 641.9
650
600
600
550
550
500 Inner city Suburban Inner city Suburban 500
High-performing inner city school Source: GAO's data analysis.
Inner City Schools We Reviewed Had More First-Year, Thus Less Experienced,
Teachers than Suburban Schools
Although the selected inner city schools' student achievement scores were
generally lower, this pattern did not appear to be related to or
consistent with per-pupil spending. That is, higher-performing schools
were not necessarily schools that were high in per-pupil spending. For
example, per-pupil spending at the highest-performing inner city school in
Fort Worth we visited was $3,058, which was higher than one selected inner
city school, lower than the other selected inner city school, and lower
than each of the suburban schools.
First-year teachers in the 24 schools we visited generally constituted a
higher percentage of the faculty in inner city schools than suburban
schools.19 First-year teachers comprised more than 10 percent of the
teaching staff in 8 of the 12 inner city schools, but the same was true in
just 4 of 12 suburban schools. However, both the percent of first-year
teachers and differences between inner city and suburban schools varied
among the 4 metropolitan areas. (See fig. 6 for the percentage of
first-year teachers by school and metropolitan area.) For example, in the
New York metropolitan area there were no first-year teachers at 2 of the
suburban schools, but at 2 inner city schools first-year teachers were 24
and 13 percent of the faculty.20 In the Fort Worth metropolitan area, 2 of
the suburban schools had almost twice the percent of first-year teachers
as the two inner city schools with the highest percent of first-year
teachers.
19
The percentage of first-year year teachers can be used as an indicator of
lower teacher quality because of their relative inexperience.
20
Information on first-year teachers was received for only 2 of the 3
selected New York City schools.
Page 20 GAO-03-234 Inner City Schools
Figure 6: Percentage of First-Year Teachers by School and Metropolitan
Area
25 Percentage of first-year teachers
Fort Worth New York Oakland St. Louis
Inner city school Suburban school High-performing inner city school
Indicates 0% of first-year teachers at suburban school
Indicates 0% of first-year teachers at high-performing inner city school
Source: GAO's data analysis.
Note: One New York school did not provide data on first-year teachers.
Notably, the percentage of first-year teachers was low at the two
high-performing inner city schools. In Oakland, the percentage of
first-year teachers at the high-performing inner city school was 6
percent, compared with 12 percent at the other two inner city schools. In
St. Louis, the high-performing inner city school had no first-year
teachers, whereas the other two inner city schools had 11 and 16 percent.
As noted earlier in the report, average teacher salaries in large part
accounted for most of the differences in school spending. The fact that
teaching staff at inner city schools were generally comprised of higher
percentages of first-year teachers is not inconsistent with the finding on
teacher salaries. The average teacher salary at a school includes the
salaries of all teachers in the school, from first-year teachers to the
most senior staff. For example, in a school with a high proportion of
first-year teachers the average teacher salary could still be higher than
that of
Enrollment Was Higher in Inner City Schools than in Suburban Schools, and
Buildings Were Older
another school because of higher proportions of tenured teachers and the
district's salary structure.
The enrollment of the 12 inner city schools we visited tended to be higher
than that of the 12 suburban schools we visited, but enrollment varied
across and within metropolitan areas.21 The national average elementary
school enrollment is 443, and schools with enrollments over 600 are
considered "large," regardless of the school's capacity.22 In three out of
the four metropolitan areas we visited, Fort Worth, New York, and Oakland,
the enrollment at the inner city schools was consistently higher than the
national average enrollment. In addition, 6 of the 12 inner city schools
we visited had enrollments over 600 students. In contrast, enrollments
exceeded 600 in only 2 of the 12 suburban schools we visited. (See fig. 7
for enrollments at the selected schools.)
21
WestEd reports that research indicates smaller schools can reduce the
effects of poverty on student achievement. See, WestEd, Are Small Schools
Better? School Size Safety & Learning, November 21, 2001, San Francisco,
CA.
22
National Center for Education Statistics, Condition of American Public
School Facilities, 1999.
Page 22 GAO-03-234 Inner City Schools
Figure 7: Student Enrollments at Selected Schools
1,400 Students enrolled
Fort Worth New York Oakland St. Louis
National average: 443 students enrolled
Inner city school Suburban school High-performing inner city school
Source: GAO's data analysis.
Among the schools we visited, most of the inner city schools were older
than 50 years, which is higher than the national average of 43 years.
Furthermore, 7 of the oldest 10 buildings were inner city schools, 2
having been built in the 19th century. In contrast, most of the suburban
schools we visited were less than 40 years old.
