06PA 12.1
PERFORMANCE AUDIT REPORT
K-12 Education: Reviewing Free-Lunch Student
Counts Used as the Basis for At-Risk Funding, Part I
A Report to the Legislative Post Audit Committee
By the Legislative Division of Post Audit
State of Kansas
November 2006
Legislative Post Audit Committee
Legislative Division of Post Audit
THE LEGISLATIVE POST Audit Committee and
its audit agency, the Legislative Division of Post
Audit, are the audit arm of Kansas government.
The programs and activities of State government
now cost about $11 billion a year. As legislators
and administrators try increasingly to allocate tax
dollars effectively and make government work more
effi ciently, they need information to evaluate the
work of governmental agencies. The audit work
performed by Legislative Post Audit helps provide
that information.
We conduct our audit work in accordance
with applicable government auditing standards
set forth by the U.S. Government Accountability
Offi ce. These standards pertain to the auditor's
professional qualifi cations, the quality of the audit
work, and the characteristics of professional and
meaningful reports. The standards also have been
endorsed by the American Institute of Certifi ed
Public Accountants and adopted by the Legislative
Post Audit Committee.
The Legislative Post Audit Committee is a
bipartisan committee comprising fi ve senators and
fi ve representatives. Of the Senate members, three
are appointed by the President of the Senate and
two are appointed by the Senate Minority Leader.
Of the Representatives, three are appointed by the
Speaker of the House and two are appointed by the
Minority Leader.
Audits are performed at the direction of
the Legislative Post Audit Committee. Legislators
or committees should make their requests for
performance audits through the Chairman or any
other member of the Committee. Copies of all
completed performance audits are available from
the Division's offi ce.
The Legislative Division of Post Audit supports full access to the services of State government for all
citizens. Upon request, Legislative Post Audit can provide its audit reports in large print, audio, or other
appropriate alternative format to accommodate persons with visual impairments. Persons with hearing
or speech disabilities may reach us through the Kansas Relay Center at 1-800-766-3777. Our offi ce
hours are 8:00 a.m. to 5:00 p.m., Monday through Friday.
LEGISLATIVE POST AUDIT COMMITTEE
Senator Les Donovan, Chair
Senator Anthony Hensley
Senator Nick Jordan
Senator Derek Schmidt
Senator Chris Steineger
Representative John Edmonds, Vice-Chair
Representative Tom Burroughs
Representative Peggy Mast
Representative Bill McCreary
Representative Tom Sawyer
LEGISLATIVE DIVISION OF POST AUDIT
800 SW Jackson
Suite 1200
Topeka, Kansas 66612-2212
Telephone (785) 296-3792
FAX (785) 296-4482
E-mail: LPA@lpa.state.ks.us
Website:
http://kslegislature.org/postaudit
Barbara J. Hinton, Legislative Post Auditor
LEGISLATURE OF KANSAS
LEGISLATIVE DIVISION OF POST AUDIT
800 SOUTHWEST JACKSON STREET , SUITE 1200
TOPEKA , KANSAS 66612-2212
TELEPHONE (785) 296-3792
FAX (785) 296-4482
E-
MAIL
: lpa@lpa.state.ks.us
November 6, 2006
To: Members, Legislative Post Audit Committee
Senator Les Donovan, Chair
Representative John Edmonds, Vice-Chair
Senator Anthony Hensley
Representative Tom Burroughs
Senator Nick Jordan
Representative Peggy Mast
Senator Derek Schmidt
Representative Bill McCreary
Senator Chris Steineger
Representative Tom Sawyer
This report contains the fi ndings, conclusions, and recommendations from our
completed performance audit, K-12 Education: Reviewing Free-Lunch Student Counts Used
as the Basis for At-Risk Funding, Part I .
The report also contains several appendices including a detailed description of the
methodology we used to determine free-lunch eligibility for our random sample, a summary
of the major free-lunch application reviews conducted by Department of Education staff,
and a comparison of school district enrollment counts to comparable U.S. Census Bureau
estimates for the 2003-04 school year.
The report includes several recommendations for the Department of Education
and for appropriate legislative committees. We would be happy to discuss these
recommendations or any other items in the report with any legislative committees,
individual legislators, or other State offi cials.
Barbara J. Hinton
Legislative Post Auditor
READER'S
GUIDE
Get the Big Picture
Read these Sections and Features:
1. Executive Summary - an overview of the questions we
asked and the answers we found.
2. Conclusion and Recommendations - are referenced in
the Executive Summary and appear in a box after each
question in the report.
3. Agency Response - also referenced in the Executive
Summary and is the last Appendix.
Helpful Tools for Getting to the Detail
* In most cases, an " At a Glance " description of the agency or
department appears within the first few pages of the main report.
* Side Headings point out key issues and findings.
* Charts/Tables may be found throughout the report, and help provide
a picture of what we found.
* Narrative text boxes can highlight interesting information, or
provide detailed examples of problems we found.
* Appendices may include additional supporting documentation, along
with the audit Scope Statement and Agency Response(s) .
Legislative Division of Post Audit
800 SW Jackson Street, Suite 1200, Topeka, KS 66612-2212
Phone: 785-296-3792 E-Mail: lpa@lpa.state.ks.us
Web: www.kslegislature.org/postaudit
EXECUTIVE SUMMARY
i
Legislative Division of Post Audit
November 2006
The National School Lunch Program provides free lunches to
students who meet poverty thresholds or participate in designated
programs . Children from families with incomes below 130% of the poverty
level qualify for free lunches. In addition, students who participate in programs
such as Food Stamps and Temporary Assistance for Needy Families, or
who've been identifi ed as migrant, homeless, or runaway, also can receive free
lunches.
Kansas distributes at-risk funding based on the number of
students eligible for free lunches in each district . Districts report a count
of free-lunch students along with other enrollment data to the Department
of Education. Department staff perform a comprehensive enrollment audit
during which they remove ineligible free-lunch students because of errors on
the applications, or because the students weren't enrolled and attending on
September 20th. Based on the at-risk weighting in the school funding formula,
school districts received $822 in at-risk funding for each free-lunch student in
2005-06. Because the at-risk weighting increases in the future, that amount
will grow to $2,021 by 2008-09.
