Title I Program: Stronger Accountability Needed for Performance of
Disadvantaged Students (Chapter Report, 06/01/2000, GAO/HEHS-00-89).

Pursuant to a congressional request, GAO provided information on: (1)
Title I services at schoolwide and targeted assistance schools; (2)
state efforts to hold schools and districts accountable for student
achievement; and (3) research and evaluations of Title I and schoolwide
programs.

GAO noted that: (1) both schoolwide and targeted assistance schools
generally offered similar services, such as tutoring, and targeted
additional services to students needing extra help; (2) however,
schoolwide schools were generally more likely than targeted assistance
schools to provide services such as extended day programs and often
chose different methods of service delivery, such as moving students in
and out of flexible groups as their achievement levels changed; (3)
educators at the high-poverty schools GAO visited generally preferred
the schoolwide approach because they believed that it allowed them to
serve more students, facilitated faculty collaboration, and allowed them
to deliver services more efficiently and effectively; (4) however, some
principals and teachers cautioned that schools adopting the schoolwide
approach need to be careful that low-achieving students still receive
the extra help they may need to improve their academic performance; (5)
many states have yet to take all the steps necessary to oversee program
operations and hold districts and schools accountable for results; (6)
states varied considerably in the frequency and focus of their efforts
to monitor compliance with Title I requirements and to oversee program
quality; (7) in addition, some states had collected extensive and
detailed information on educational outcomes, but most states had
substantially less information on educational outcomes and on
disadvantaged students in general; (8) the majority of states had
established criteria to determine whether schools and districts were
performing satisfactorily; (9) however, these criteria were sometimes
confusing or vague and were based solely on the performance of the
student population as a whole, without reference to the achievement of
specific subgroups of children, such as students from low-income
families or students with limited English proficiency; (10)
consequently, states are not yet in a position to ensure accountability
for the educational outcomes of disadvantaged students, the children
that remain central to the mission of the Title I program; (11) limited
data and methodological problems have made it difficult to draw firm
conclusions about whether Title I in general-- and schoolwide programs
in particular--are effective in improving educational outcomes; and (12)
because schools and districts have considerable discretion in spending
their Title I dollars and are not required to report the specific
services provided, it has been difficult for researchers to isolate the
effect of specific services.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  HEHS-00-89
     TITLE:  Title I Program: Stronger Accountability Needed for
	     Performance of Disadvantaged Students
      DATE:  06/01/2000
   SUBJECT:  Accountability
	     Education program evaluation
	     Schools
	     Disadvantaged persons
	     Students
	     Performance measures
	     Economically depressed areas
	     State-administered programs
	     Academic achievement
	     Aid for education
IDENTIFIER:  North Carolina
	     Pennsylvania
	     Rhode Island
	     Tennessee
	     Texas
	     Dept. of Education Title I Program

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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/HEHS-00-89

6

14

Title I Program Targeted Primarily--but Not Exclusively--to
High-Poverty Schools 14

High-Poverty Schools May Choose the Schoolwide Option, Which
Focuses on All Students in the School 17

States Provide Program Accountability and Program Oversight 19

Education Provides States With Assistance and Oversight 21

Objectives, Scope, and Methodology 22

Additional Opportunities in High-Poverty Schools

27

Schoolwide and Targeted Assistance Schools Generally Offered
Similar Services but Varied in Service Delivery Methods 27

Schoolwide Approach Viewed as Benefiting More Students 29

Accountable for Outcomes of Disadvantaged Students

32

States Are Responsible for Ensuring Local Compliance; Education
Provides Oversight and Technical Assistance 32

Focus and Frequency of Program Oversight Efforts Varied
Across States 33

States Have Made Uneven Progress in Collecting and Reporting
Assessment Data Essential to Accountability 35

States' Annual Progress Criteria Include Only Assessment Results
for Overall Student Population, Limiting Accountability for Disadvantaged
Students 44

Limited

46

Research Efforts Have Provided Limited Information on Program Effectiveness
46

Data Limitations Present Challenges in Evaluating Program
Effectiveness 48

Consideration, Agency Comments, and Our Evaluation

50

Conclusions 50

Recommendations 51

Matter for Congressional Consideration 51

Agency Comments and Our Evaluation 51

Appendix I: Comments From the Department of Education

54

Appendix II: Major Contributors to This Report

57

Table 1: Examples of Key Title I Requirements 20

Table 2: Characteristics of Schools Visited 24

Table 3: Principals Reporting Use of Extended-Time Learning Opportunities 28

Table 4: Assessment Design Issues 37

Figure 1: Distribution of Schools by Poverty Level and Title I Status 15

Figure 2: Example of District-School Title I Allocation Process 17

Figure 3: Frequency of States' On-Site Monitoring 35

Figure 4: Other Outcome Measures Gathered by States 40

Figure 5: Number of States That Disaggregate Assessment and
Other Outcome Data, as of Fall 1999 41

Figure 6: Number of States Reporting Disaggregated Assessment
Data Through the Internet 42

CCD Common Core of Data

SERFF Study of Education Resources and Federal Funding

Health, Education, and
Human Services Division

B-282653

June 1, 2000

The Honorable James M. Jeffords
Chairman
The Honorable Edward M. Kennedy
Ranking Minority Member
Committee on Health, Education, Labor, and Pensions
United States Senate

The Honorable Christopher J. Dodd
Ranking Minority Member
Subcommittee on Children and Families
Committee on Health, Education, Labor, and Pensions
United States Senate

The Title I program was established in 1965 to help schools meet the needs
of economically and educationally disadvantaged students. Title I has
traditionally directed its funds to those students who are lowest-achieving
or at highest risk for school failure. In 1994, the reauthorization of Title
I established new provisions encouraging the use of the schoolwide option
and increasing accountability for the educational outcomes of all children.
You requested us to provide information about Title I services at schoolwide
and targeted assistance schools, state efforts to hold schools and districts
accountable for student achievement, and research and evaluations of Title I
and schoolwide programs.

This report was prepared under the direction of Harriet C. Ganson, Assistant
Director, Education, Workforce, and Income Security Issues, who may be
reached at (202) 512-9045 if you or your staff have any questions. Major
contributors to this report are listed in app. II.
Marnie S. Shaul, Associate Director
Education, Workforce, and
Income Security Issues

Executive Summary

Title I, the largest federal elementary and secondary education program, was
established in 1965 to help schools meet the needs of economically and
educationally disadvantaged students. Title I has traditionally directed its
funds--approximately $7.9 billion in fiscal year 1999--to those students who
are lowest-achieving or at highest risk for school failure (targeted
assistance). The 1994 reauthorization1 of Title I expanded the focus of the
program by increasing the number of schools eligible to use their Title I
funds to improve the school as a whole (a schoolwide program). In 1994, the
Congress also established new provisions aimed at creating greater
accountability for educational outcomes--not just for disadvantaged
children, but for all children. The addition of this broader focus to Title
I--particularly the growth in schoolwide programs--caused concern for some
educators and policymakers, who feared that some disadvantaged students
might lose services they would otherwise have received.

In light of the current discussions regarding reauthorization of Title I,
the Chairman and Ranking Minority Member of the Senate Committee on Health,
Education, Labor and Pensions and the Ranking Minority Member of that
Committee's Subcommittee on Children and Families asked GAO to examine how
the changes to Title I, particularly the growth in schoolwide programs, have
affected Title I's focus on disadvantaged children. GAO reviewed (1) the
services provided under Title I schoolwide and targeted assistance programs,
(2) states' efforts to ensure compliance with key Title I requirements and
hold districts and schools accountable for educational outcomes, and (3)
research on and evaluation of Title I overall and schoolwide programs in
particular.

Title I allocations vary considerably across schools, depending on, among
other factors, the amount of money the district receives, the number of
children in poverty, and how the school district chooses--with certain
restrictions--to allocate its Title I dollars to individual schools.
Although high-poverty schools are more likely to receive Title I funding
across the nation, some high-poverty schools receive no Title I funds and
many schools with below-average poverty rates do receive Title I dollars.
According to 1999 Department of Education (Education) data, an estimated 54
percent of the approximately 91,000 public elementary and secondary schools
nationwide received Title I funds, and about one-third of the schools that
received Title I funds have poverty rates near or below the national
average.

Prior to 1994, only schools with poverty rates of 75 percent or higher were
eligible to choose the schoolwide option, but the 1994 reauthorization
lowered the eligibility threshold to 50 percent. Currently, about half of
Title I schools are eligible to choose the schoolwide option, and a large
majority--an estimated 82 percent--of eligible schools have chosen this
approach. The remaining Title I schools--those that are not eligible or
choose not to adopt the schoolwide approach--are called targeted assistance
schools and are required to target their Title I dollars to low-achieving
students within the school.

In addition to expanding the availability of the schoolwide option, the 1994
Title I reauthorization broadened the accountability provisions. States are
now expected to hold districts and schools accountable for educational
outcomes. States are to collect and publicly report assessment data, develop
criteria to determine whether schools and districts are performing
satisfactorily,2 and take actions to improve the performance of
low-performing schools and districts. Then, to ensure that all types of
students are making progress, states are to collect and report assessment
results by six specified student categories--gender, racial and ethnic
group, English proficiency status, disability status, migrant status, and
economic status. States had until the 1997-1998 school year to develop
content and performance standards. Education, as authorized by statute,
extended the deadline for performance standards for many states to coincide
with the deadline for assessments, which must be completely finalized by the
2000−2001 school year.

Education has responsibility for general oversight of Title I. As part of
its oversight, Education conducts on-site reviews of each state's program
every 3 to 4 years. This is done to assess whether states are adequately
monitoring how schools implement Title I requirements--those that are
financial and programmatic and those that are specifically related to
outcomes. In addition, Education provides technical assistance regarding the
interpretation of Title I--requirements and issues related to overall
educational quality. This assistance takes a variety of forms such as
conferences, forums, and ongoing staff contacts. Finally, Education conducts
research on the effects of services provided under Title I.