In addition to the physical condition of the buildings, playground
facilities in the inner city schools differed greatly from facilities in
the suburban schools. Inner city schools we visited were less likely to
have playground equipment and expansive play areas. For example, the
playgrounds in St. Louis suburban schools all had green fields and a
variety of playground equipment. In this same metropolitan area, only one
of the inner city schools had any playground equipment and at the other
two schools asphalt lots were the single outdoor recreational facility.
Figure 8 shows the playgrounds of an inner city school and a suburban
school in the St. Louis metropolitan area.
Figure 8: Playgrounds of an Inner City School in St. Louis and a
Neighboring Suburban School
Inner City Schools Had Less Library and Technological Support than
Suburban Schools
Overall, the inner city schools we visited had fewer library books per
child and were less likely to have a computer laboratory than suburban
schools. Most of the suburban schools visited were below the national
average of 2,585 books per 100 students-7 of the 12 schools had more than
2,000 books per 100 students. However, only 3 of the inner city schools
visited had more than 2,000 books per 100 students.23 For example, in New
York City, the 3 selected inner city schools had fewer than 1,000 library
books per 100 students, whereas the 3 selected suburban schools had more
than 2,000 library books per 100 students and one had more than 3,000.
Notably, the high-performing inner city school in St. Louis had 2,813
library books per 100 students, more than any of the suburban schools we
visited in that area. Similarly, the high-performing inner city school in
Oakland had 2,244 books per 100 students, which was more than the other
two Oakland inner city schools and 2 of the 3 selected suburban schools.
Furthermore, only 7 of the 12 selected inner city schools had a full-time
librarian, whereas all but one suburban school had a full-time librarian.
(See fig. 9 for the number of library books per 100 students at selected
schools.)
23
The number of library books reported is not precise. The reported number
is based upon data that include both counts provided by some schools and
best estimates provided by librarians from other schools.
Page 25 GAO-03-234 Inner City Schools
Figure 9: Number of Library Books per 100 Students at Selected Schools
4,000 Books per 100 students
Fort Worth New York Oakland St. Louis
National average: 2,585 books per 100 students
Inner city school Suburban school High-performing inner city school
Source: Prepared by GAO from school site visits.
Our site visits also revealed a difference between inner city and suburban
schools in terms of the presence of a computer laboratory. Eleven of the
12 suburban schools we visited had a computer laboratory, whereas 8 of the
inner city schools visited had such a facility. Among schools with
computer laboratories, however, the ratio of students to laboratory
computers was similar among inner city and suburban schools.
In-School Parental Involvement Differed between Selected Inner City and
Suburban Schools
Parents of children attending the suburban schools we visited were more
involved in on-site school activities than parents of inner city
children.24 According to the suburban school principals, parental
involvement in their schools was typically very high and included
participation in volunteer activities, attendance at parent-teacher
conferences, and providing financial support to the school. Parent
volunteerism at suburban schools could be quite substantial. For example,
parents at one suburban school in the Oakland metropolitan area provided
24,000 hours of volunteer time during the school year. Inner city
principals characterized parents as concerned and interested in their
children's education, though less likely to attend parent-teacher
conferences and volunteer in school. A number of inner city principals we
interviewed also noted that while parents generally wanted to help their
children succeed in school, they often lacked the necessary finances,
skills, or education to offer additional assistance beyond that offered by
the school.
Conclusions
Our findings suggest that spending differences between the inner city
schools and suburban schools in our review do exist, but these differences
for the most part depend upon the metropolitan area. In some metropolitan
areas, inner city schools spent more per pupil whereas in others suburban
schools spent more per pupil. Spending differences, regardless of
metropolitan area for the most part, seemed to be the result of
differences in salaries and student to teacher and staff ratios. However,
the very heavy concentration of poverty in inner city schools may place
them at a spending disadvantage, even when spending is equal. In addition,
the suburban schools, as well as the high-performing inner city schools we
visited, generally had more experienced teachers, lower enrollments, more
library books per child, and more parental in-school volunteer activities
than the other inner city schools in this study. These factors are
important to consider in improving the performance of inner city schools.
24
Research has indicated the importance of parental involvement to student
achievement. The National Conference of State Legislatures reported on a
comprehensive review of 66 studies that examined the correlation between
parent involvement and student success and concluded that parent
involvement, not income or social status, was the most accurate predictor
of student success. See National Conference of State Legislatures,
Improving Student Achievement, July 2001, citing Anne T. Henderson and
Nancy Berla, A New Generation of Evidence: The Family is Critical to
Student Achievement (Washington, D.C., Center for Law and Education,
1995).
Agency Comments
We provided a draft of this report to the Department of Education for
review and comment. Education's Executive Secretariat confirmed that
department officials had reviewed the draft and had no comments.