In 2005-06, Kansas districts received almost $111 million in at-risk
funds for about 135,000 students identifi ed as eligible for free lunches .
That amount was more than double the previous year's amount because of
increases in at-risk funding per student.
About 17% of free-lunch students in our Statewide random sample
were ineligible, costing the State an additional $19 million in at-risk funds .
Of the 500 free-lunch students in our random sample, 85 students weren't
eligible, primarily because households under-reported their income. Many
households are able to under-report their income and still receive free lunches
because federal law requires school district offi cials to accept their applications
at face value. Projecting our results to all free-lunch students, we estimate
the State paid almost $19 million in at-risk funds for nearly 23,000 ineligible
students in 2005-06.
Based on our survey of district offi cials, about 6,900 students
Statewide may have been eligible for free lunches but their families didn't
apply. According to district offi cials, most eligible families who don't apply are
either too embarrassed to do so, or are concerned about the confi dentiality of
their applications.
Question 1: Does the Count of Free-Lunch Students Used for
At-Risk Funding Accurately Refl ect the Number of
Students Who Are Eligible for the Program?
EXECUTIVE SUMMARY
LEGISLATIVE DIVISION OF POST AUDIT
.................. page 4
.................. page 5
.................. page 7
.................. page 8
................ page 10
Overview of the National School Lunch Program and
Distribution of At-Risk Funding in Kansas
ii
EXECUTIVE SUMMARY
Legislative Division of Post Audit
November 2006
The free-lunch counts used for at-risk funding also may include
a number of students the Legislature didn't intend to fully fund . The
Department has developed an alternative at-risk funding application so
districts can get at-risk funds for students attending schools that don't
provide lunches. This includes a number of non-traditional students for
which at-risk funding may not have been intended. For example, our review
of eight alternative schools that don't serve lunch showed that 127 of the 319
free-lunch students in those schools were age 20 or older.
Districts also receive the full amount of at-risk funding for part-time
students (primarily kindergartners) because the State doesn't prorate the
funding. For example, in the Topeka school district, at-risk funding based on
an FTE count rather than a headcount would have saved the State $340,000
in 2005-06. Because of future increases in at-risk funding, those savings
could grow to about $840,000 by 2008-09.
We identifi ed additional problems with the Department's free-
lunch reviews that, if addressed, could produce a more accurate count .
In 2005-06, the child nutrition team didn't report roughly 1,850 ineligible
free-lunch students it knew about to the Department's fi scal auditors.
Auditors could have removed them from the at-risk count, saving the State
$1.5 million in at-risk funding. Also, the Department's fi scal auditors didn't
audit every school in the six largest districts, potentially missing about 100
ineligible students in 2005-06. Lastly, because the Department's two teams
don't coordinate their reviews, the same free-lunch application may be
reviewed several times by Department staff.
Question 1 Conclusion
Question 1 Recommendations
For 2003-04, Kansas had 54,000 more free-lunch students than
adjusted U.S. Census estimates would suggest . The Census data
suggest that approximately 76,000 children in Kansas were at or below
130% of the federal poverty level in 2003-04, compared with almost 130,000
free-lunch students.
The free-lunch count is signifi cantly higher than the adjusted
Census estimate, primarily because the count includes many ineligible
students . Based on our results from Question 1, we estimated that 22,000
of the almost 130,000 free-lunch students in 2003-04 weren't eligible, which
is almost half the 54,000-student difference between the free-lunch count
and the Census Bureau estimate. Other factors include what age groups are
counted, and whether foster care children are included.
Question 2: How Does the Number of Free-Lunch Students
Reported by Districts Compare with Poverty Estimates Compiled
By the U.S. Census Bureau?
................ page 11
................ page 13
................ page 16
................ page 16
................ page 18
................ page 19
EXECUTIVE SUMMARY
iii
Legislative Division of Post Audit
November 2006
The Census Bureau's district-level poverty estimates have
several limitations because of the way they're produced . The Census
Bureau estimates are less accurate for certain populations, such as rural
communities or transitory families, which affects poverty measurements.
In addition, the Census Bureau counts children in the districts where they
live, not in the districts where they're enrolled. Lastly, the census poverty
estimates have a signifi cant lag time and may become less accurate the
further they get from the 10-year census count.
Question 2 Conclusion
Appendix A: Scope Statement
Appendix B: LPA Methodology Used to Determine
Free-Lunch Eligibility
Appendix C: Summary of the Major Free-Lunch
Application Reviews Conducted by
Kansas Department of Education Staff
Appendix D: Comparison of the School District
Enrollment Counts to the U.S. Census Bureau
Populations Estimates, by School District
Appendix E: Agency Response
Appendix F: Changes Made to the Audit Report on December 18, 2006
................ page 20
................ page 22
................ page 24
................ page 26
................ page 30
................ page 31
................ page 42
................ page 45
This audit was conducted by Katrin Osterhaus, Allen Bartels, Dan Bryan, and Heidi Zimmerman.
Scott Frank was the audit manager. If you need any additional information about the audit's
fi ndings, please contact Katrin at the Division's offi ces. Our address is: Legislative Division of Post
Audit, 800 SW Jackson Street, Suite 1200, Topeka, Kansas 66612. You also may call us at (785)
296-3792, or contact us via the Internet at LPA@lpa.state.ks.us.
Notice to the Reader
On December 18, 2006, Legislative Post Audit made changes to pages 7 and 10 of this report
to correct an error that was identifi ed after the report was released in November 2006.
A "strike-and-add" version of those changes is presented in Appendix F so the reader can see
how the report was changed. All legislative committees and all agency offi cials that received
copies of the initial report were sent a copy of these changes.
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
1
Since the State's school fi nance formula was passed in 1992, State
funding for at-risk services has been distributed to school districts
based on the number of students in each district that qualify for
the federal free-lunch program. Funding for at-risk services has
increased signifi cantly in recent years. The 2005 Legislature
nearly doubled the weighting factor for free-lunch students,
bringing total at-risk funding Statewide to almost $111 million for
the 2005-06 school year.
In addition, the multi-year funding plan passed by the 2006
Legislature will more than double at-risk funding through the
2008-09 school year. As funding for at-risk services increases,
the number of students who qualify for the free-lunch program
becomes an increasingly important factor in the State's school
fi nance formula.