Both schoolwide and targeted assistance schools generally offered similar
services, such as tutoring, and targeted additional services to students
needing extra help. However, schoolwide schools were generally more likely
than targeted assistance schools to provide services such as extended day
programs and often chose different methods of service delivery, such as
moving students in and out of flexible groups as their achievement levels
changed. Educators at the high-poverty schools GAO visited generally
preferred the schoolwide approach because they believed that it allowed them
to serve more students, facilitated faculty collaboration, and allowed them
to deliver services more efficiently and effectively. However, some
principals and teachers cautioned that schools adopting the schoolwide
approach need to be careful that low-achieving students still receive the
extra help they may need to improve their academic performance.

Many states have yet to take all the steps necessary to oversee program
operations and hold districts and schools accountable for results. States
varied considerably in the frequency and focus of their efforts to monitor
compliance with Title I requirements and to oversee program quality. In
addition, some states had collected extensive and detailed information on
educational outcomes, but most states had substantially less information on
educational outcomes and on disadvantaged students in general. The majority
of states had established criteria to determine whether schools and/or
districts were performing satisfactorily. However, these criteria were
sometimes confusing or vague and (with only one state as an exception) were
based solely on the performance of the student population as a whole,
without reference to the achievement of specific subgroups of children, such
as students from low-income families or students with limited English
proficiency. Consequently, states are not yet in a position to ensure
accountability for the educational outcomes of disadvantaged students, the
children that remain central to the mission of the Title I program.

Limited data and methodological problems have made it difficult to draw firm
conclusions about whether Title I in general--and schoolwide programs in
particular--are effective in improving educational outcomes. Because schools
and districts have considerable discretion in spending their Title I dollars
and are not required to report the specific services provided, it has been
difficult for researchers to isolate the effect of specific services.
Education plans to expand its existing data collection to include
information specific to Title I and schoolwide programs. These changes,
combined with other actions to improve the completeness and quality of
existing data, could facilitate research improvements.

Opportunities in High-Poverty Schools

Title I schoolwide and targeted assistance schools generally offered similar
educational services. In the schools GAO visited, both schoolwide and
targeted assistance schools directed specific services (such as tutoring) at
students they had identified as needing additional assistance. However,
according to recent surveys conducted by Education, schools using the
schoolwide approach were more likely than targeted assistance schools to
offer programs with extended instructional time, such as after school,
weekend, or summer school programs.

Principals and teachers at the schools GAO visited said that the targeted
assistance approach could lead schools to choose convenient, but less
preferable, service delivery methods compared with the schoolwide approach.
One Education survey found that 80 percent of targeted assistance elementary
schools used pull-out programs, which take students out of their regular
classrooms to provide additional tutoring or assistance; only 53 percent of
the schoolwide schools used this approach. The pull-out option provides a
convenient method for tracking students who received services and
demonstrating that funds were used for students formally identified as
eligible under Title I. While the schoolwide schools GAO visited sometimes
used the pull-out strategy, they often preferred other approaches such as
moving students in and out of flexible groups as their achievement levels
changed.

Many district and school officials also said that the schoolwide approach
allowed them to serve more students, fostered increased collaboration, and
offered greater flexibility. In the high-poverty schools GAO visited,
principals and teachers said that because most or all of their students
suffered from educational disadvantages, the schoolwide approach allowed
them to address learning deficiencies for a greater number of students, as
well as to prevent such deficiencies. However, some school and district
officials cautioned that schools adopting the schoolwide approach need to be
careful that low-achieving students receive the extra help they need to
improve their academic performance.

District and school staff said that they believed the schoolwide approach
helped principals and teachers develop unified goals, engage in shared
decision-making, and assume responsibility for school improvement. State,
district, and school officials also stated that the schoolwide approach
allowed schools to use Title I funds without the need to direct these funds
to specific students. Consequently, schools could use their federal dollars
to employ a wider variety of instructional strategies. While those schools
that have chosen the schoolwide option see a number of advantages in doing
so, evidence regarding the overall effectiveness of the schoolwide approach
is inconclusive, as discussed later.

Accountable for Educational Outcomes of Disadvantaged Students

State efforts to monitor compliance and oversee program quality varied
dramatically. Forty-seven states reported that they include on-site visits
as part of their monitoring procedures, and many of these states reported
that they rely solely on on-site visits to monitor certain key program
requirements. The average time between visits ranged from 2 years or less
(for 6 states) to more than 7 years (for 17 states). Three states reported
that they made no on-site visits, while three states reported that they
visited all their school districts each year.

Title I gives states considerable flexibility in deciding how to develop and
implement assessments of students' performance. Consequently, there are
major differences in states' assessments, including the format, content, and
difficulty of the assessments, as well as whether all students participated
in state-sponsored assessments. Moreover, while some states provided
extensive, detailed, and timely assessment data that school and district
officials found useful in improving instruction, other states collected or
reported less data on student achievement. Only about one-third of the
states have collected disaggregated data by all six required categories.

States varied in how they set criteria to judge whether schools and
districts have been performing satisfactorily. While some states had
developed clear and specific criteria for assessing adequate yearly progress
for schools and districts, other states had outlined only vague objectives.
The law does not require states to include criteria for adequate yearly
progress that are based on disaggregated data, and only one state has chosen
to establish criteria that include the performance of disadvantaged
students. Without these data, it will be more difficult for states to hold
districts and schools accountable for the achievement of disadvantaged
students, and to discern whether achievement gaps between disadvantaged and
nondisadvantaged students are closing.

Research efforts to determine the effect of Title I on student achievement
have provided only limited information. Recent evaluations suggest that
Title I may have had a limited positive effect on overall student
achievement in high-poverty schools; less information is available
specifically on the effect of schoolwide programs. Education plans to expand
its data collection efforts in ways that may facilitate future research on
program effectiveness for both Title I overall and schoolwide programs.
These data could later be combined with other information about student
achievement or school characteristics to allow for more comprehensive
analyses of the effectiveness of Title I and schoolwide programs. However,
while linking key databases may facilitate more comprehensive analyses,
additional work is needed to improve data quality and consistency,
especially for key variables, such as school poverty rate.

GAO makes two recommendations to the Secretary of Education. First,
Education should conduct additional activities to facilitate the exchange of
information and best practices among states so they can identify ways to
improve the timeliness and specificity of their assessment data, the
collection and reporting of disaggregated assessment data, and the clarity
of their criteria for adequate yearly progress. Second, Education should
implement additional measures to improve research on the effectiveness of
specific services in both schoolwide programs and targeted assistance
schools. Such measures could include expanding and improving current data
collection efforts so that comprehensive analyses could be conducted linking
program characteristics to services and student outcomes and/or developing
an evaluation design for a study or set of studies of educational services
that would include national representation of both schoolwide and targeted
assistance schools.

To hold schools and districts accountable for improving the performance of
disadvantaged students and to help educators, parents, and others discern
whether achievement gaps between disadvantaged and nondisadvantaged students
are closing, the Congress should consider requiring that states' criteria
for progress, as expressed in their definitions of adequate yearly progress,
apply specifically to disadvantaged children as well as to the overall
student population.

GAO obtained comments on a draft of this report from the Department of
Education. Education agreed with most of the report's findings and
conclusions but did not cite specific actions it would take in response to
our recommendations.

Introduction

Title I of the Elementary and Secondary Education Act (Title I), the largest
federal program for elementary and secondary education, targets its $7.9
billion budget primarily to schools with a relatively high percentage of
students from low-income families. Nearly half of Title I schools operate on
a schoolwide basis--that is, the school can use its Title I dollars in
combination with other funds to improve the school as a whole. The remaining
schools that receive Title I funds are expected to target these funds to
lower-achieving students within the school. In 1994, the Congress
established new requirements whereby states (in addition to their general
oversight responsibilities) were to collect and report information on
educational outcomes and hold schools and districts accountable for results.

High-Poverty Schools

Title I was established in 1965 to help schools--especially high-poverty
schools--meet the needs of disadvantaged students. Children from low-income
families often face obstacles that can reduce their chances for success in
learning. For example, children from low-income families are more likely to
be exposed to drug abuse, violence, and unhealthy living conditions. In
addition, parents in economically disadvantaged families sometimes have
limited education and involvement in their children's learning. However, the
effects of poverty on student achievement are not confined to students who
happen to be poor themselves. Research shows that students in schools with
high poverty rates are more likely to be low performers, independent of
their own family background. These effects may be related to lower
expectations on the part of teachers in high-poverty schools, lack of highly
qualified teachers or sufficient resources in these schools, or other
factors.

In recognition of these special challenges, the Title I program targets
primarily high-poverty schools. Although high-poverty schools are more
likely to receive Title I funding, some high-poverty schools receive no
Title I funds and many schools with below-average poverty rates do receive
Title I dollars. According to 1997-98 Education data, an estimated 54
percent of all schools have poverty rates near or below the national
average, compared to about one-third of Title I schools (see fig. 1).

Note: Under Title I, school districts have some discretion in selecting the
measure used for schools' poverty rates. The most commonly used measure has
been students' eligibility for free and reduced-price lunches.

Source: Jay Chambers, Joanne Lieberman, Tom Parrish, and others, Study of
Educational Resources and Federal Funding (Washington, D.C.: U.S. Department
of Education, Planning and Evaluation Service, 1999).