We are sending a copy of this report to the Secretary of Education. We
will make copies available to others upon request. In addition, the report
will be available at no charge on GAO's Web site at http:// www.gao.gov.
If you or your staff have any questions about this report, please call me
at (202) 512-7215. See appendix III for other staff acknowledgments.
Sincerely yours,
Marnie S. Shaul, Director Education, Workforce, and Income Security Issues
Appendix I: Objectives, Scope, and Methodology
The objectives of our study were to provide information on similarities
and differences between (1) per-pupil spending in selected inner city and
suburban schools and (2) other characteristics that may relate to student
achievement, such as, teacher experience, school enrollment, educational
materials, physical facilities, and parental involvement. To address the
first objective, we reviewed the literature on spending differences,
interviewed experts about the issues and approaches to measuring spending
data, and collected spending and related school data on 42 inner city and
suburban schools. To address the second objective, we examined the
literature, interviewed experts about relationships between student
achievement and school characteristics, and visited 24 inner city and
suburban schools to collect information on student achievement, the
quality and availability of educational materials, the condition of the
buildings and facilities, and type and extent of parental involvement.
This appendix discusses the scope of the study, criteria for selecting
metropolitan areas and schools, and the methods employed to describe and
explain observed spending differences.
Scope This study focused on similarities and differences between inner
city schools and suburban schools. This is different and distinct from a
study of similarities and differences between urban and suburban schools,
or urban and suburban districts, as urban schools and districts generally
include a wider range of poverty than inner city schools. This study
covered selected inner city and suburban schools in seven metropolitan
areas.
Metropolitan Area Selection Metropolitan areas were purposively selected
to reflect diversity on the basis of geography
and size. We used geographic areas from the
Northeast, Midwest, South, and West. Three
size categories were used: (1) very large, (2)
large, and (3) medium. We defined these by
population.
o Very large: areas where the central city of
a metropolitan area had a population of more
than 1 million residents;
o Large: areas where the central city of a
metropolitan area had a population between
500,000 and 1 million residents;
o Medium: areas where the central city of a
metropolitan area had a population between
250,000 and 500,000 residents.
Defining Inner City and Suburbs
The metropolitan areas selected for inclusion in the study were Boston,
Chicago, Denver, Fort Worth, Miami, New York, Oakland, and St. Louis.
Inner city and suburban schools in Miami were dropped from the study
because the district did not provide the necessary data. (See table 4 for
the selected metropolitan areas.)
Table 4: Metropolitan Areas Selected for Study
Category Metropolitan area Geographic location City population
Very large Chicago Midwest 2,896,016
New York Northeast 8,008,278
Large Boston Northeast 589,141
Denver West 554,636
Fort Worth South 534,694
Medium Miami South 362,470
Oakland West 399,484
St. Louis Midwest 348,189
Source: GAO's data analysis and 2000 Census.
For this study, in consultation with experts, we defined "inner city" as a
contiguous geographic area that (1) had a poverty rate of 40 percent or
higher, (2) was located within the "central core" of a city with a
population of at least 250,000 persons, and (3) the city is the central
city of a metropolitan with a population of at least 1 million persons.
We defined suburb as the geographic area that is (1) outside the
boundaries of a central city with a population of at least 250,000
persons,
(2) inside the boundaries of the metropolitan statistical area (SMSA) of
the central city, as defined by the Office of Management and Budget and
used by the census, and (3) the metropolitan area has a population of at
least 1 million persons.
In total, we collected spending data on 42 schools, 21 inner city and 21
suburban public elementary schools in seven metropolitan areas, and
gathered information on (1) school-level per-pupil spending and federal
revenues, and (2) school, teacher, other staff, and student
characteristics for the 2000-01 school year. In addition, we conducted
site visits at 24 of the selected schools. These schools were located in
the New York, St. Louis, Fort Worth, and Oakland metropolitan areas. We
visited them in order to obtain supplementary information on
characteristics that might affect student achievement, such as facilities,
educational materials, and types of parental involvement.
School Selection
The study was designed to compare "typical" inner city and "typical"
suburban schools, rather than those schools with extreme poverty or
wealth. We consulted with experts about our design. We used the factors
described below to select typical schools. Our goal was to make
comparisons that would reflect likely differences, if any, between the
inner city and suburban schools in a given metropolitan area.
To select the inner city schools, we (1) consulted with local experts in
each metropolitan area to identify the geographic area of the central city
of the SMSA generally considered the inner city, (2) calculated census
child poverty rates for each census tract within the inner city area,
(3)
retained identified census tracts with census child poverty rates
higher than 40 percent, (4) ranked the census tracts by poverty
rate, and
(5)
identified the three inner city census tracts closest to the 50th
percentile, that is, the median poverty census tracts of the inner
city.1 We then selected the public elementary school that served
those census tracts, but purposely excluded schools that were
special schools, for example, magnet schools, science academies,
etc.