Recently, legislators have seen information indicating the number
of students districts report as eligible for the free-lunch program
varies signifi cantly from poverty estimates prepared by the U.S.
Census Bureau. In addition, information compiled by the State
Department of Education, as well as fi ndings in the Legislative
Post Audit cost study, indicate the number of students eligible for
free lunch has little relationship to the number of students who
actually receive at-risk services.
This information has raised concerns among some legislators about
the validity of using the number of students qualifying for free
lunch to measure poverty within each district, and for some it also
calls into question the appropriateness of using free-lunch counts
as the basis for distributing State at-risk funding.
This audit was conducted in two parts. This fi rst part answers the
following questions:
Does the number of free-lunch students used for at-risk
funding accurately refl ect the number of students who are
eligible for the program?
How does the number of free-lunch students reported by
school districts compare with poverty estimates compiled
by the U.S. Census Bureau?
1.
2.
K-12 Education: Reviewing Free-Lunch Student Counts
Used as the Basis for At-Risk Funding, Part I
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
2
To answer the fi rst question, we reviewed federal and State
eligibility requirements for the free-lunch program. We
interviewed Department of Education staff from the Fiscal Audit
team and the Child Nutrition and Wellness team to understand
their roles in verifying the eligibility of free-lunch students.
We verifi ed the eligibility status of a random and representative
sample of 500 free-lunch students. For families who participate in
food stamp or temporary assistance programs with the Department
of Social and Rehabilitation Services (SRS), we verifi ed their
participation in these programs. For families who applied for free-
lunch based on income, we compared the households' reported
income information to tax records and quarterly wage reports at
the Departments of Revenue and Labor. We then projected our
results to all free-lunch students Statewide. We also surveyed
offi cials in each school district to estimate the potential number of
under-reported students in the free-lunch count.
To answer the second question, we reviewed information and
interviewed U.S. Census Bureau staff to understand how the
poverty data are compiled and what they measure. We compared
the student free-lunch counts and Census poverty data for the
most recent four years available. We interviewed Department of
Education offi cials and Census staff to identify factors that account
for the large differences between those sets of numbers. For
quantifi able factors, we calculated their impact in order to account
for as much of the discrepancy between the free-lunch count and
Census Bureau estimates as possible. Lastly, we analyzed free-
lunch counts from Kansas and other states in relation to Census
estimates over time, to determine if there were any cyclical
patterns to the disparities.
A copy of the complete scope statement for the audit request
approved by the Legislative Post Audit Committee is included in
Appendix A .
The second part of this audit will answer the last question in the
scope statement:
How does the number of free-lunch students in Kansas
compare with the number of students who receive at-risk
services?
In conducting this audit, we followed all applicable government
auditing standards set forth by the U.S. Government
Accountability Offi ce, except that, because of time constraints,
3.
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
3
we didn't test the computer data the Department supplied to us
regarding district verifi cation results. These data are used to
provide information about the results districts found during their
verifi cation work, and are used only to supplement our audit work.
In addition, the reader should be aware that the income tax data
used throughout the audit are self-reported, and there was no way
for us to assess the accuracy of the information. Although there's
no way to know how accurate these data are, we've compared the
information to quarterly wage data from the Department of Labor
when we could. We found no indication that the data were grossly
or systematically wrong. Thus, the information presented in this
report about our projections of ineligible free-lunch counts should
be viewed as an indicator, and not as absolute fact. It is unlikely,
however, that it is so grossly or systematically inaccurate as to
affect our fi ndings and conclusions. Our fi ndings begin on page 7,
following a brief overview on the National School Lunch Program
and the distribution of at-risk funding in Kansas.
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
4
The National School Lunch Program is a federally assisted
meal program operating in almost 100,000 public and nonprofi t
private schools and residential-care institutions. It provides
nutritionally balanced, low-cost or free lunches to more than 27
million children each day. The U.S. Department of Agriculture
administers the Program at the federal level.
Children from families with incomes
below 130% of the poverty level, or
who participate in certain designated
programs, qualify for free lunches . In
addition, students whose families have
incomes between 130% and 185% of the
poverty level are eligible for reduced-price
lunches. Figure OV-1 shows the total
Kansas headcount enrollment for the 2005-
06 school year, broken down by lunch
status.
Students can qualify for free lunches in two
ways:
Some students are eligible for free lunches when they
participate in one of several designated programs . The
designated programs include Food Stamps, Temporary Assistance
for Needy Families (TANF), and Head Start. Students also are
categorically eligible when they've been identifi ed as migrant,
homeless, or runaway.
Many students are directly certifi ed for free lunches by their school
district. To do that, parents supply districts with a letter from the
Department of Social and Rehabilitation Services (SRS) that shows
the family was approved for Food Stamps or TANF benefi ts. SRS
also provides school districts with a list of Food Stamp and TANF
recipients that they can use to match and certify additional students
who don't provide the letter to the district.
Other students are eligible for free lunches when their families
report income below 130% of the poverty threshold . These
families are supposed to report all household members and their
incomes on an application for free lunches. District offi cials review
the applications and compare the income to a set of eligibility
guidelines published by the U.S. Department of Agriculture.
Families whose income is less than the threshold are eligible and
their children can receive free lunches. Children who have been
placed in foster care establish their eligibility in a similar fashion,
except only the child's income is taken into consideration.
Overview of the National School Lunch Program and Distribution of
At-Risk Funding in Kansas
The National School
Lunch Program Provides
Free Lunches to
Students Who Meet
Poverty Thresholds
Or Participate in
Designated Programs
!""#$%&
$' %"
!$%%
"&$"(%!#$!%%"!$
!%'
)!*+
,-$#$!%.$!'
!$")$$.!""#$*
!""#"$%
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
5
Federal law requires school districts to verify the information
on a small sample of approved applications each year . For
most applications, verifi cation involves contacting the person who
fi lled out the application and requesting some form of support for
the income listed. This documentation might include pay stubs, a
tax return, or a letter from the applicant's employer. If the families
fail to provide any documentation, or if the documentation they
submit reveals they aren't eligible for free lunches, the students'
free lunch is removed.