Title I dollars are distributed to local school districts through state
education agencies. The amount each school district receives is determined
by a complex formula that incorporates, among other factors, the average
per-pupil expenditure in the state, the number of children in poverty, and
previous allocations to the state and the district. Once Title I dollars
reach the local school district, district officials distribute the funds to
the schools. A school is Title I eligible if its school attendance area has
a poverty rate that is at least equal to the district average rate or 35
percent (whichever is less). District officials are required to follow
certain rules in allocating Title I funds to schools based on the number of
low-income students. First, districts must provide Title I funding to all
schools with a poverty rate of 75 percent or higher before providing Title I
funding to any school with a poverty rate below 75 percent. In addition,
district officials must provide funding to schools with the same grade span
in rank order of poverty rate.3 Even within these requirements, however,
district officials have some discretion as to which schools get Title I
dollars and how much each school receives. Districts may choose to
concentrate their Title I funds on their highest-poverty schools and limit
school eligibility to a poverty level that is higher than the districtwide
average. Districts may give schools different amounts per poor child so long
as schools with higher poverty rates receive higher allocations per poor
child than schools with lower poverty rates.4 Fig. 2 provides an example of
how this allocation process works.

aDollar amounts shown are allocations per low-income child.

on All Students in the School

Once funding has been distributed to the school level, Title I has
traditionally expected schools to direct the funds to students who are
low-achieving or at highest risk for school failure. However, a provision
known as the schoolwide program allows a school to spend its Title I funds
to improve the school as a whole, rather than targeting Title I funds to
low-achieving students. The 1994 reauthorization of Title I increased the
number of schools eligible to use Title I funds on a schoolwide basis.
Before 1994, only schools with poverty rates above 75 percent were eligible
to choose the schoolwide option, but the 1994 reauthorization lowered the
eligibility threshold to 50 percent. In addition, under a schoolwide
program, the school is encouraged to take an integrated approach, combining
federal resources with other funds to implement a school plan, instead of
viewing each program in isolation.

The 1994 reauthorization encouraged the use of the schoolwide option and
provided for waivers so that some otherwise ineligible schools could also
adopt a schoolwide program. Consequently, the number of schools operating
schoolwide programs has increased dramatically--from 3,903 in the 1993-94
school year to 19,701 in 1999. As of 1999, about half of Title I schools
were eligible to choose the schoolwide option, and a large majority--an
estimated 82 percent--of eligible schools have chosen the schoolwide
approach. The remaining Title I schools--those that are not eligible or
choose not to adopt the schoolwide approach--are called targeted assistance
schools. These schools are required to target their Title I dollars to
low-achieving students within the school.

Schoolwide schools had many of the same characteristics as high-poverty
schools in general, because eligibility for schoolwide status is determined
by poverty rate. In general, schools with poverty rates of 50 percent or
greater are more likely to be elementary schools, to be located in urban
areas, and to have a higher percentage of nonwhite students. Schools that
adopted the schoolwide approach shared these characteristics.

States varied a great deal in the proportion of schools choosing the
schoolwide option. Some states had a greater proportion of schoolwide
programs than would be expected on the basis of the number of high-poverty
schools in that state, while other states had a smaller proportion of
schoolwide programs. For example, while 18 percent of New Jersey schools had
poverty rates of 50 percent or greater, only 4 percent of all NewJersey
schools (and 11 percent of Title I schools in New Jersey) operated
schoolwide programs. In contrast, 43 percent of schools in South Carolina
had poverty rates of 50 percent or greater, and many of these schools
operated schoolwide programs--39 percent of schools in the state and 90
percent of Title I schools.5

Factors such as funding allocations, state and district support for
schoolwide programs, and local circumstances may have contributed to the
variation across states in the proportion of schools choosing schoolwide
programs. For example, the proportion of all public schools in a state that
received Title I funds varied considerably across states--from 20 percent in
Nevada to 77 percent in Mississippi. A state where Title I funds were
allocated to a greater proportion of schools, particularly schools with
lower poverty rates, would likely have fewer schoolwide programs and more
targeted assistance schools compared with a state where Title I dollars were
concentrated in fewer schools. In addition, adoption of schoolwide programs
may have varied because of state and district policies encouraging the
schoolwide approach. According to school and district officials, other local
or regional circumstances may also have played a role in the decision to
choose a schoolwide program.

The 1994 legislation gave the states additional responsibilities designed to
increase Title I's focus on educational outcomes for all students. Under the
legislation, states were to take the major responsibility for holding
schools and districts accountable for student achievement. Specifically,
states were required to develop content standards to describe what students
need to know and performance standards to describe expected proficiency in
at least the core subject areas of reading and mathematics. States were then
to establish assessments to measure how students are doing in relation to
the content and performance standards.

Each state must report the results of assessments, as well as its other
outcome data, to schools, districts, and the public. In addition to the
overall results, states must collect and publicly report assessment results
by six specified student categories: gender, racial and ethnic group,
English proficiency status, disability status, migrant status, and economic
status. States may also select other outcome measures (such as graduation
rates) to capture important facets of school performance that cannot be
measured through assessments. States must also develop criteria for
determining whether schools and districts are performing satisfactorily.
These criteria, collectively called adequate yearly progress, are designed
to help states identify low-performing schools and districts. States may
provide additional assistance, or invoke penalties or sanctions, to help and
motivate low-performing schools to improve.

The deadline for implementing content and performance standards was the
1997-98 school year. Education, as authorized by statute, extended the
deadline for performance standards for many states to coincide with the
deadline for assessments, which must be completed by the 2000-2001 school
year. During the transition period, while states are developing standards,
assessments, and adequate yearly progress benchmarks, they can use an
interim process for assessing schools and districts.

The new requirements for assessment, data collection and reporting, and
accountability augment rather than replace Title I's long-standing financial
and programmatic requirements. Under Title I, states have a significant role
in ensuring that schools and districts comply with all these requirements.
Title I also contains a number of requirements that address a variety of
important program objectives, many of them directly related to protecting
the interests of disadvantaged children. For example, requirements are
directed not only at ensuring that Title I funds are targeted to
high-poverty schools in accordance with the law, but also at promoting
parental involvement in education--a factor that educators agree is
particularly important to the success of poor and low-achieving students.
Examples of some of Title I's key requirements are given in table 1.

           Objective                            Requirement
                                 Districts that get more than $500,000 in
                                 Title I funds must spend 1 percent of
                                 their Title I allocations for parental
                                 involvement activities.

 Getting parents more involved   Title I schools are required to have a
 in their children's education   school-parent compact, which describes the
                                 school's responsibility and the ways in
                                 which parents will be responsible for
                                 supporting their children's learning, such
                                 as monitoring attendance, homework
                                 completion, and television watching.
                                 Title I schools that choose a schoolwide
                                 program must produce a schoolwide plan
                                 that includes eight components required by
                                 law. Among these components are strategies
 Ensuring quality schoolwide     and activities to address the specific
 planning that addresses the     needs of disadvantaged children.
 needs of disadvantaged students
                                 Title I schools are expected to implement
                                 their schoolwide plans, using any approach
                                 they choose, as long as the plan contains
                                 the required components.
                                 Districts must meet maintenance-of-effort
                                 requirements--that is, aggregate state and
                                 local education expenditures for the
                                 preceding year generally may not be less
                                 than 90 percent of the expenditures for
                                 the year before.

 Assuring continuation of state  Schools are expected to use federal funds
 and local funding levels        to supplement, not supplant, funds that
                                 would be available in the absence of
                                 federal funds for the education of
                                 students in Title I schools.

                                 State and local funds must be used to
                                 provide services in Title I schools that
                                 are comparable to services those funds are
                                 providing in other schools.
                                 Districts must allocate federal funds in
 Ensuring that Title I funds are accordance with Title I regulations that
 directed to high-poverty        target high-poverty schools.
 schools and disadvantaged
 students in accordance with the Targeted assistance schools must direct
 law                             Title I funds primarily to services for
                                 students identified as lowest achieving or
                                 at risk of school failure.

The Congress and Education anticipated that the new accountability
requirements, combined with Title I's financial and programmatic
requirements, would help all schools improve student achievement. Title I's
programmatic and financial requirements would help ensure that important
building blocks--such as strong parental involvement and additional funding
for high-poverty schools--were in place. The availability of assessment and
other outcome data at the local level would provide useful information to
help schools identify and work on areas of weakness. The public distribution
of these outcome data would provide a powerful incentive for improvement,
and using these data to identify low-performing schools and districts would
create added accountability. The success of this process, however, depends
on states' ability to conduct careful oversight; collect and report
detailed, specific, and valid outcome data; and implement clear progress
criteria for schools and districts. In addition, states must collect and use
the disaggregated assessment data--in combination with their oversight of
key program requirements--to protect the interests not only of the general
student population, but also of the disadvantaged students targeted under
the Title I program.

Although the states have the major responsibility for ensuring that local
schools and districts comply with key Title I financial and programmatic
requirements, Education oversees how states implement Title I requirements
related to outcomes, and whether states sufficiently monitor how schools and
districts implement Title I's financial and programmatic requirements.
Education is also responsible for providing technical assistance to the
states, not only on the interpretation of Title I requirements, but also on
other broader issues related to overall educational quality.

As a major component of its oversight process, Education conducts on-site
reviews for each state.6 Site visits are generally conducted on a rotating
cycle every 3 to 4 years, with review teams spending about 5 business days
in each state they visit.7 Education requires states to submit a
self-assessment prior to the on-site visit to help focus the review. In
addition to meeting with state education department officials, review teams
generally visit a few local districts and schools.

In keeping with Education's technical assistance role, Education staff told
us that they review state programs from a problem-solving rather than
strictly a compliance perspective. The agency has also shifted from
conducting separate reviews that focused on only a single federal program to
broader, integrated reviews--including Title I reviews--looking at themes
that cut across several federal programs. For example, themes have included
parental involvement and professional development, which are important
components of several federal programs. Although the theme is not the sole
focus of the monitoring visit, Education staff said that areas related to
the theme may be emphasized more. Education officials told us that the
process for conducting on-site reviews is evolving and they expected to make
additional changes in the future. While Education staff expressed support
for the broader, cross-cutting review process, several individuals were
concerned that this broader process may result in Title I compliance issues
being monitored less thoroughly. Education officials said that they plan to
supplement the integrated review process with program-specific reviews to
address this concern.