Where possible, we attempted to include one high-performing inner city
school in each metropolitan area we visited. We used Dispelling the Myth,
an Education Trust (EdTrust) database of high-poverty, high-performing
schools, for this selection. Dispelling the Myth is an ongoing EdTrust
project to identify high-poverty and high-minority schools that have high
student performance or have made substantial improvement in student
achievement. We identified schools in that database with a student poverty
rate greater than 50 percent and an overall achievement score on the most
recent state reading assessment test above the 50th percentile. Because
the EdTrust database used free and reduced lunch eligibility as its
1
In two metropolitan areas, New York and Oakland, local experts identified
three distinct inner city areas. In these two metropolitan areas, the
census tract in each inner city area closest to median level poverty was
selected. Two of the three selected inner city schools in New York-schools
selected from Harlem and the Bronx-had poverty rates above 40 percent and
were located in inner city areas. These schools, however, were selected on
the basis of per capita income, which was the selection methodology
employed during the early design phase of the study and subsequently
replaced by the median poverty rate approach. The schools were retained,
however, for data efficiency purposes and because their child poverty
rates were consistent with that of schools that would have been selected
in their stead.
criterion for poverty,2 we further verified that the school was located in
an inner city census tract as defined by this study serving an area with a
census child poverty rate greater than 40 percent. We purposely excluded
schools that were special schools, for example, magnet schools, science
academies, etc. Inner city schools from the St. Louis and Oakland
metropolitan areas met these criteria. The identified high-performing
inner city school in St. Louis replaced a selected school. The identified
high-performing inner city school in Oakland, however, was a school that
would have been selected through the described census tract approach and
was, therefore, treated similarly to the other selected inner city
schools. (See table 5 for the selected inner city census tracts and child
poverty rates.)
Child poverty can be measured by (1) census data or (2) the number of
children eligible for free or reduced-price lunch. The subsidized lunch
program provides a looser definition of poverty than census poverty data.
The number of students eligible for subsidized lunches is roughly double
the number meeting the census poverty definition. Nonetheless, according
to the Department of Education, the subsidized lunch program provides the
best available source of data on low-income students at the school level.
Page 32 GAO-03-234 Inner City Schools
Table 5: Selected Inner City Census Tracts and Child Poverty Rates
Metropolitan area Census tract Child poverty rate
Boston 0611 48.5%
0814 49.8%
0924 50.9%
Chicago 6106 58.6%
6812 58.9%
4001 59.0%
Denver 0011.02 51.3%
0007.02 52.0%
0010 52.2%
Fort Worth 1046.04 51.0%
1050.06 51.2%
1061.02 52.1%
New Yorka 209.01 42.9%
0395 52.4%
65 56.5%
Oakland 4054 44.9%
4088 46.8%
4024b 49.6%
St. Louis 1212c 85.0%
1104 54.8%
1243 54.9%
Note: Child poverty rates were computed using 1990 census data.
a
Census tracts are from three separate counties: CT 209.01 (New York); CT
0395 (Kings); CT 65 (Bronx).
bCensus tract contained identified high-performing inner city school.
c
Census tract 1112 (54.6% child poverty) was replaced by identified
high-performing inner city school in census tract 1212.
Source: GAO's data analysis.
To select suburban schools, we (1) collected census child poverty rates
for all school districts in the defined suburban area outside the central
city of the selected metropolitan area and within the same state as the
central city; (2) ranked by census child poverty rates in the suburban
school districts; and (3) identified the three suburban school districts
closest to the 50th percentile, that is, the median suburban school
districts, based upon child poverty rates. We dropped districts that were
contiguous or had a 5 to 17-year-old population of less than 500 and
replaced them with the district with the next closest median level child
poverty that did not have any of these attributes.
For those districts, we selected the elementary school of the district. If
more than one elementary school served the school district, we selected
the elementary school in the district with the median child poverty rate
(as determined by free and reduced lunch eligibility) for elementary
schools in that district. (See table 6 for the child poverty rates for the
selected suburban school districts.)
Table 6: Selected Suburban School Districts' Child Poverty Rates
Metropolitan area Selected school district poverty rate
District 1 District 2 District 3
Boston 6.5% 6.6% 6.9%
Chicago 4.5% 4.6% 4.6%
Denver 8.8% 10.1% 11.0%
Fort Worth 12.0% 12.3% 13.9%
New York 4.9% 5.1% 5.2%
Oakland 7.5% 7.9% 8.4%
St. Louis 9.3% 10.5% 10.8%
Note: Child poverty rates were computed using 1995 census child poverty
estimates for school districts.