State at-risk funding is based on a headcount of free-lunch students
as of September 20th of each year. For the 2005-06 and 2006-
07 years, school districts who receive a large infl ux of military
families during the year will be allowed a second count on
February 20th to account for the additional students.
Auditors remove ineligible students and add eligible students
to the free-lunch count . The Department's fi scal auditors perform
a comprehensive enrollment audit at each district to verify all
enrollment counts (including free lunch), and establish State
funding for various components of the school funding formula.
To check the free-lunch counts, auditors review students' free-
lunch documentation for completeness and accuracy. In addition,
auditors verify that free-lunch students were enrolled and attending
as of September 20th.
Figure OV-2 details the adjustments made by the Department's
auditors in 2005-06. As the fi gure shows, the auditors removed
some students and added others, for a net reduction of almost 1,000
students Statewide.
Kansas Distributes
At-Risk Funding
Based on the Number
of Students Eligible for
Free Lunches in
Each District
&
'(
)*+,!"
-#.--
#!/
0(,-$#$1$!
10(,-$#$1$!
&+,!"
-/.--#
!$/
"&$"(%!#$%"%"!$
%!$
2
%&',-$#$
$$
"$!%$
0($#"$(!$!$
$$ 33
$$ -!/$!"!#$$!$$"$! /$! $!
&$$#$!
'
)!*1"(
!%$%!#+
,-$#$!%.$!'
!
(012
3*(0
!""#"$%
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
6
Most free-lunch students removed by the Department's auditors
either weren't eligible for free lunches or weren't enrolled and
attending on September 20th . About half the nearly 1,500 students
auditors removed from the free-lunch count had an incomplete,
missing, or incorrectly processed application and weren't eligible
for free lunches. The remaining students the auditors removed were
eligible for free lunches, but didn't meet the enrollment and attendance
requirements, had been claimed by another district, or weren't
approved for free lunches until after the September 20th deadline.
According to the audit team director, auditors add free-lunch students
to the count when they come across applications for reduced-price or
denied applications that should have been approved for free lunches.
This happens when districts don't separate the free-lunch applications
from the others, and auditors end up reviewing all applications.
School districts received $822 in at-risk funding for each free-lunch
student in 2005-06 . The fi nal, audited free-lunch count is referred to
as the "at-risk" count for funding purposes. Figure OV-3 shows how
the Statewide at-risk funding was determined for the 2005-06 school
year.
The multi-year funding plan passed by the 2006 Legislature increases
at-risk funding each year through the 2008-09 school year. As shown
in Figure OV-4 , at-risk funding per free-lunch student will more than
double over the next several years--from $822 in 2005-06 to $2,021
by 2008-09.
!""
#$$
%&'"
!"
!!
"!
!
#$%
&! !!
"!!
'(()(*'((+(,
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
7
In the 2005-06 school year, Kansas public school districts received
almost $111 million in at-risk funding for nearly 135,000 free-lunch
students. About 17% of our Statewide random sample of 500 free-lunch
students were ineligible--primarily because families under-reported
their income. That means the State paid almost $19 million in at-risk
funds for about 23,000 ineligible students last year. On the other hand,
based on our survey of school district offi cials, about 6,900 students
Statewide may have been eligible for free lunches, but their families
didn't apply, mostly because they're embarrassed to reveal their
fi nances.
The free-lunch counts used for at-risk funding also may include a
number of students the Legislature didn't intend to fund, including adult
students in alternative schools and part-time students. We also identifi ed
additional problems with the Department of Education's free-lunch
reviews that, if addressed, could produce a more accurate count. These
and related fi ndings are discussed in the sections that follow.
As described in the Overview, Kansas distributes at-risk funding based
on the headcount of students in each school district who are determined
to be eligible for free lunches as of September 20 each year. In 2005-
06, the State provided almost $111 million in at-risk funding for about
135,000 free-lunch students Statewide. Figure 1-1 shows the total
number of free-lunch students and total amount of at-risk funding
provided Statewide over the past fi ve years. As the fi gure shows, at-
risk funding per student more than doubled in 2005-06, dramatically
increasing the total at-risk funding for that year.
!
"!
!"#$%& '&(
)*"
" +,"-"&
$%& '&(.
ANSWER IN BRIEF
Question 1: Does the Count of Free-Lunch Students Used for At-Risk Funding
Accurately Refl ect the Number of Students Who Are
Eligible for the Program?
In 2005-06, Kansas
Districts Received Almost
$111 Million in At-Risk
Funds for About 135,000
Students Identifi ed as
Eligible for Free Lunches
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
8
We selected a random sample of 500 students from the Department's
audited free-lunch counts for the 2005-06 school year. School districts
sent us copies of the free-lunch applications or other related materials
for each student showing why they were determined to be eligible.
As discussed in the Overview, students can become eligible for free
lunch in two ways:
Students are "categorically" eligible when they participate in a
qualifi ed program, such as Food Stamps or Temporary Assistance for
Needy Families, or when they have been determined to be runaway,
homeless or migrant.
Students are "income" eligible when families' reported incomes are
below the 130% poverty threshold.
To verify categorically eligible students, we confi rmed that they
participated in a qualifi ed program or had a certain status. For
example, if a student received free lunches because his family also
received food stamps, we reviewed documentation to confi rm that
status.
To verify income-eligible students, we reviewed tax return data from
the Department of Revenue and quarterly wage reports from the
Department of Labor to verify their reported income. If the verifi ed
income was more than the threshold for receiving free lunches we
counted the student as ineligible. Children in foster care can become
eligible based on income because only the student's income is
considered. For those students, we confi rmed with SRS that they were
in foster care.
This decision-making process involved a
complex series of steps. Those steps are
described in more detail in Appendix B .
Of the 500 free-lunch students in our
random sample, 85 students (17%)
weren't eligible, primarily because
households under-reported their
income . Figure 1-2 summarizes our
fi ndings, broken down by how the
student originally was determined to be
eligible.
As the fi gure shows 80 of the 85 students
we found to be ineligible were originally
approved for free lunches because of
About 17% of Free-
Lunch Students in Our
Statewide Random
Sample Were Ineligible,
Costing the State an
Additional $19 Million in
At-Risk Funds
#
$%&'(
)
*(
+
*(
("&')*+
$
,#
-
!'.