In its leadership role, Education can encourage states to comply with Title
I requirements. Education sponsors conferences and forums to allow states to
share information, provides written guidance to states, and discusses issues
with state officials during on-site monitoring reviews, as well as other
means of providing technical assistance for program improvement. Education,
for example, sponsored three regional meetings on assessment systems and has
provided the states with some written guidance on progress criteria.
Education is currently providing technical assistance to states in
developing their progress criteria and final assessment systems.

The Chairman and the Ranking Member of the Senate Committee on Health,
Education, Labor and Pensions and the Ranking Minority Member of that
Committee's Subcommittee on Children and Families asked us to examine the
implications of the 1994 changes in the Title I program, particularly the
growth of schoolwide programs. Specifically, the objectives of this study
were to describe

ï¿½ the services offered under the schoolwide and targeted assistance
approaches and the extent to which schoolwide and targeted assistance
schools provide special assistance to low-achieving or disadvantaged
children within the school;

ï¿½ how states and the federal government exercise general oversight for the
Title I program; and how they measure, report, and create accountability for
educational outcomes, both for the student population as a whole and for
low-achieving and disadvantaged students in particular; and

ï¿½ the evaluations that have been conducted to examine whether Title I and
schoolwide programs are contributing to improved student achievement.

To address these objectives, we obtained and analyzed information from a
variety of sources at the federal, state, district, and school levels. To
obtain data on characteristics of and services provided by schoolwide and
targeted assistance schools, we reviewed and analyzed three national
databases collected by Education:

ï¿½ the Common Core of Data (CCD), which contains descriptive information on
the approximately 91,000 public elementary and secondary schools in the
United States, as reported by the 50 states;

ï¿½ the Study of Education Resources and Federal Funding (SERFF), which
contains information on school characteristics and services from a
nationally representative sample of over 700 schools and 180 school
districts in the 1998 school year; and

ï¿½ the School-Level Implementation of Standards-Based Reform: Findings From
the Follow-up Public School Survey on Education Reform, administered in 1998
to a nationally representative sample of about 1,600 school principals.

To obtain information on services provided under both targeted assistance
and schoolwide programs and on the implementation of accountability
requirements, we conducted on-site visits and interviewed state officials,
district staff, principals, and teachers in five states: North Carolina,
Pennsylvania, Rhode Island, Tennessee, and Texas. We chose these states to
illustrate a variety of characteristics, including the number and percentage
of schoolwide schools, the status of state assessments, the growth in the
school-age population in the past 5 years, the state share of education
expenditures, and Title I allocations per student. We visited 3 districts
each in North Carolina, Pennsylvania, Tennessee, and Texas, and 1 in Rhode
Island, for a total of 13 districts. In each district we interviewed the
Title I director and other officials as appropriate, including
superintendents, school support staff, and district staff involved in
assessment. We selected these districts on the basis of several criteria,
including their size and location. The districts we visited ranged in size
from just under 1,000 to nearly 70,000 students and from 49 to 89 percent of
students in poverty.

Similarly, we visited between 1 and 3 Title I schools within each of these
districts, for a total of 21 schools (15 elementary schools, 5 middle
schools, and 1 high school). At each school we interviewed the principal and
toured the facility; at the majority of schools, we also spoke with teachers
and other school staff. We selected these schools on the basis of their
poverty rates, grade levels, size, status as schoolwide or targeted
assistance, diversity of student population, and other characteristics (see
table 2).

                                 Student
                                 racial     Poverty
 SchoolState Grade Number of   breakdowna    rate    Schoolwide   Targeted
             span  students                                     assistance
                                  (%)      (percent)
                               90 black

 1     N.C.  K-5   318         7 white     95        X

                               3 Hispanic
                               41 white

                               38 black
 2     N.C.  K-5   903                     48                   X
                               11 Hispanic

                               10 Asian
                               95 black
 3     N.C.  K-5   199                     79        X
                               5 white
                               59 black
 4     N.C.  PK-5  247                     69        X
                               39 white
                               47 black

 5     Penn. K-5   663         31 white    81                   X

                               20 Hispanic
                               39 black

 6     Penn. K-5   738         38 white    74                   X

                               23 Hispanic
 7     Penn. K-6   180         100 white   69        X
                               47 black

 8     Penn. 6-8   663         47 white    47                   X

                               5 Asian
 9     Penn. K-5   355         99 black    77        X
                               35 Hispanic

                               26 black
 10    R.I.  K-5   447                     85        X
                               26 white

                               13 Asian
                               55 Hispanic

                               25 black
 11    R.I.  6-8   891                     85        X
                               13 white

                               7 Asian
                               53 white
 12    Tenn. K-4   814                     49                   X
                               42 black
                               52 white
 13    Tenn. K,    366                     93        X
             5-6
                               42 black
 14    Tenn. K-8   164         100 white   63        X
                               50 white
 15    Tenn. K-5   535                     67        X
                               49 black
                               96 Hispanic
 16    Tex.  PK-5  884                     97        X
                               3 black
                               96 Hispanic
 17    Tex.  6-8   1024                    88        X
                               3 black
                               88 Hispanic
 18    Tex.  9-12  1664                    82        X
                               9 black
                               43 black

 19    Tex.  PK-2  284         40 white    68        X

                               17 Hispanic
                               36 white

 20    Tex.  6-8   572         33 black    71        X

                               29 Hispanic
                               61 Hispanic
 21    Tex.  PK-4  417                     79        X
                               36 white

aThe student racial breakdown for each school may not add up to 100 percent.
In some cases, schools had racial groups represented by only a few students
and these groups were not included. In other cases, the total did not equal
100 percent due to rounding.

We also surveyed the Title I directors in all 50 states in December 1999 and
January 2000 to obtain information about how the states measured educational
outcomes (including not only their state assessment systems but other
outcome measures), the level of detail at which they publicized this
information, their definitions of "adequate yearly progress" for schools and
districts under Title I, and how they monitored the Title I program at the
district and school levels. To obtain additional information on the data
states make available to the public, we used a standardized data collection
instrument to review the content of the 50 state education agencies'
Internet web sites for information on assessment, other outcome measures,
and other information at the school and district levels. To obtain
information on state outcome data, accountability systems, and Title I
program monitoring, we interviewed officials from five states and reviewed
documents they provided. We also reviewed the general literature on state
accountability systems, assessments, and educational outcomes.

To obtain information on current research on Title I, Title I services, and
schoolwide programs, we conducted a comprehensive literature search and
review. We interviewed Education researchers to obtain information on their
current studies and future evaluation plans.

We also interviewed staff in the Office of Compensatory Education (which
administers Title I), as well as federal officials representing federal
comprehensive technical assistance centers, regional education laboratories,
and equity assistance centers. In addition, we interviewed representatives
of major education associations and other experts. Finally, in addition to
reviewing Title I law and regulations, we also reviewed and analyzed other
Education documents, including program guidance, monitoring manuals,
training materials, and reports from completed monitoring visits.

Our work was done between April 1999 and February 2000 in accordance with
generally accepted government auditing standards.

Educators Believe Schoolwide Approach Can Provide Additional Opportunities
in High-Poverty Schools

Schoolwide and targeted assistance schools often provided similar
educational and support services. However, many of the educators we
interviewed told us they preferred the schoolwide approach because they
believed it provided disadvantaged students additional opportunities to
improve academic achievement, especially in high-poverty schools. Both
schoolwide and targeted assistance schools aimed special services at
children who needed extra help. However, according to an education survey,
schoolwide schools were more likely to use methods that increased
instructional time (such as before-school, after-school, and summer school
programs). Principals and teachers in high-poverty schools generally
supported the schoolwide approach, primarily because they believed it helped
them serve a greater number of students. At the same time, some educators
expressed concern that schoolwide schools need to take care that broader
efforts to improve student performance do not result in decreased services
to those low-performing students who need special assistance. Moreover,
evidence regarding the overall effectiveness of the schoolwide approach is
inconclusive.

Services but Varied in Service Delivery Methods

Title I schoolwide and targeted assistance schools often provided similar
educational services for disadvantaged students and their families. In the
schools we visited, both schoolwide and targeted assistance schools directed
specific services (such as tutoring) to students identified as needing
additional assistance. After-school programs were the most common type of
extended instruction, and 7 of the 21 schools we visited used Title I funds
to support after-school programs. Principals told us that the after-school
programs consisted primarily of tutoring for low-achieving students and were
generally available 1 or 2 days per week. Three of the schools that operated
after-school programs also provided the students with other services, such
as snacks and transportation home.

Most of the schools we visited--both schoolwide and targeted
assistance--used Title I funds for activities designed to more closely
involve parents in their children's education. These activities included
parent-teacher meetings, workshops, parent advisory councils, and
home-school liaisons. For example, one targeted assistance elementary school
we visited used Title I funds to support workshops that provided parents
with information about Title I services; ideas for helping their children
with school work at home; and summer learning activities for their children,
including a recommended reading list. A schoolwide school we visited used
Title I funds to provide tables and chairs so that parents can have lunch
with their children in a private area set aside near the school cafeteria.
Some schoolwide and targeted assistance schools also provided family
literacy programs for parents and students. For example, one school district
we visited operated a family literacy center in collaboration with a local
community college. Throughout the year, the center offered basic skills and
parenting workshops to any parent with children attending the local public
schools.

In several schools, Title I funds were also used to pay the salaries of
home-school liaisons who linked schools with parents, students, and the
community. In one targeted assistance elementary school, for example, the
principal told us that the home-school liaison made about 1,100 home visits
during the past school year to discuss with parents issues concerning their
children such as attendance, homework, and other classroom or behavioral
matters. At another schoolwide school we visited, the home-school liaison
not only made home visits but also held monthly meetings for parents at
school, regularly contacted parents to increase their involvement in school
activities, coordinated the activities of parent volunteers at the school,
and published a school newsletter. The liaison also established
relationships with community members, recruiting a local business owner who
then donated both time and money to the school.