Source: GAO's data analysis.
Data Collected From 42 selected schools we obtained detailed information
for the 2000-01 school year on (1) school spending and federal revenues,
(2) staffing and teacher experience, and (3) student characteristics. The
practical difficulties of conducting any data collection effort may
introduce errors, commonly referred to as nonsampling errors. For example,
difficulties in how a particular question is interpreted or in the sources
of information that are available can introduce unwanted variability into
the results. We took steps in the development of the instrumentation, the
data collection, and the data editing and analysis to minimize these
errors. We pretested our data collection instrument with the Boston school
district and called individual district officials to clarify answers.
Completed instruments were examined for inconsistencies, and follow-up
calls were made to districts to clarify imprecise responses or data that
were unusually different from other respondent data.
* School spending data included (1) instructional staff salaries, (2)
certified professional staff salaries, (3) administrative staff
salaries, (4) operations staff salaries, (5) education materials
and supplies spending, and
* (6) building maintenance and repair spending. In addition, schools
reported federal sources of revenue.
* School, staff, and student information included numbers of (1)
regular education teachers, special education, English as a second
language instructional staff, and other specialized instructional
staff, for example, art teachers, reading teachers; (2) regular
education teacher assistants, special education teacher assistants,
and other instructional staff teacher assistants, for example, art
teacher assistants, reading teacher assistants;
* student support professional and nonprofessional staff by job
title;
* administrators and administrative assistants by job title; (5)
operations staff by job title; (6) the number of first-year
teachers; (7) total enrollment;
* number of students with disabilities and number of students
with limited English proficiency; (9) race and ethnicity of
students; and (10) the number of students eligible for free
and reduced lunch.
Data on student achievement, facilities, educational materials, and
parental involvement that may contribute to academic achievement were
obtained from site visits to 12 inner city and 12 suburban schools. We
developed a site visit protocol and pretested it at site visits to inner
city and suburban schools in the New York and Baltimore metropolitan
areas.
We obtained information on student achievement. In Fort Worth, we used
Grade 3 reading scores on the Texas Assessment of Academic Skills. In New
York, we used Grade 4 scores on the State English Language Arts
Assessment. In Oakland, we used Grade 4 reading scores on the Stanford 9
test. In St. Louis, we used Grade 3 Communication Arts scores on the
Missouri Assessment Program. In each metropolitan area, we contrasted the
achievement scores of the selected schools to the state average.
Depending upon data, information was collected as a dichotomous variable
(yes/no), date or period of time, number, or ranked scale assessment. (See
table 7 for school site visit information collected, assessment measure,
and description of the measurement scale.)
Methodology to Analyze Differences in Spending and Factors Accounting for
Spending Differences
Table 7: School Characteristics, Assessment Measure, and Measurement
Description
Category Assessment Measurement description
Facilities
Age of building Date Year of construction
Renovations Date Year of most recent renovation
Ancillary buildings Yes/no Presence of auxiliary classrooms
Classroom size Square feet Size of 2nd Grade classrooms
Special classrooms Yes/no Presence of special classrooms and
description a GAO assessment
Playgrounds Yes/no GAO assessment (1- 4 scale) and
Condition of facilities Scale description
Educational materials
Age of textbooks Date Year of purchase Computers Number Total computers in
building
GAO assessment (1- 3 scale) and
Modernization Scale description
Principal assessment (1-5 scale)
and
School supplies Scale description
Library Number Number of books
Parental involvement
Principal assessment (1-5 scale)
and
School activities Scale description
Donate/raise money Yes/no Principal assessment
Volunteer Yes/no Principal assessment
PTA participation Yes/no Principal assessment
Note: Scale is a subjective assessment.
a
Special classrooms include gymnasium, auditorium, cafeteria, art room,
music room, science room, and gardens, and were separately noted.
Source: GAO site visit data collection protocol.
For each metropolitan area, per-pupil spending3 for each of the three
inner city schools and three suburban schools were ordered and paired,
that is, the lowest spending inner city school was paired with the lowest
spending suburban school, the middle spending inner city school was paired
with the middle spending suburban school, and the highest spending inner
city school was paired with the highest spending suburban school.
3
Spending includes personnel salaries, building maintenance and repair, and
educational materials and supplies. Some expenditures, such as
transportation and district overhead, are, therefore, not included in
spending.
Page 36 GAO-03-234 Inner City Schools
To examine factors that explained differences in school spending, we
conducted regression analysis. Regression analysis is a statistical
methodology that measures the relationship between one variable and one or
more other variables.