$
-
.'&/*+0&&1
')"2/)"&3
,
#
#-
403
##-
%
,
-./
-
012+1
1
)12
!
!"
035"2&'+"6)"7+8"
'
&''&'""78"&'"'"1
'&9:"/'11')&''&;
"6)"'1*&'"
8":+'"
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
9
their families' low income. In virtually every case, we found that
the family had under-reported the household income. For a majority
of these households the under-reporting was substantial. Of the 80
ineligible students:
62 households earned income that was 20% or higher than the free-lunch
income limit.
12 of the 62 households earned income that was more than double the
free-lunch income limit.
These results are consistent with a 2004 study by the U.S. Department
of Agriculture which found that 18% of the students in its sample were
ineligible.
As Figure 1-2 also shows, four of the ineligible students were listed
as foster care children on the free-lunch application. SRS offi cials
had no record of three of these students being in foster care. For
one additional foster child, the district received and approved the
application after the offi cial September 20 count date.
The cases described in the accompanying box
illustrate some of the scenarios we encountered.
We conducted a series of statistical analyses to
see if students from different types of districts
were more likely to be ineligible. We compared
districts of different sizes, location, and poverty,
but didn't fi nd any signifi cant differences.
Many households are able to under-report
their income and still receive free lunches
because federal law requires school district
offi cials to accept their applications at face
value . The National School Lunch Program
specifi cally states that district offi cials must
determine if a student is eligible for free
lunches based solely on the income information
presented on the application. Even if offi cials
know a family is lying about their income,
offi cials still are required to approve the
application. (They can require the family to
submit pay stubs or other documentation to
verify the information after the application
was approved.) By contrast, we were able to
use data sources districts don't have access to
Examples of Ineligible Students From Our
Statewide Random Sample
In all, 85 of the 500 free-lunch students we selected in a Statewide
random sample actually weren't eligible for free lunches. The
following examples illustrate some of the scenarios we encountered
for ineligible students.
Applicants who didn't report the earnings for all the
members of their household --One application we reviewed
showed several household members, with only one person
listed as a wage earner. The total household income on the
application was a little more than $20,000. When we looked
up the household members at the Department of Revenue, we
found tax returns for fi ve household members totaling more
than $168,000 in annual income--well above the $42,000
eligibility threshold for free lunches.
Self-employed applicants who didn't report their full
income --Because the monthly income for many self-
employed people is very irregular, they are allowed to submit
their tax records from the previous year to establish their
annual income. We reviewed one application where the
applicant claimed part of his income from his business. When
we looked at the applicant's full tax return, we found more than
$100,000 in total income.
Applicants who claimed the student was in foster care,
when they weren't --Special rules exist for foster care
children that make it very easy for them to qualify for free
lunches. All the applicant has to do is check a box to indicate
the child is in foster care. We reviewed 16 such applications,
and checked with SRS offi cals to see if these students actually
were in foster care. SRS offi cials had no record of the child
being in foster care in three of these 16 cases.
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
10
such as tax returns and quarterly wage reports to determine students'
eligibility for our random sample.
In 2005-06, the State paid almost $19 million in additional at-
risk funds to school districts for nearly 23,000 ineligible students .
When the results from our random sample are projected to the nearly
135,000 free-lunch students Statewide, we estimate that almost 23,000
free-lunch students who were counted for at-risk funding (17%) really
weren't eligible. The State paid $822 in at-risk funding per free-
lunch student in 2005-06, meaning it provided almost $19 million in
additional funds for these students.
Under the school fi nance plan the Legislature passed during the 2006
session, the amount of State at-risk funding per free-lunch student will
increase signifi cantly over the next three years. Figure 1-3 shows the
estimated additional cost to the State for 23,000 ineligible students for
that time frame. As the fi gure shows, if the error rate we identifi ed
stayed the same over the next three years, the additional cost to the
State would increase from almost $19 million in 2005-06 to just more
than $46 million in 2008-09.
A common thought among many school district offi cials is that, while
some families under-state their income in order to get free-lunches,
a large number of families who would be eligible don't apply. To
estimate how many of these students there might be, we surveyed
district offi cials in all 300 districts. We received completed surveys
from 123 districts--a 41% response rate.
Offi cials from 123 districts estimated that about 3,800 potentially
eligible students didn't apply for free lunches. Using those results,
we estimated there may be as many as 6,900 potentially eligible
students Statewide. Had those families applied and been approved for
free lunches, it would have cost the State an additional $5.7 million
in at-risk funding in 2005-06. The reader should keep in mind that
!""#"$
!""$"4
!""4"5
!""5"+
1$ 24&$&$!
!%$0
'
'
'
'
1$ 2&
)$$
5
5
5
5
*6(!!.54#
6
7585
7!48#
7-485
7/$8!
-
91:(0);<
3!!.54#6
!""#"$!""5"+%
)!*1"(
!%$%!#
!!" $$ + ,-$#$ !%.$!
/0!))'
%
Based on Our Survey of
District Offi cials,
About 6,900 Students
Statewide May Have Been
Eligible for Free
Lunches, But Their
Families Didn't Apply
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
11
these projections are based
on assessments provided by
school district offi cials, and
there's no way for us to verify
the accuracy of their estimates.
According to district
offi cials, most eligible
families who don't apply
are either too embarrassed
to do so, or are concerned
about the confi dentiality of
their applications . We asked
district offi cials to identify the
main reasons why families
don't apply for free lunches
when they would be eligible
to receive them. Those results
are summarized in Figure 1-4 .
Because of actions the Legislature took during the 2005 and 2006
legislative sessions, at-risk funding per student will increase from
less than $400 in 2003-04 to more than $2,000 by 2008-09. Because
of the dramatic increase in funding for each student, policies that
allow districts to count additional students will cost the State
signifi cantly more in the future. In this section, we discuss two
policies that allow districts to include students in their free-lunch
counts that the Legislature may not have intended to fully fund.
The Department of Education has developed an alternative at-
risk funding application so school districts can receive at-risk
funds for students who attend schools that don't provide lunches .
Many school districts have set up alternative school settings to
better reach and provide services to non-traditional students, such as
pre-kindergarten children, high school drop-outs, delinquents, and
pregnant teenagers.