While schoolwide and targeted assistance schools generally provided similar
services for their students, schoolwide schools tended to offer a wider
array of programs. In responding to Education's recent surveys, school
officials reported that schools using the schoolwide approach were more
likely than targeted assistance schools to offer programs with extended-time
learning opportunities (see table 3).

 Extended-time         Schoolwide schools       Targeted assistance schools
 programs              (percentage)             (percentage)
 Before school         18                       14
 After school          53a                      36a
 Weekend               7a                       3a
 Summer school         39                       36

aStatistically significant.

Source: Camilla Heid and Ann Webber, School-Level Implementation of
Standards-Based Reform: Findings from the Follow-Up Public School Survey on
Education Reform (Washington, D.C.: U.S. Department of Education, Planning
and Evaluation Service, 1999).

State, district, and school officials also told us that the schoolwide
approach allowed schools to adopt instructional methods they considered more
appropriate for all of their students. For example, one schoolwide
elementary school we visited used interactive computer software designed to
help teachers improve students' reading abilities. The computer program
helped teachers identify students' strengths and weaknesses so that students
were able to receive individualized assistance in areas of identified need.
Under the schoolwide approach, this school could use Title I funds to
purchase and implement the computer software for the whole class, without
targeting it specifically to the lower- achieving students. Similarly, staff
at one elementary school told us they preferred to keep all their students
together in the same classroom and move them in and out of flexible groups
as their achievement levels changed. These fluid groups allowed students to
obtain long- or short-term remedial help as needed within the classroom,
without being formally identified as Title I eligible. School officials also
described classroom settings in which students were placed in small groups,
combining strong and weak performers, and encouraged to help each other with
reading.

In contrast, targeted assistance schools were more likely to rely on
"pull-out" programs, using Title I funds to pay for teachers who provide
remedial instruction targeted to specific students. The pull-out option can
provide a convenient method for tracking who received Title I services and
how the funds were expended, allowing targeted assistance schools to more
easily demonstrate that funds were used for those students formally
identified as eligible under Title I. However, while the schoolwide schools
we visited sometimes used the pull-out strategy, they preferred other
approaches.

Under the schoolwide approach, schools can provide Title I services to more
students, some of whom would not be eligible under targeted assistance. In
the high-poverty schools we visited, many educators viewed this broader
focus as a primary advantage of the schoolwide approach. They explained that
most or all of their students suffered from educational disadvantages and
that the schoolwide approach allowed them to prevent as well as address
learning deficiencies. For example, one high-poverty, urban elementary
school we visited used its Title I funds to help implement broad changes in
the school curriculum. To help students improve their critical thinking
skills, the school integrated a science theme throughout the curriculum,
using new instructional methods such as hands-on activities and computers
with science-related software programs. The principal told us the schoolwide
approach provided more opportunities to develop strategies that could
benefit a greater number of students.

In contrast, targeted assistance schools more often relied on pull-out
programs to provide remedial assistance to low-achieving students. For
example, one Education survey found that 80 percent of targeted assistance
schools used pull-out programs, while 53 percent of schoolwide schools used
this approach.8 Although pull-out programs provide intensive tutoring to
low-performing students, they serve only a limited number of students each
school year and may exclude other low or marginally performing students in
need of additional instruction. One elementary school we visited, for
example, used a well-known pull-out program to provide individualized
reading instruction to low achievers in the early grades. In this school,
one reading specialist provided one-on-one tutoring to about 10 children
throughout the school year. While the number of students who receive these
services will vary according to the number of reading specialists available,
some district and school officials told us they did not adopt this specific
program because it is expensive and can serve only a small number of
students.

Many district and school officials also told us they favored schoolwide
programs because this approach offered other advantages, such as fostering
increased collaboration and greater flexibility. Before adopting a
schoolwide approach, schools are required to develop school plans that
describe strategies to improve academic achievement, both for low-achieving
students and for the rest of the school. School officials told us that
school staff participating in these planning activities worked together in
new ways, creating better coordination and communication between
instructional programs. For example, staff told us that during the planning
process, classroom strategies such as team teaching were developed because
of the teachers' willingness to try different approaches that may help
improve all students' academic performance.

In addition to serving more students, the schoolwide approach can give
schools greater flexibility in choosing educational strategies to improve
student achievement. For example, one rural elementary school we visited
implemented an approach that relied on reducing class size to give students
more individualized attention, while another school invested in professional
development to help teachers implement state standards. Other schools we
visited purchased computers and software that allowed students to progress
at their own pace. In all of these schools, principals told us that the
flexibility of the schoolwide program was an important factor in encouraging
them to adopt the changes.

Some district and school officials cautioned that schools that adopt the
schoolwide approach need to be careful to make sure that low-performing
students receive the extra help they need to improve their academic
performance and not be lost in the overall program. For several district and
state officials, this concern was especially important for schools with
relatively lower poverty rates and fewer disadvantaged children.

Nearly all of the schools GAO visited had high poverty rates and used the
schoolwide approach. Most schools chose the schoolwide approach because
school officials believed there were a number of advantages in doing so.
However, research on the overall effectiveness of the schoolwide approach,
as discussed in chapter 4, is inconclusive.

Most States Not Positioned to Hold Schools and Districts Accountable for
Outcomes of Disadvantaged Students

Although Education is responsible for overseeing states' implementation of
Title I, states continue to play a central role in program oversight and
accountability at the school and district levels. Historically, states have
been responsible for ensuring that districts and schools comply with Title I
programmatic and financial requirements, such as protecting the integrity of
federal funds and ensuring quality schoolwide planning that addresses the
needs of disadvantaged students. Under the oversight of Education, states
are expected to collect and report data on assessment outcomes and to hold
schools and districts accountable for these outcomes.

States varied dramatically in their efforts to oversee schools' and
districts' compliance with Title I's programmatic and financial
requirements. States focused their reviews on different requirements and
adopted different oversight methods. In addition, while some states made
annual oversight visits to schools and districts, others never made visits.
States also varied in how they reviewed schools' and districts' compliance
with Title I requirements and the degree to which they implemented Title I's
data collection and reporting requirements. Some states provided extensive,
detailed, and timely assessment data that school and district officials
found useful in improving instruction. Other states collected and reported
less information on assessments, including results for economically
disadvantaged students--information that is required beginning in the
2000-2001 school year. Furthermore, states' criteria for determining whether
schools and districts have performed satisfactorily have, with one
exception, been based solely on the performance of the student population as
a whole, without reference to the performance of specific subgroups of
children. Although the law does not require disaggregated data for yearly
progress assessments, lack of these data makes it difficult for states to
hold districts and schools accountable for the achievement of disadvantaged
students, and to discern whether achievement gaps between disadvantaged and
nondisadvantaged students are closing.

Provides Oversight and Technical Assistance

Under the oversight of Education, states play two distinct roles in ensuring
accountability for the Title I program. First, the states must ensure that
local schools and districts comply with the programmatic and financial
requirements that apply to them. Second, the states are responsible for
collecting and reporting data on assessment outcomes and devising criteria
to hold schools and districts accountable for these outcomes.

In both these roles, states have considerable discretion in how they carry
out their responsibilities and are generally free to develop their own
oversight methods and procedures. States also have discretion in designing
and gathering outcome data. The states set standards that determine the key
content children are expected to know. States can choose from a variety of
assessments--from multiple-choice standardized exams to collected portfolios
of student work--to measure student achievement as related to these
standards. States can also collect and report other educational outcome
measures. Finally, states have broad latitude in how they establish criteria
for determining whether schools and districts are performing
satisfactorily--for example, whether schools and districts will be evaluated
solely on the basis of assessment results or whether criteria will include
other measures of educational outcomes, such as graduation and dropout
rates.

Education is responsible for assessing whether states implement Title I
requirements related to assessments and sufficiently monitor schools' and
districts' implementation of Title I's financial and programmatic
requirements. Education also provides technical assistance to the states on
Title I regulations and on program quality issues.

States

States varied in how they focused their program oversight efforts. Like the
federal Department of Education, some states have approached their oversight
process from a technical assistance rather than strictly a compliance
perspective. For example, one state we visited used peer reviewers to help
schools and districts interpret data on teaching practices and other key
areas and placed less emphasis on compliance with financial requirements. In
contrast, another state focused more on district-level program
administration. Similarly, while the majority of states (34) conducted
reviews that covered multiple federal programs, some states focused their
review process on Title I alone.

States differed not only in their general approach to monitoring, but also
in the methods they used in their Title I reviews. These methods included
visits to districts and schools, reviews of state-required annual reports
and self-assessment documents, and reviews of districts' financial audits.9
For the most part, states depend on districts' data gathering and reporting,
even though they may be incomplete in some areas. Consequently, on-site
visits have generally been viewed as important because they provided state
officials with the opportunity to look directly at program implementation.
Forty-seven states reported that they included visits to school districts as
part of their monitoring procedures, and many of these states reported that
they relied solely on on-site visits to monitor certain key requirements.
For example, 26 states reported reviewing the components of schoolwide plans
only during on-site visits.

States varied dramatically in how frequently they conducted on-site reviews.
For example, while most states made on-site visits to schools and districts,
the average time between visits to districts ranged from 2 years or less
(for 6 states) to more than 7 years (for 17 states), as fig. 3 shows. Three
states reported that they made no on-site visits at all, while three states
reported that they visited all their school districts each year. The fact
that some states relied solely on site visits for oversight of key financial
and programmatic requirements, and made only very infrequent visits, raises
questions about whether these states conduct sufficient oversight.

Source: GAO survey of the 50 state education agencies.