In our regression model, we tried to determine the extent to which total
per-pupil spending at a selected individual school could be explained by
(1) average teacher salary at the school, (2) adjusted student-teacher
ratio at the school,4 (3) the ratio of students to student support staff
at the school, and (4) annual spending at the school on building
maintenance and repair.
The variables in the model were defined as follows:
o Total per-pupil spending-total dollars spent by the school in the
2000-01 school year divided by total enrollment.5
o Average teacher salary-total salary expenditure for teachers at the
school divided by the number of teachers. Teacher salary was used in
the regression to capture the salary structure at the school.6
o Adjusted student-teacher ratio-total enrollment adjusted for students
with special educational needs divided by the total certified
instructional staff. Adjusted enrollment differed from total
enrollment in that the adjusted enrollment included an additional
weight of 100 percent for each child receiving special education
instruction at the school and 50 percent for students with limited
English proficiency. Adjusted enrollment was used to capture the
direct higher spending by the school for students with special needs.
Teachers included: regular classroom teachers, special education
teachers, teachers of students with limited English proficiency, art
teachers, music teachers, physical education teachers, reading
teachers, teachers for the gifted and
4
Enrollment was weighted to account for students with disabilities and
students with limited English proficiency in order to more accurately
gauge the school's student-teacher ratio.
5
Total enrollment was calculated as the enrollment of the school on October
1, 2000.
6
It was assumed that across schools the salaries of other employees in the
school would be "structurally" related to the salaries of teachers. That
is, if teachers at a particular school earn on average a higher salary,
then other employees at the school, such as operations staff and
administrators, would similarly earn higher salaries.
talented, science teachers, and computer laboratory teachers.7 Teaching
assistants and paraprofessionals were not included because their direct
involvement with instruction was not always certain.
o The ratio of students to student support staff at the school was
computed by dividing the total enrollment by the total certified
professional staff. Support staff was not adjusted for students with
special needs because it was assumed that at the school level support
staff to student time is less dependent upon the disability of the
child. Total certified professional staff included: administrators,
health providers, and certified staff providing services to students.8
o Spending on building maintenance and repair at the school included
contracted maintenance and repair and salary expenditures for building
custodians and maintenance workers for the 2000-01 school year. (See
table 8 for the regression results for factors explaining differences
in per-pupil spending at the selected schools.)
7
Classroom and instructional-service paraprofessionals were not included.
8
Operation staff and clerical staff were not included.
Table 8: Regression Results for Factors Explaining Differences in
Per-Pupil Spending at Selected Schools
Independent variable Coefficient Standard error t-score Significance
Constant 3024.888 678.076 4.461
Teacher salary 7.718E-02 .011 7.295 .000a
Weighted student teacher-89.375 30.934 -2.889 .007a ratio
Student-support staff ratio -5.134 2.165 -2.372 .024a
Maintenance and repair 2.067E-03 .001 1.988 .055
Dependent variable: Per-pupil spending R = 0.854 F = 21.536 sig. = 0.000
aS
ignificant at the 0.05 level. Source: GAO's data analysis.
Appendix II presents selected data on the 42 schools examined in the seven
metropolitan areas, as well as additional information obtained from site
visits at 24 schools.
Appendix II: School Profiles
This appendix contains three tables of school-level information collected
from selected inner city and suburban schools in seven metropolitan areas.
Table 9 contains student characteristic information. Student
characteristic information includes enrollment, child poverty measured by
the census, percent of students with disabilities, percent of students
with limited English proficiency, and percent of children that are
minority.
Table 10 contains actual spending per child, then spending per child at
low, medium, and high weights for selected schools in seven metropolitan
areas. Table 11 includes information on the percent of first-year
teachers, federal dollars per child, and federal dollars as a percent of
total spending.