Many of these schools offer classes online, at night, or on weekends,
and therefore don't provide lunches to the students. Because no
lunch is available, students attending these schools can't apply for
free lunches under the National School Lunch Program, even if their
income is low enough to qualify.
To let school districts count these students for at-risk funding
purposes, the Department has developed an alternative application
that closely mirrors the federal application for free lunches.
)*=
0(
9(
((>2
)
,!6$3$$!/"%
0 !%
#0
#$
,!6$3$$!/"%
0 !%
!%$"$(!
$--3!2$!!%%"$$!!
#-"$
,!7$3$$!--"(0
$($!!-!
8/"&!
!!$!"!&"!/$!
& $&!/#$
/
,!6$3$$!$2$$#$!!
$--3!2
8/"&&0
9!$3!%$-!&#
-$&
3""!$-$!0
-!$
3""!3$!
"$ #(! $(%"$--"'
)!*1
/(!% !!" $$- !"'
/
)*((2(
>*=0(
The Free-Lunch Counts
Used for At-Risk Funding
Also May Include a
Number of Students the
Legislature Didn't Intend
To Fully Fund
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
12
According to a 1997 Attorney General opinion, the alternative
application is legal because districts can receive at-risk funding as
long as the student is eligible for a free lunch, even if the student
doesn't actually receive the lunch.
Statewide information on the funding districts receive through
these alternative applications wasn't readily available. Therefore,
we obtained 2005-06 enrollment information from a sample of
eight alternative education schools in four school districts that
don't serve lunch. We compared this information with data
from the Department to determine how many of these students
districts were allowed to count for at-risk funding. Our results are
summarized in Figure 1-5 . As the fi gure shows, the four districts
in this sample received at-risk funding for 319 students attending
these schools, even though those students can't actually receive
free lunches.
9
?(0);
:0!"
-+
0);
?@
#
A0);A(0:3
!""#"$%
)$$
3!!#-"$$"$$/$
2%&--"$!3$#$!0"&
0"%!
%" '
)!*1"(
!%$%!#
!!" $$ $+ ,-$#$!%.$!'
B*!#-
B*!#+>
B*!$"*
B*-"5&
! !-!&#
%!!
$$!"
$$
"$8"" &
:$
&!$
&:$
.)!!"
&:$
0!3)!!"
;%% !"! %!
!# !!""
! !-!&#
%!!
$$!"
$$
! !$
!%
(!"$
2 $$
,0(1"$$/
)!!"
! !-!&#
%!"$
"
!:!$(
&:$
! !-!&#
%!!
$$!"
$$
;%% !"! %!
$"
0
.+
,0(&
:$
;%% ! $!#$$
!%-&$&"
-!$$"!-!$
"<$$
!4
!4@
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
13
Allowing districts to use the alternative at-risk funding application
for students who don't receive free lunches creates two potential
problems:
Many of the students in these settings are adults for whom the
State normally wouldn't provide at-risk funding for anyway .
The alternative application helps ensure districts don't lose at-risk
funding when students are in alternative settings. However, because
those settings include many adults who normally wouldn't be funded,
districts actually can gain additional at-risk funding. For example, in
the eight alternative schools we reviewed, 127 of the 319 free-lunch
students were over age 20.
Neither school districts nor the Department are required to verify
the income information on any of these alternative applications .
Because the alternative applications aren't actually for free lunches,
they aren't covered by the required verifi cation process we described
in the Overview. Without the possibility that the application will
be selected for verifi cation, there's a greater risk that the income
information on these applications could be manipulated in order to
generate more at-risk funding.
Districts receive the full amount of at-risk funding for part-
time students because the State doesn't prorate the funding . By
statute, the State is required to distribute at-risk funding based on the
headcount of free-lunch students in each school district. This means
that districts receive a full share of at-risk funding ($822 per student
in 2005-06) for every student, even if they attend school only part
time. Most of these are kindergarten students.
To determine the fi nancial impact of using the free-lunch headcount,
we compared the 2005-06 headcount in the Topeka school district
to the number of free-lunch FTE students. That year, Topeka had
about 7,150 free-lunch students (headcount) who generated $5.9
million in at-risk funds. Had the count of part-time free-lunch
students been prorated, Topeka would have had about 6,735 free-
lunch students (FTE), who would have generated $5.5 million in
funding--a difference of about $340,000. Because the State's
funding for at-risk students will increase signifi cantly over the next
several years, the difference in Topeka would cost about $840,000 in
2008-09.
Two teams within the Department review the applications for free
lunch to ensure that the students' eligibility has been determined
correctly:
The Department's fi scal audit team reviews the applications to ensure
districts receive the right amount of at-risk funding.
We Identifi ed Additional
Problems With the
Department's Free-Lunch
Reviews That, If
Addressed, Could
Produce a More
Accurate Count
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
14
The Department's child nutrition team also reviews the applications to
fulfi ll certain requirements of the National School Lunch Program.
In this section, we discuss how these two teams' work could be
better coordinated to produce a more accurate free-lunch count for
at-risk funding.
In 2005-06, the child nutrition team didn't report roughly 1,850
ineligible free-lunch students it knew about to the Department's
fi scal auditors . The child nutrition team becomes aware of students
who aren't really eligible for free-lunches in two ways:
First, the child nutrition team oversees school districts' efforts
to verify the fi nancial information on a sample of free-lunch
applications, and compiles the results of those reviews . In
2005-06, school districts identifi ed 1,839 ineligible students and
reported their names to the child nutrition team. Of these, 14% were
determined ineligible because the applicant failed to respond to
repeated requests for supporting documentation.
Second, the child nutrition team identifi es students who aren't
eligible for free-lunches through its own reviews . The team
conducts several reviews of free-lunch applications to fulfi ll certain
federal requirements. During one of these reviews for the 2005-06
school year, the team identifi ed 17 additional students who weren't
eligible for free-lunches and had been missed by the Department's
fi scal auditors.
The Department's child nutrition team didn't report the roughly
1,850 ineligible students to the fi scal auditors, who could have
removed them from the at-risk funding counts. For these students,
the State overpaid districts $1.5 million in at-risk funding in 2005-
06.
In 2005-06, the Department's fi scal auditors didn't audit every
school in the six largest districts, potentially missing about 100
ineligible students . The fi scal auditors review 100% of the free-
lunch applications in all but the six largest school districts (Wichita,
Olathe, Shawnee Mission, Kansas City, Topeka, and Blue Valley).