Assessment Data Essential to Accountability

States determine the content covered in state standards and how student
performance is measured. States varied in the types of assessments used (for
example, multiple-choice tests, questions requiring written responses, or
alternative assessments such as portfolios of student work), the test
length, the frequency with which assessments are done, and the analytic
techniques used. States, districts, and schools also have some discretion
over whether certain students (such as those with limited English
proficiencies) are included in state-sponsored assessments. Nearly all
states have collected some type of assessment or outcome data; however, few
have obtained data specifically on certain groups of disadvantaged students.
Furthermore, although most states publicly reported assessment data for the
overall student population of the district or school, many have not reported
similar data on different groups of disadvantaged students.

Full Participation

States are responsible for developing content and student performance
standards as well as assessments to measure students' proficiencies. Every
state except one has adopted content standards and established assessments;
some states have established assessments in only reading and mathematics,
while others also require assessment in areas such as science and social
studies. In two states, the state allowed local districts to choose their
own assessments; however, if districts choose different assessments, results
may not be comparable.10 In the remaining 48 states, the state education
agency has required the same assessment on a statewide basis so results can
be compared across all districts and schools.

Because so much of schools' and districts' accountability hinges on state
assessments, it is important to ensure that these tests serve as valid
measures of student achievement. States have made choices about the type,
format, length, and difficulty level of the tests, and each of these choices
can have implications for the usefulness of the assessment. For example, an
assessment with a large number of multiple-choice items that focus on basic
skills can provide extensive information to help schools improve the
performance of low-achieving students, but might be less useful in promoting
continuous improvement in schools with higher levels of academic
achievement. An assessment that requires written responses to a smaller
number of detailed and more difficult questions might better measure
critical thinking skills but cover a narrower range of topics and provide
less information on what material low-achieving students have mastered.
Table 4 shows several important design issues, their implications for the
usefulness of an assessment in accomplishing various purposes, and how
states have responded to these issues.

   Issue          Options           Implications        States' responses
           States must choose
           either a              Norm-referenced
           norm-referenced       tests can provide
           assessment, which     national comparison
           may show a student's  information and may
           or school's results   be easier to use     27 states use a
           compared with a       because they are     norm-referenced test
                                 commercially         in at least one
 Type of   national average, or  available. However,  subject; 38 states
 test      a                     a                    use a
           criterion-referenced
           assessment, which     criterion-referenced criterion-referenced
           measures knowledge    test may do a better test in at least one
           in relationship to    job of measuring     subject.
           specific criteria     students' knowledge
           (such as may be       compared to specific
           incorporated in the   elements of the
           state standards).     state standards.
                                 States with very
                                 detailed and
                                 specific standards
                                 may find it
                                 difficult to cover
                                 the wide range of
                                 topics and also
                                 include enough items
                                 to provide a valid
                                 measure of knowledge Although we did not
                                 in any one area. If  specifically examine
           States must choose    state tests have too variations in test
 Test      the number of items   few items, they may  length as part of our
 length    they include in       not adequately test  study, it seems
           their assessments.    the students'        reasonable to expect
                                 knowledge of the     that length of
                                 range of topics.     assessments varies
                                 However, tests with  across states.
                                 an excessive number
                                 of items may result
                                 in students spending
                                 unnecessary time
                                 taking tests rather
                                 than being engaged
                                 in classroom
                                 instruction.
                                 Traditional,
                                 multiple-choice
                                 questions can be
                                 machine-scored, and
                                 the results of the
                                 scoring do not
                                 depend on the
                                 individual grading
                                 the test. Some       Only 4 states rely
           States must choose    experts prefer       exclusively on
           from more             formats that require traditional
           traditional           students to generate multiple-choice
           objective-style       their own response.  assessments in their
           questions (like       Assessments with     core subjects, 44
 Test      multiple-choice),     such questions,      included
 format    questions requiring   however, can be more open-response formats
           written responses,    costly and difficult in their assessments,
           and alternative       to score. Similarly, and 2 required the
           assessments (such as  portfolios may be    use of portfolios
           portfolios of         preferred because    statewide.a
           student work).        they reflect the
                                 student's work over
                                 time rather than on
                                 one test day;
                                 however, obtaining
                                 consistent,
                                 objective scoring of
                                 portfolios can also
                                 be time-consuming
                                 and expensive.
                                 A more difficult
                                 assessment that sets
                                 a high standard may
                                 motivate schools and
                                 districts to
                                 concentrate on
                                 challenging material
                                 but provide less     Only limited
                                 information to use   information is
                                                      available, but one
           States must choose    to help              study, which reviewed
 Difficultythe overall           low-achieving        tests used in several
 level     difficulty level of   students improve. If states, found wide
           the assessment.       the lowest-achieving variation in
                                 students can answer
                                 only a very few      difficulty level
                                 questions correctly, among 8th grade
                                 the test cannot help mathematics tests.
                                 teachers distinguish
                                 between material
                                 these students have
                                 mastered and the
                                 material they have
                                 not.

aAs noted previously, two states leave the format of the assessment up to
local school districts. For more detailed information on the use of
open-response format questions on state assessments, see Jerald, p. 62.

Both federal legislation and Education's guidance have strongly emphasized
that, to the extent practicable, all students should participate in state
assessments. If certain students do not take the test, it impairs the
state's ability to hold schools and districts accountable for those
students' achievement. However, in some circumstances it may be considered
appropriate to exempt an individual student or group of students from
state-sponsored assessments. Students with severe disabilities, for example,
may be working toward different educational goals than the general
population, and thus it may not be considered appropriate to measure their
progress against state standards.11

States varied in how they attempted to strike the balance between including
as many students as possible and making sure that students were not
subjected to inappropriate assessments. In our survey, 45 states reported
that some students were exempted; however, over half (24) were unable to
provide any information on precisely how many students were exempted.12
Similarly, of the 41 states that reported district- or school-level
assessment data through their Internet site, only 20 provided any additional
information about the extent to which students participated in or were
excluded from assessments, and even fewer states (4) provided information
that specified the reasons for student exclusions. Without such information,
state officials and outside observers cannot determine whether state
assessment results reflect achievement for all students or only for a
selected group. Furthermore, if different numbers or types of students are
excluded in different schools, school-to-school comparisons may become less
meaningful.13

Students; Few States Collected Other Outcome Data for These Students

States varied in the type and amount of data they collected from schools and
districts on assessment results and other outcome measures for disadvantaged
students. Some states collected information on assessment results
specifically of disadvantaged students. Fewer states indicated that they
collected information on other educational outcomes, such as dropout or
attendance rates of these students.

Title I requires states to collect assessment results separately for
specific subgroups of students when final assessments are in place; however,
many states have not yet collected all these data. As of fall 1999, only 17
states--less than half--had disaggregated their assessment results by all
six required categories (gender, racial and ethnic group, migrant status,
disability status, English proficiency status, and economic status). The
most frequently omitted category was migrant status, perhaps reflecting the
limited outcome information generally available on migrant students.14 Many
states also did not disaggregate assessment data by economic status.

In addition, 38 states reported that they gathered other outcome data not
required by Title I legislation. The most commonly used measures were
student attendance, dropout, and graduation rates, as shown in fig. 4. Only
4 states indicated that they disaggregated these data by the six categories
required for assessment data; 14 states did not disaggregate their other
outcome data by any of these categories (see fig. 5).

Source: GAO survey of the 50 state education agencies.

Data, as of Fall 1999

Source: GAO survey of the 50 state education agencies.

Data for Various Subgroups of Disadvantaged Students

While most states report assessment results to the public, many states do
not separately report data for various subgroups of disadvantaged students.
In the 1994 Title I reauthorization legislation, the Congress recognized
that publicity could be an important tool for holding schools and districts
accountable for student performance. By making test results available to the
public, states give parents, educators, and the community information on how
their schools are doing and provide an incentive for schools and districts
to improve their results. Nearly all states (48) said they made their
assessment data available to the public; fewer (34) reported their other
outcome data. State survey responses indicated that the majority of states
report assessment information to the public through press releases and by
making information available on states' Internet web sites.

While most states publicly reported assessment and other outcome data for
the overall population of the district or school, some states did not report
such data by specific groups of disadvantaged students. For example, one
state official told us that although the state keeps disaggregated
assessment data, his office does not use them. States identified the
Internet as one of the most widely used means of publicizing assessment
results, and it is a natural choice for distributing the more voluminous
information from disaggregated assessments. However, very few states used
the Internet to report disaggregated data, as fig. 6 illustrates.

Source: GAO review of the Internet web sites of the 50 state education
agencies.

Assessment and other outcome data clearly play an important role in holding
schools and districts accountable for student performance in general and for
disadvantaged students in particular. However, staff at many of the
districts and schools we visited also emphasized how these data helped them
improve instruction. District officials and principals told us that they
used assessment and other outcome measures to focus attention and effort on
improving student achievement. For example, several schools we visited
displayed signs, banners, or bulletin boards showing their progress in
improving assessment results. District and school staff also reported using
assessment information to identify and address weaknesses in instruction.
One school we visited used a detailed analysis of the state assessment to
identify key mathematics concepts students needed to work on, and set aside
a 15-minute period each day for students to play special math games to
develop and reinforce these specific skills.

School and district officials emphasized that data must be detailed and
timely to be useful in improving instruction. However, the information
reported by states was sometimes limited or late. School and district staff
found assessment and other outcome data especially useful when broken down
by teacher or objective, but only 12 states provided assessment data by
teacher, 23 by objective, and 11 by individual test item. In addition, in
two of the five states we visited, district and school officials told us
that the results of the state assessments were not provided to them in a
timely manner--making it difficult for them to make significant
instructional changes to address students' needs.15

Overall Student Population, Limiting Accountability for Disadvantaged
Students

In addition to requiring that states collect and report data on educational
outcomes, Title I requires states to establish criteria to evaluate whether
schools and districts are performing satisfactorily--that is, making
"adequate yearly progress" toward improving student achievement. States have
chosen different ways to define criteria for determining adequate yearly
progress. Some states defined their criteria on the basis of a fixed-target
assessment score for all schools, while others required schools to show
improvement compared with previous results. For example, one state required
that 75 percent of all students in grades 4 and 8, and 85 percent of all
students in grade 11, meet proficiency levels on the state assessment. In
another state, for a school to meet the criteria, at least 50 percent of the
students had to score at a satisfactory level in mathematics and reading for
all grades tested for 2 consecutive years, or the number of students scoring
at a satisfactory level had to increase by at least 5 percent from one year
to the next. Although states are not required to establish their final
adequate yearly progress criteria until the 2000-2001 school year, a
majority of states reported these criteria in response to our survey of the
50 states.