Table 9: School-Level Student Characteristics for Selected Schools in
Seven Metropolitan Areas
Percent with
Metropolitan Inner Percent Percent with limited Percent
city/ child English
area suburban Enrollment poverty disabilities proficiency minority
Oakland Inner 745 45 5 64 100
city
Oakland Inner 312 50 9 73 97
city
Oakland Inner 1,238 47 6 41 100
city
Oakland Suburb 402 8 8 15 64
Oakland Suburb 877 8 0 4 38
Oakland Suburb 460 8 8 3 16
St. Louis Inner 163 85 12 0 100
city
St. Louis Inner 292 55 13 0 100
city
St. Louis Inner 499 55 8 0 90
city
St. Louis Suburb 602 11 18 3 2
St. Louis Suburb 391 11 5 0 8
St. Louis Suburb 459 9 11 1 1
Fort Worth Inner 760 52 6 17 79
City
Fort Worth Inner 555 51 3 10 99
city
Fort Worth Inner 937 51 3 15 94
city
Fort Worth Suburb 413 12 18 2 11
Fort Worth Suburb 392 12 6 5 14
Fort Worth Suburb 373 14 17 13 85
New York Inner 484 56 9 22 100
city
New York Inner 645 52 11 18 98
city
New York Inner 630 43 6 3 100
city
New York Suburb 457 5 16 9 28
New York Suburb 553 5 9 3 n/a
New York Suburb 536 5 9 0 4
Boston Inner 712 51 21 0 96
city
Boston Inner 193 50 9 0 99
city
Boston Inner 250 49 17 12 71
city
Boston Suburb 386 7 12 0 3
Boston Suburb 979 7 15 0 1
Boston Suburb 335 7 8 3 3
Denver Inner 562 52 12 52 94
city
Denver Inner 372 52 13 19 94
city
Denver Inner 468 51 12 32 94
city
Denver Suburb 407 9 13 0 7
Denver Suburb 292 10 8 11 51
Denver Suburb 623 11 6 10 41
Chicago Inner 466 59 9 0 100
city
Chicago Inner 900 59 14 5 50
city
Chicago Inner 692 59 12 0 100
city
Chicago Suburb 503 5 17 1 11
Percent with
Metropolitan Inner Percent Percent with limited Percent
city/ child English
area suburban Enrollment poverty disabilities proficiency minority
Chicago Suburb 401 5 8 2 8
Chicago Suburb 280 5 6 5 13
Source: GAO's data analysis.
Table 10: Spending Per Pupil and Spending Per Pupil at Low, Medium, and
High Weights for Selected Schools in Seven Metropolitan Areas
Spending Metropolitan Inner city/ per pupil area suburban (unweighted)
Spending Spending Spending per pupil, low per pupil, per pupil, high
weight medium weight weight
Oakland Inner city 1 $4,022 $3,313 $2,442 $1,939
Oakland Inner city 2 $4,100 $3,226 $2,321 $1,818
Oakland Inner city 3 $2,810 $2,340 $1,804 $1,471
Oakland Suburb $4,975 $4,498 $4,104 $3,778
Oakland Suburb $2,470 $2,421 $2,316 $2,220
Oakland Suburb $4,849 $4,454 $4,223 $4,016
St. Louis Inner city $5,337 $4,168 $3,244 $2,656
St. Louis Inner city $6,311 $5,144 $4,287 $3,675
St Louis Inner city $3,162 $2,675 $2,230 $1,912
St. Louis Suburb $4,395 $3,705 $3,444 $3,218
St. Louis Suburb $3,467 $3,240 $3,085 $2,943
St. Louis Suburb $3,106 $2,769 $2,620 $2,486
Fort Worth Inner city $2,984 $2,542 $2,049 $1,718
Fort Worth Inner city $3,282 $2,879 $2,364 $2,007
Fort Worth Inner city $3,058 $2,658 $2,157 $1,817
Fort Worth Suburb $4,040 $3,396 $3,156 $2,947
Fort Worth Suburb $4,391 $4,060 $3,777 $3,534
Fort Worth Suburb $4,246 $3,546 $3,180 $2,885
New York Inner city $7,263 $5,966 $4,704 $3,887
New York Inner city $6,057 $4,929 $3,956 $3,307
New York Inner city $5,904 $5,150 $4,401 $3,843
New York Suburb $6,566 $5,616 $5,218 $4,877
New York Suburb $8,637 $7,874 $7,523 $7,205
New York Suburb $7,218 $6,623 $6,390 $6,173
Boston Inner city $6,731 $5,197 $4,366 $3,764
Boston Inner city $4,998 $4,223 $3,567 $3,087
Boston Inner city $5,770 $4,568 $3,752 $3,185
Boston Suburb $4,277 $3,802 $3,638 $3,487
Boston Suburb $4,433 $3,863 $3,683 $3,519
Boston Suburb $4,737 $4,357 $4,149 $3,962
Denver Inner city $3,240 $2,537 $1,898 $1,520
Denver Inner city $3,865 $3,113 $2,495 $2,084
Denver Inner city $3,852 $3,089 $2,410 $1,979
Denver Suburb $3,313 $2,931 $2,783 $2,649
Denver Suburb $5,404 $4,894 $4,496 $4,161
Denver Suburb $2,810 $2,578 $2,374 $2,202
Chicago Inner city $4,482 $3,743 $3,089 $2,629
Chicago Inner city $4,529 $3,615 $2,948 $2,490
Chicago Inner city $3,595 $2,939 $2,425 $2,065
Chicago Suburb $3,975 $3,430 $3,274 $3,132
Spending Spending Spending Spending
Metropolitan Inner city/ per pupil per pupil, per pupil, per pupil,
low high
area suburban (unweighted) weight medium weight
weight
Chicago Suburb $3,053 $2,812 $2,705 $2,606
Chicago Suburb $3,216 $2,996 $2,858 $2,734
Source: GAO's data analysis.