Because of limited resources, the auditors review the applications at
a sample of schools in those districts--one-third of the elementary
schools, half of the middle schools, and all of the high schools, and
other special attendance centers.
In 2005-06, this sampling strategy left 147 schools with about
25,000 free-lunch students unaudited. For these six districts, we
used the auditors' results in the schools they did audit to estimate
the number of students who likely were ineligible in the schools
they didn't audit. These results are shown in Figure 1-6 .
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
15
As the fi gure shows, the Department's auditors may have missed
about 100 ineligible students in these six districts. Because the
sample wasn't random, these results aren't precise, but they do
provide a general indicator of the number of ineligible students
who were missed.
While this may have cost the State about $80,000 in 2005-06,
increases in the at-risk funding per student in the future could
drive that fi gure to about $196,000 in 2008-09. According to the
Department's audit team director, it would take one additional
auditor at a cost of $60,000 (salary and benefi ts) to audit all the
schools in these six districts.
Because the Department's two teams don't coordinate their
reviews, the same free-lunch applications may be reviewed
several times by Department staff . In all, the fi scal audit team
and the child nutrition team conduct three separate reviews that
involve checking the approved free-lunch applications to ensure
students are really eligible. Appendix C contains more detailed
information about each review.
Although the reviews are conducted for different purposes, it's
reasonable to assume that once a State offi cial has reviewed an
application and determined it was processed correctly, it doesn't
need to be reviewed a second or third time by another offi cial.
If the review efforts were streamlined and coordinated between
teams, it potentially could free up staff time to review other
program components, or to review more schools each year.
!"
!"
#$
"
%$&'$$
#
$%&'$$
()
&$$(()*(+$&&,(
&,,,,-&&,(-.,/
&(0123$*4(-*5,&,
*(6$()$
,(/
'*
+
,,-
$..%.*/
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
16
The most important factors that cause school district offi cials
to mistakenly approve a large number of ineligible students for
free lunches are outside the districts' control. Even though many
families mistakenly under-report or even purposefully lie about
their income on the free-lunch application, federal law requires
school districts to accept those applications at face value. That's
because under the National School Lunch Program concerns about
fraud and abuse are secondary to the goal of making sure students
who need free meals get them.
Although school districts do work to verify the information on
some of the applications, they're not authorized to look at the tax
returns and wage reports we could. This means the free-lunch
counts always will be overstated, no matter how diligently school
district and Department of Education offi cials enforce the rules
of the Program. Still, as we've shown, there are some things the
Department could do to make the counts more accurate.
To make the free-lunch count as accurate as possible, the
Department of Education should do the following:
require districts to verify a random sample of alternative
"at-risk" applications for students who attend schools
in alternative settings that don't provide lunches. The
verifi cation should be done in the same manner as
traditional free-lunch applications are verifi ed under the
National School Lunch Program.
ensure its child nutrition team shares the free-lunch
eligibility fi ndings from the school district's reviews and its
own reviews with the fi scal audit team, and ensure that the
fi scal staff adjust the free-lunch counts accordingly.
To increase the effi ciency of free-lunch eligibility reviews, the
Department of Education should do the following:
create a system that allows Department staff to indicate
they have reviewed a student's free-lunch status. This
might be done by adding a signature space at the bottom
of the free-lunch application form for Department staff to
document their review.
discourage fi scal auditors and child nutrition staff from
reviewing applications that have already been reviewed
and approved by someone from the Department.
1.
a.
b.
2.
a.
b.
CONCLUSION
RECOMMENDATIONS
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
17
after implementing the effi ciency-related recommendations
in 2(a)-2(b), identify any remaining resources needed to
perform annual enrollment audits in all schools.
To ensure the Legislature only provides at-risk funding for
those students it intended, the House Select Committee on
School Finance and the Senate Education Committee should
consider amending State law to:
institute an age limit for free-lunch students for purposes of
at-risk funding.
change the at-risk funding count from a headcount to an
FTE count.
c.
3.
a.
b.
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
18
In 2003-04, Kansas had 54,000 more free-lunch students than
adjusted fi gures from the U.S. Census Bureau would suggest.
The primary reason for this difference is that the free-lunch
count includes approximately 22,000 ineligible students for that
year, based on the proportion of ineligible students discussed
in Question 1. The Census Bureau's district-level poverty
estimates also have several limitations, including diffi culties in
accurately measuring important populations, signifi cant lag time
in publishing fi gures, and decreasing accuracy as they get further
from the 10-year Census count. These and related fi ndings are
discussed in the sections that follow.
As required by the U.S. Constitution, the U.S. Census Bureau
counts all people living in the United States every 10 years.
Statistical estimates are used to update the population count in
the intervening years. As part of the 10-year count, the Census
Bureau gathers a variety of demographic data from a sample of
households, including information about their earnings.
Each year, the Census Bureau combines the demographic data
from the 10-year count, the intervening population estimates, and
information from government assistance programs to come up
with estimates of the number of people who live below the federal
poverty level within a variety of geographic areas.
Among the breakdowns available each year is an estimate of the
number of children between the ages of 5 to 17 in each school
district who live below the federal poverty level. The most current
estimates for school districts are for the 2003-04 school year.
Because students are eligible for free lunches if their household
incomes are less than 130% of the federal poverty level, we
adjusted these Census estimates upwards (from 100% to 130%) by
using other data from the Census Bureau.
The Census data suggest that approximately 76,000 children
in Kansas were at or below 130% of the federal poverty level
in 2003-04, which was 54,000 fewer than the free-lunch count .
Figure 2-1 compares the adjusted Census Bureau estimates to
the Statewide free-lunch counts from 2000-01 to 2003-04. More
detailed information for each school district is contained in
Appendix D .
Question 2: How Does the Number of Free-Lunch Students Reported by Districts
Compare with Poverty Estimates Compiled by the U.S. Census Bureau?
ANSWER IN BRIEF
In 2003-04, Kansas Had
54,000 More Free-Lunch
Students Than Adjusted
U.S. Census Estimates
Would Suggest
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
19
-
3)12
!33&+40'
-5 !"