Several states have very lengthy and complex adequate yearly progress
criteria that can be difficult for school and district staff to understand
and interpret. Officials in two districts we visited, in two different
states, told us that they had encountered difficulties interpreting their
assessment results in light of the states' expectations of performance.
Other criteria we reviewed also appeared confusing or vague. If the school
and district staff do not fully understand the criteria on which their
performance will be judged, it is likely that they will have difficulty
developing strategies for improvement. In addition, states have--with only
two exceptions--defined adequate yearly progress solely in terms of
assessment results, without including other educational outcome measures
such as graduation, attendance, or dropout rates. These other outcome
measures are also important in considering schools' and districts'
performance, yet are generally not included in the yearly progress criteria.
Because adequate yearly progress is a new concept, some states have had
difficulty developing their criteria.

Annual progress criteria are not required by law to be disaggregated by
subgroups of students, and only one state has chosen this approach. As a
result, schools and districts may not be held accountable for closing
achievement gaps over time among subgroups of students but only for
improving the average overall student achievement. Because subgroups of
students can have poor results on assessments even if the overall student
body is performing at higher levels, disadvantaged children may fall further
behind even in schools making satisfactory progress according to states'
criteria. Basing progress criteria on each student subgroup may
unnecessarily complicate states' evaluation of school and district
performance. However, in Texas, the state has included two specific
subgroups--economically disadvantaged and race/ethnicity--in its definition
of adequate yearly progress to improve accountability for the achievement of
disadvantaged students.

Evaluations of Title I and Schoolwide Programs Have Been Limited

Nationally representative evaluations of the effectiveness of Title I are
needed because state assessment data on the progress of disadvantaged
students are currently limited and state data on overall student achievement
vary in quality and type. However, research on the effectiveness of the
Title I program is also currently limited, particularly for the more recent
schoolwide approach. Previous research suggests that Title I may have a
slightly positive effect on student achievement in high-poverty schools.
Research on the general effectiveness of the schoolwide approach for
improving student achievement in high-poverty schools is inconclusive.

Conducting research on the effectiveness of Title I presents challenges,
including the difficulty of discerning the effect of Title I on the regular
school program. Education's planned improvements in data collection show
some promise for facilitating better research, but these efforts will need
to be expanded to ensure that specific educational programs and services can
be linked to achievement--both for disadvantaged students and all students.

Effectiveness

Title I may have a limited positive impact on overall student achievement in
high-poverty schools, according to the two most recent congressionally
mandated studies completed by Education. Education's most recent national
assessment generally concluded that, since 1992, national performance in
reading and mathematics has improved for 9-year-olds in the highest-poverty
public schools--those schools most likely to receive Title I funds.16
However, the report also noted that a substantial achievement gap remains
between students in the highest- and lowest-poverty schools. Education
concluded in its earlier mandated study that Title I may have helped
educational achievement but did not bring the participating students up to
the level of their classmates.17 Other research findings are based on data
that are not nationally representative; these studies provide either weak
positive evidence or mixed evidence about the effectiveness of Title I.

Because of data and methodological limitations, neither the national
evaluations nor smaller studies have been able to support any conclusions on
whether schoolwide programs are effective. The earlier national evaluation
did identify five high-performing, high-poverty schools, three of which had
adopted schoolwide programs; however, because there were so few schools,
researchers could not draw conclusions on the basis of this finding.
Similarly, another Education report on four case studies found that while
administrators and faculty in schoolwide schools praised the schoolwide
option, "the schools were not producing remarkable levels of student
achievements."18 Similar to what we discussed in chapter 2, evidence on the
effectiveness of schoolwide programs was often based solely on the assertion
of the school's principal or Title I coordinator, according to a recent
research synthesis.19 Several studies of schoolwide programs within
individual school districts (including Philadelphia, Houston, and
Minneapolis) have provided mixed results. For example, in Minneapolis the
gap in math achievement between lower achieving and other students was
smaller in schoolwide than in targeted assistance schools, but the opposite
was true in Houston.20 Education has an ongoing study of Title I programs
which may shed some light on factors affecting student achievement in
high-poverty schools; however, it has a limited scope and is not nationally
representative.21

Effectiveness

The lack of comprehensive, detailed, and nationally representative data on
Title I schools and services has made program evaluation especially
challenging. For example, researchers have been generally unable to clearly
link Title I services to student achievement because the available data are
not sufficiently comprehensive. As we found in our site visits, schools use
the flexibility they have under Title I to spend funds on a wide variety of
goods and services, some of which are similar or identical to goods and
services offered at other schools but not supported with federal funds. This
makes it even more difficult to distinguish the effect of Title I from the
effect of the overall educational program of the school.

Researchers also face challenges linking differences in student achievement
to Title I rather than to other factors such as differences in the resources
available to schools. To link Title I to student achievement, researchers
need sufficiently detailed data to either select comparable schools to study
or to statistically estimate the effect of program services on student
achievement separately from other external factors. However, available data
do not contain the information needed to select comparable schools or to
statistically adjust for other factors. For example, Education's major
database on U.S. public schools, the Common Core of Data (CCD), does not
currently provide information on whether a school receives Title I funds or
has a schoolwide program; the CCD provides only incomplete information on
school poverty rates. Without these data, researchers cannot use the
database to identify schoolwide and targeted assistance schools with similar
poverty rates.

Other challenges exist in evaluating the effectiveness of educational
services in both schoolwide and targeted assistance programs. One is
selecting an appropriate measure of student achievement. And, as discussed
in chapter 3, measures of student achievement are not comparable across
states.

In addition, no clear consensus has emerged on what criteria should be used
to determine whether Title I is effective. For example, the results of one
of the earlier evaluations indicated that Title I students learn at
approximately the same rate as their more advantaged peers, but that gaps in
initial achievement tend to remain.22 Some individuals have interpreted
results from the Prospects Study as evidence that the Title I program may
not be effective. Others disagree, pointing to work which suggests that
without intervention, gaps in achievement between disadvantaged and other
students tend to grow over time.

To meet some of these challenges, Education plans to expand its data
collection to facilitate research on the effectiveness of Title I, including
schoolwide programs. For example, Education officials plan to collect
information on the Title I and schoolwide program status of the nation's
schools as part of its CCD. This information could be combined with student
assessment and services data to facilitate more comprehensive analyses of
Title I effectiveness. In addition, Education is planning to expand its
examination of the results of state assessments--including disaggregated
data--to assess students' performance. While these assessment data will not
be comparable across states, an analysis of student achievement on state
assessments is important because it represents the primary criterion the
Title I legislation applies to define school success.

Education's efforts to extend its data collection and analysis show promise
for enabling better research, but these actions will still not fully address
the challenges to determining program effectiveness. Education's analysis of
disaggregated state data, for example, may provide helpful information in
the future. However, the value of this type of analysis will be limited
because not all states have fully disaggregated their assessment data.
Moreover, Education's planned efforts will only partially address the need
for more comprehensive data on school characteristics, resources, and
services provided. For example, Education's plans to add Title I information
to its CCD will be valuable, but data quality and consistency issues (such
as missing data on poverty and other key variables) must be addressed.23
Similarly, detailed data on educational services, from a sufficiently large
number of both schoolwide and targeted assistance schools, will be needed to
assess program effectiveness.

Conclusions, Recommendations, Matter for Congressional Consideration, Agency
Comments, and Our Evaluation

Educating children has traditionally been largely a state and local
responsibility. However, through the Department of Education, the federal
government has a role in supporting quality education for disadvantaged
children and the responsibility for overseeing the nearly $8 billion annual
federal investment in the Title I program. In its program monitoring,
technical assistance, and research roles, Education can help the states hold
schools and districts accountable for the academic achievement of
disadvantaged students. Since the 1994 reauthorization of Title I, states
have made some progress in improving their systems for holding districts and
schools responsible for the academic achievement of all students, including
disadvantaged students. However, some areas are still in need of attention.
Education, in its leadership role, has provided some assistance to states in
developing standards, assessments, and criteria for adequate yearly
progress; however, a majority of states appear to need additional help, for
example, in collecting and reporting disaggregated assessment data.

Beginning in the 2000-2001 school year, states are required to collect and
report assessment data disaggregated by subgroups of students, including the
economically disadvantaged. Only about one-third of the states currently
collect this information, and it is unlikely that all of the states will
meet the deadline for collecting and reporting disaggregated assessment
data. Without disaggregated data, test results for the whole student
population can mask the results of disadvantaged students and prevent states
and districts from identifying schools that may not be meeting the
educational needs of disadvantaged students.

While the majority of states provided us with their criteria for adequate
yearly progress, some of the criteria were confusing or vague. Moreover,
only one state's criteria for adequate yearly progress are based on
disaggregated data of subgroups of students. These criteria are needed to
hold schools and districts accountable for improving the performance of
disadvantaged students and to discern whether achievement gaps between
disadvantaged students and nondisadvantaged students are narrowing over
time.

Finally, Education's research can provide states, districts, and schools
with important information to help them improve instruction. Comprehensive
evaluation efforts are needed to determine which services and service
delivery methods are most effective in closing the achievement gaps among
students.