Table 11: Percent of First-Year Teachers, Federal Dollars Per Pupil, and
Federal Dollars as a Percent of Total Spending at Selected Schools in
Seven Metropolitan Areas
Metropolitan Inner city/ area suburban
Percent Federal Federal dollars first year dollars as percent of teachers
per pupil total spending
Oakland Inner city 12 $455
Oakland Inner city 6 $745
Oakland Inner city 12 $294
Oakland Suburb 0 $179
Oakland Suburb 15 $17
Oakland Suburb 0 $10
St. Louis Inner city 0 $1,493
St. Louis Inner city 16 $2,581
St Louis Inner city 11 $875
St. Louis Suburb 5 $193
St. Louis Suburb 17 $106
St. Louis Suburb 4 $107
Fort Worth Inner city 11 $510
Fort Worth Inner city 4 $966
Fort Worth Inner city 11 $586
Fort Worth Suburb 19 $313
Fort Worth Suburb 4 $189
Fort Worth Suburb 18 $591
aa
New York Inner city a
aa
New York Inner city 24
New York Inner city 13 $669
New York Suburb 0 $268
New York Suburb 3 $180
New York Suburb 0 $14
Boston Inner city 12 $448
Boston Inner city 31 $329
Boston Inner city 6 $578
Boston Suburb 0 $130
Boston Suburb 0 $169
Boston Suburb 0 $117
Denver Inner city 9 $673
Denver Inner city 0 $428
Denver Inner city 19 $716
Denver Suburb 19 $111
Denver Suburb 14 $129
Denver Suburb 8 $93
Chicago Inner city 15 $623
Chicago Inner city 8 $480
Chicago Inner city 6 $659
Percent Federal Federal dollars
Metropolitan Inner city/ first year dollars as percent of
area suburban teachers per pupil total spending
Chicago Suburb 0 $228 6
Chicago Suburb 0 $263 9
Chicago Suburb 0 $71 2
a
Data not provided by the New York schools. Source: GAO's data analysis.
Appendix III: GAO Contacts and Staff Acknowledgments
Harriet Ganson, (202) 512-7042
GAO Contacts
Peter Minarik, (202) 512-7230
In addition to those named above, Elisabeth Anderson, Shannon McKay,
Eve Veliz, and Sarit Weisburd made key contributions to this report. Luann
Moy provided important methodological contributions to the review of the
research. Patrick DiBattista also provided key technical assistance.
(130111)
GAO's Mission
The General Accounting Office, the investigative arm of Congress, exists
to support Congress in meeting its constitutional responsibilities and to
help improve the performance and accountability of the federal government
for the American people. GAO examines the use of public funds; evaluates
federal programs and policies; and provides analyses, recommendations, and
other assistance to help Congress make informed oversight, policy, and
funding decisions. GAO's commitment to good government is reflected in its
core values of accountability, integrity, and reliability.
Obtaining Copies of GAO Reports and Testimony
The fastest and easiest way to obtain copies of GAO documents at no cost
is through the Internet. GAO's Web site (www.gao.gov) contains abstracts
and full- text files of current reports and testimony and an expanding
archive of older products. The Web site features a search engine to help
you locate documents using key words and phrases. You can print these
documents in their entirety, including charts and other graphics.
Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as "Today's Reports," on its
Web site daily. The list contains links to the full-text document files.
To have GAO e-mail this list to you every afternoon, go to www.gao.gov and
select "Subscribe to daily E-mail alert for newly released products" under
the GAO Reports heading.
Order by Mail or Phone
The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:
U.S. General Accounting Office 441 G Street NW, Room LM Washington, D.C.
20548
To order by Phone: Voice: (202) 512-6000
TDD: (202) 512-2537
Fax: (202) 512-6061
To Report Fraud, Contact:
Web site: www.gao.gov/fraudnet/fraudnet.htm
E-mail: [email protected]
Automated answering system: (800) 424-5454 or (202) 512-7470
Federal Programs
Jeff Nelligan, managing director, [email protected] (202) 512-4800 U.S.
General Accounting Office, 441 G Street NW, Room 7149 Washington, D.C.
20548
*** End of document. ***