#$
,,
,
,,
, $$
$$
$$
#
####
####
$####
,####
#####
####
####
$####
####
###
###
## #
6(1
!:"&
."03
03<'
#-'11'8/8
'&9:"/'11');"6)"
'1*&'""=."
!"
Figure 2-1 also shows that the difference between the Census
Bureau's estimates and the free-lunch counts has been growing. In
2000-01, there were approximately 36,000 more free-lunch students
than the Census data would suggest. By 2003-04, the difference had
grown to just more than 54,000.
One possible explanation for this trend is that the free-lunch counts
and Census poverty estimates are relatively close when each 10-
year count is completed, but grow further apart each year until the
next 10-year count. To test this theory, we analyzed free-lunch and
Census poverty data for all 50 states from 1987 to 2003, but found no
evidence of a cyclical trend.
To identify and quantify the reasons for the differences between the
adjusted Census Bureau poverty estimates and the free-lunch count
in Kansas, we talked to offi cials from the Kansas Department of
Education and the Census Bureau, reviewed a number of studies, and
performed additional analyses on various school district data. Our
results are summarized in Figure 2-2 on the next page.
As the fi gure shows, we were able to account for over half (29,000)
of the 54,000 student discrepancy. By far the most signifi cant factor
we identifi ed was the inclusion of ineligible students in the free-
lunch counts. Based on our results from Question 1, we estimate that
22,000 of the almost 130,000 free-lunch students in 2003-04 weren't
eligible.
The Free-Lunch Count
Is Signifi cantly Higher
Than the Adjusted
Census Estimate,
Primarily Because the
Count Includes
Many Ineligible Students
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
20
/
/
- &"
#$
0" 1'&2
&3
1 &""4
&#
"&#&(&#
5
!# &#
"(
"6"(3/
&"
7
5
!#"&#
"8
& "31&#/
&"#8
&# #"
&#"(
5
%&
/ 0"&(
&#
1",&6"
# 3#
9
5
#
!# &#
1#&6&&
&& 3/
&"
9&#"8
5
'(
!# &#
"8
16:#83
" &#
&("1'
!/
"
&"'&(
5
)%"
$ 1
&&2
&3
# &#
1"#&(&#
" 8,"&&,"&"&,("
36&&
"&#& ""4,68#6#
;&(" ,&"##8"
&(&&"(,3
1&"3
"1&&
5
'"*
+**,
-.
).$
""""#8
& /
0"
&" "*"
" #
""(**+**/
%&(0"*"122/
34
,##5
67(
+'
8$*"
5
,#*+**
/
The next most signifi cant factor was the age of students (6,500
students). The Census Bureau estimates only include students
from ages 5 to 17; the free-lunch counts include students of ages
three and up.
While the Census Bureau poverty estimates provide a lot of
potentially useful information, that information has limitations.
First, it's important to remember that because of the way the
Census fi gures are created, they are estimates and don't represent
actual counts of the number of children in poverty.
The Census Bureau's
District-Level Poverty
Estimates Have
Several Limitations
Because of the Way
They're Produced
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
21
Bureau offi cials don't guarantee the accuracy of the fi gures. More
importantly, we found several limitations when using these estimates
to measure student poverty:
The Census Bureau estimates are less accurate for certain
populations, which affects poverty measurements . The Census
Bureau gathers extensive demographic data from a sample of
households when it conducts the 10-year count. Certain problems with
how this demographic information is gathered for certain populations
can have an impact on the picture of student poverty:
Rural Populations--Because rural areas have signifi cantly fewer
people, the Census Bureau may not receive enough responses
during the 10-year count to accurately estimate the demographic
make-up of people in those areas.
Transitory Populations--As we noted in Figure 2-2 , the Census
Bureau has diffi culty counting populations that frequently move,
including undocumented workers, migrants, and homeless.
People from such populations are more likely to live in poverty.
Children in Large Families or Unconventional Living
Arrangements--According to the Census Bureau, the 1990
Census missed more than two million children. Although offi cials
believe the 2000 Census was more accurate, the Census Bureau
still has trouble counting children accurately. One problem is
the Census form--it only had enough space for six household
members, meaning children from larger families may have been
left out. Another problem was that children in unconventional
living arrangements--for example those living with another family
member--may not have been included.
Foster Care Children—According to the Census Bureau, unrelated
foster care children under the age of 15 aren't counted as part
of the foster family's household because the survey asks for
information on related children.
The Census Bureau counts children in the districts where they
live, not in the districts where they're enrolled . Open enrollment
policies and virtual schools create opportunities for students to attend
different school districts than the ones in which they live. Because the
Census Bureau isn't able to take this into account when it develops its
estimates, the estimate for some districts can become very distorted.
The box on the next page illustrates this problem by comparing the
Census estimate of the number of students in certain districts with the
actual enrollment fi gures for those districts.
The Census poverty estimates have a signifi cant lagtime . As
mentioned earlier, the latest available district-level poverty estimate
was for 2003.
The Census Bureau estimates may become less accurate
the further they get from the 10-year count . The demographic
PERFORMANCE AUDIT REPORT
Legislative Division of Post Audit
November 2006
22
information gathered during the 10-year count serves as the
foundation of the district-level poverty estimates. The estimates
are updated each year, using a set of statistical techniques.
Census Bureau offi cials acknowledge that those estimates may
become less accurate as they get further away from the 10-year
count.
Many states, including Kansas, use student poverty as a proxy
for the number of at-risk students within a school district.
The number of students who are eligible for free lunches
offers a timely and convenient measure of student poverty
that is linked to federal poverty guidelines. Unfortunately,
as our work in Question 1 shows, a signifi cant number of
the students included in those counts aren't eligible for free
lunches.
Poverty estimates published by the U.S. Census Bureau offer
a credible alternative to free-lunch counts for measuring
CONCLUSION
The Census Bureau May Signifi cantly Underestimate
the Number of Children Enrolled in Many School Districts
To test the reliability of the U.S. Census Bureau's estimates of the total number of all school-aged
children (not just low income) in each school district, we compared those estimates to the total public-
school enrollment in each district for 2003-04. Because the Census estimates include all children
between the ages of 5 and 17 living in the district (including those who attend private schools), we
would expect the Census counts to be higher than the publ