We make two recommendations to the Secretary of Education. First, Education
should conduct additional activities to facilitate the exchange of
information and best practices among states so they can identify ways to
improve the timeliness and specificity of their assessment data, the
collection and reporting of disaggregated assessment data, and the clarity
of their criteria for adequate yearly progress. Second, Education should
implement additional measures to improve research on the effectiveness of
specific services in both schoolwide programs and targeted assistance
schools. Such measures could include expanding and improving current data
collection efforts so that comprehensive analyses could be conducted linking
program characteristics to services and student outcomes, and/or developing
an evaluation design for a study or set of studies of educational services
that would include national representation of both schoolwide and targeted
assistance schools.

To hold schools and districts accountable for improving the performance of
disadvantaged students and to help educators, parents, and others discern
whether achievement gaps between disadvantaged and nondisadvantaged students
are closing, the Congress should consider requiring that states' criteria
for progress, as expressed in their definitions of adequate yearly progress,
apply specifically to disadvantaged children, as well as to the overall
student population.

We provided a draft of this report to the Department of Education for review
and comment. Without commenting on the recommendations, Education agreed
that annual progress criteria based solely on the performance of the student
population as a whole may be inadequate to hold schools accountable for the
performance of disadvantaged students. However, Education said that such
accountability does not necessarily require separate criteria for each of
the six subgroups for which states currently must report assessment data. We
did not say that separate criteria should be used for each of the student
subgroups and we have modified the report to make this clear. Education also
commented that some states may respond to including specific groups of
disadvantaged students in their progress criteria by lowering standards. We
continue to believe that it is important for states to include subgroups in
their adequately yearly progress criteria because this information will help
them hold schools and districts accountable for the performance of
disadvantaged students.

Education also emphasized in its comments that states are not required to
publicly report disaggregated data until the 2000-2001 school year. We had
pointed this out several times in the report but had also found that many
states are not yet positioned to meet this requirement. The large number of
states that have not fully implemented assessments and disaggregated the
results raises serious concerns about whether states will be able to provide
the detailed data necessary to ensure accountability for the performance of
disadvantaged students by the 2000-2001 school year.

The Department agrees with our position that it is difficult to separate the
effects of Title I from the effects of state and local efforts. Education
said that reviewing how schoolwide programs spend Title I funds would be
insufficient to evaluate program effectiveness. We did not recommend this
approach. Rather, we recommended that Education improve research on the
effectiveness of specific services designed to address the needs of
disadvantaged students. Research that links these services to educational
outcomes can help determine which services are most effective for raising
student achievement. This approach does not preclude other types of studies,
such as Education's National Assessment of Title I, that examine the overall
performance of students in high-poverty schools. However, research on the
effectiveness of specific services will be especially useful to principals
and school district officials, in our opinion, because it can help them
better leverage their Title I dollars to promote student achievement.

We reported that the lack of comprehensive, detailed, and nationally
representative data on Title I schools and services has made program
evaluation especially challenging. Education said that the agency has
gathered detailed information on Title I through several nationally
representative studies, including the Public School Surveys on Education
Reform, the Study of Education Resources and Federal Funding, and the
National Longitudinal Survey of Schools. We reviewed these data and found
that the information they provide is valuable, but did not contain
sufficient detail for researchers to assess the effectiveness of specific
services. For example, while Education has collected some data on services,
its surveys generally did not include sufficient information to allow
researchers to control for differences in resources and student populations
across schools. Education's current efforts to extend its data collection
and analysis show promise. However, additional steps will be needed to
improve research on the effectiveness of federally supported services to
address the educational needs of disadvantaged students.

Finally, Education provided other comments regarding technical aspects of
the report, which we incorporated where appropriate. (See app. I for
Education's comments.)

Comments From the Department of Education

Major Contributors to This Report

Harriet C. Ganson, Assistant Director, (202) 512-9045
Sarah L. Glavin, Senior Economist, (202) 512-7180
Timothy E. Hall
Sheila R. Nicholson
Ellen K. Schwartz
John G. Smale
James P. Wright

(104972)

Table 1: Examples of Key Title I Requirements 20

Table 2: Characteristics of Schools Visited 24

Table 3: Principals Reporting Use of Extended-Time Learning Opportunities 28

Table 4: Assessment Design Issues 37

Figure 1: Distribution of Schools by Poverty Level and Title I Status 15

Figure 2: Example of District-School Title I Allocation Process 17

Figure 3: Frequency of States' On-Site Monitoring 35

Figure 4: Other Outcome Measures Gathered by States 40

Figure 5: Number of States That Disaggregate Assessment and
Other Outcome Data, as of Fall 1999 41

Figure 6: Number of States Reporting Disaggregated Assessment
Data Through the Internet 42
  

1. The Improving America's Schools Act of 1994 (P.L. 103-382).

2. According to the 1994 Title I legislation, states must develop criteria
to ascertain whether schools and/or districts are making adequate yearly
progress. States may choose different ways to define these criteria, such as
using a fixed target score for all schools or requiring schools to show
improvement over time compared with previous results.

3. If a district chooses to fund any school with a poverty rate below 35
percent, each school selected for Title I funding must receive a minimum
amount per low-income student under the "125 percent rule." This minimum
amount is equal to the total district allocation under Title I divided by
the total number of low-income students in the entire district, multiplied
by 125 percent.

4. District and school officials may have additional flexibility in
determining allocations for two reasons: (1) they can choose from different
measures of poverty in determining school poverty rates; and (2) they may be
able to affect the measured poverty rate in specific schools. Under Title I
rules, as long as the same measure of poverty is used across schools,
districts may use poverty rates derived from a number of data sources,
including children eligible for free or reduced-price lunches (the most
common measure), children eligible to receive medical assistance under the
Medicaid program, or other measures. Some districts and schools can also
take steps to increase the accuracy of the measured poverty rate. In one
district we visited, the district sponsored billboards in low-income
neighborhoods urging parents to sign their children up for free or
reduced-price lunches. Similarly, at a school in another district, the
principal told us that she made special efforts to get students to return
their applications for the lunch program so that her school would exceed the
75 percent threshold and receive Title I funds.

5. School poverty rates were taken from Education's Common Core of Data and
schoolwide data were obtained from GAO's 50-state survey.

6. Education's site visits focus on how the states are fulfilling their
responsibilities. States also make site visits to districts and schools to
monitor compliance with key Title I requirements.

7. The four states receiving the most money are reviewed more frequently.

8. Heid and Webber.

9. Under the Single Audit Act, recipients of federal funds may be required
to have an annual single audit of the funds received from all federal
programs to ensure that federal dollars are spent in compliance with
applicable requirements.

10. In one of these states, a large majority of districts have chosen to
administer the same test. For more information about these states' systems,
see Craig D. Jerald, "The State of the States," Education Week: Quality
Counts 2000 (January 13, 2000), p. 62.

11. Federal law expressly allows some students to be excluded from
assessment. The decision to exclude a student with a disability is left to
the student's Individualized Education Program team, which comprises the
student's parents, teachers, district officials, and others. Students with
disabilities may also (at the discretion of the team) receive
accommodations--such as being allowed extra time or having part of the
assessment presented orally--to help them participate in the assessment.

12. Forty states excluded some students with limited English proficiency,
and 40 states excluded some students with disabilities. In addition, 18
states sponsored alternative assessments for students with limited English
proficiency (such as a Spanish-language version), and 18 states sponsored
alternative assessments for students with disabilities.

13. Although we did not collect information about the extent of
accommodations for students with disabilities, a recent research study
confirms that similar issues arise when states, schools, and districts
differ in their practices in granting accommodations. For more information
about this issue, see Anne Lewis, 1998 CRESST Conference Proceedings,
Comprehensive Systems for Educational Accounting and Improvement: R&D
Results, National Center for Research on Evaluation, Standards, and Student
Testing (CRESST), CSE Technical Report 504 (Los Angeles: University of
California, June 1999).

14. For more information on these issues, see Migrant Children: Education
and HHS Need to Improve the Exchange of Participant Information
(GAO/HEHS-00-4, Oct. 15, 1999).

15. In a number of districts we visited, district officials had implemented
assessment programs of their own to better meet their needs for detailed and
timely achievement data. Because these district-administered assessments
were scored quickly and analyzed in detail, schools could identify problems
and tailor solutions to the individual child, classroom, or school. Nine
states made at least some questions from previous assessments available
through their Internet site.

16. U.S. Department of Education, Office of the Under Secretary, Planning
and Evaluation Service, Promising Results, Continuing Challenges: The Final
Report of the National Assessment of Title I (Washington, D.C.: 1999).

17. Abt Associates, Inc., Prospects: Final Report on Student Outcomes ,
prepared under contract to the Planning and Evaluation Service, U.S.
Department of Education (Washington, D.C.: U.S. Department of Education,
Apr. 1997).

18. Sam Stringfield, Mary Ann Millsap, Rebecca Herman, and others, Urban and
Suburban/Rural Special Strategies for Educating Disadvantaged Children:
Final Report , prepared for the U.S. Department of Education (Apr. 1997).

19. Kenneth K. Wong and Stephen J. Meyer, "Title I Schoolwide Programs: A
Synthesis of Findings From Recent Evaluation," Educational Evaluation and
Policy Analysis (Summer 1998), pp. 115-36.

20. Kenneth K. Wong, Gail L. Sunderman, and Jaekyung Lee, "Redesigning the
Federal Compensatory Education Program: Lessons From the Implementation of
Title I Schoolwide Projects," National Center on Education in the Inner
Cities Review , Temple University (Aug. 1997).

21. See Brenda Turnbull, Megan Welsh, Camilla Heid, and others, The
Longitudinal Evaluation of School Change and Performance (LESCP) in Title I
Schools: Interim Report to Congress , prepared for the U.S. Department of
Education (July 1999).

22. Abt Associates, Inc., Prospects: Final Report on Student Outcomes .

23. U.S. Department of Education, National Center for Education Statistics,
Evaluation of the 1996-97 Nonfiscal Common Core of Data Surveys Data
Collection, Processing, and Editing Cycle (Washington, D.C.: Feb. 1999).
*** End of document. ***