Child Care and Early Childhood Education: More Information
Sharing and Program Review by HHS Could Enhance Access for
Families with Limited English Proficiency (17-AUG-06,
GAO-06-807).
Questions have been raised about whether parents with limited
English proficiency are having difficulty accessing child care
and early education programs for their children. Research
suggests that quality early care experiences can greatly improve
the school readiness of young children. GAO was asked to provide
information on (1) the participation of these children in
programs funded through the Child Care and Development Fund
(CCDF) and Head Start, (2) the challenges these families face in
accessing programs, (3) assistance that selected state and local
entities provide to them, and (4) actions taken by the Department
of Health and Human Services (HHS) to ensure program access. To
obtain this information, GAO analyzed program and national survey
data, interviewed officials in 5 states and 11 counties, held 12
focus groups with mothers with limited English proficiency, and
interviewed experts and HHS officials.
-------------------------Indexing Terms-------------------------
REPORTNUM: GAO-06-807
ACCNO: A61014
TITLE: Child Care and Early Childhood Education: More
Information Sharing and Program Review by HHS Could Enhance
Access for Families with Limited English Proficiency
DATE: 08/17/2006
SUBJECT: Aid for education
Child care programs
Children
Data collection
Disadvantaged persons
Education
Eligibility determinations
Families
Foreign languages
Hispanics
Immigrants
Parents
Preschool education
Subsidies
Non English speaking
Head Start Program
HHS Child Care and Development Fund
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO Product. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
******************************************************************
GAO-06-807
United States Government Accountability Office
Report to Congressional Requesters
GAO
August 2006
CHILD CARE AND EARLY CHILDHOOD EDUCATION
More Information Sharing and Program Review by HHS Could Enhance Access for
Families with Limited English Proficiency
GAO-06-807
CHILD CARE AND EARLY CHILDHOOD EDUCATION
More Information Sharing and Program Review by HHS Could Enhance Access
for Families with Limited English Proficiency
What GAO Found
HHS's Child Care Bureau (CCB) did not have information on the total
enrollment in CCDF programs of children whose parents had limited English
proficiency, but data collected by its Office of Head Start in 2003 showed
that about 13 percent of parents whose children were in Head Start
reported having limited English proficiency. The most recent (1998)
national survey data showed that children of parents with limited English
proficiency were less likely than other children to receive financial
assistance for child care from a social service or welfare agency or to be
in Head Start, after controlling for selected characteristics.
Eighty-eight percent of these children were Hispanic, and their results
differed from Asian children.
Likelihood of Selected Outcomes for Children of Parents with Limited
English Proficiency, after Controlling for Other Factors
Compared to similar children of parents proficient in English
All children of parents with limited Hispanic children of Asian children
of English parents with limited parents with limited
proficiency English proficiency English proficiency
Receipt of financial No significant assistance for child care Less likely
Less likely difference
Head Start Less likely Less likely More likely
Source: GAO analysis of Early Childhood Longitudinal Study, Kindergarten
Class of 1998-99.
Analysis of data from focus groups and site visit interviews held by GAO
revealed that mothers with limited English proficiency faced multiple
challenges, including lack of awareness of available assistance, language
barriers during the application process, and difficulty communicating with
English-speaking providers. Some of the challenges that low-income parents
with limited English proficiency experienced, such as lack of
transportation and shortage of subsidized child care slots, were common to
other lowincome families.
The majority of state and local agencies that we visited offered some oral
and written language assistance, such as bilingual staff or translated
applications. Agencies in the majority of locations visited also made
efforts to increase the supply of providers who could communicate with
parents. Officials reported challenges in serving parents with limited
English proficiency, such as difficulty hiring qualified bilingual staff.
Some officials indicated that additional information on cost-effective
strategies to serve this population would facilitate their efforts.
HHS issued guidance, translated materials, and provided technical
assistance to grantees to help them serve children of parents with limited
English proficiency. The Office of Head Start reviewed programs'
assessments of their communities' needs and conducted formal monitoring
reviews, but could not ensure that review teams consistently assessed
grantees' performance on the standards related to language access. CCB
reviewed states' plans on the use of CCDF funds generally and investigated
specific complaints, but had no mechanism for reviewing how and whether
states provide access to CCDF subsidies for eligible children of parents
with limited English proficiency.
United States Government Accountability Office
Contents
Letter Results in Brief BackgroundChildren of Parents with 1 3
Limited English Proficiency Were Less Likely than Other 7 15
Children to Participate in Subsidized Programs Parents 21
with Limited English Proficiency Faced Multiple 25
Challenges That May Have Limited Their Children's 32
Participation in Federally Funded Child Care and Early 38
Education Programs Selected Agencies Took Some Steps to 39
Assist Parents with Limited English Proficiency but 40
Reported Challenges in Serving Them HHS Provided
Assistance to Grantees on Serving Children of Parents
with Limited English Proficiency, but Gaps Remain in Its
Program Review Efforts Conclusions Recommendations for
Executive Action Agency Comments and Our Evaluation
Appendix I Scope and Methodology 42
Appendix II Analyses of the Effects of Limited English Proficiency 51
on Child Care and Early Education Patterns
Appendix III Comments from the U.S. Department of Health and Human 66
Services
Appendix IV GAO Contact and Staff Acknowledgments 70
Tables Table 1: Characteristics of Data Sources Examined Table 43
2: Selected Characteristics of Site Visit Counties Table 47
3: Composition of Focus Groups 49
Table 4: Differences in the Percentages and Odds of Receiving Any
Nonparental Care, by Parents' English Proficiency Status,
Race or Ethnicity, and Both, among Preschool-Aged
Children 54
Table 5: Differences in the Percentages and Odds of Receiving
Financial Assistance for Child Care, among Those in Any
Prekindergarten Care, by Parents' English Proficiency
Status, Race or Ethnicity, and Both 57
Table 6: Percentages of English Proficient Parents and Parents
with Limited English Proficiency, by Race or Ethnicity 58
Table 7: Percentages of English Proficient Parents and Parents
with Limited English Proficiency, by Family Income 58
Table 8: Percentages of English Proficient Parents and Parents
with Limited English Proficiency, by Education 59
Table 9: Percentages of English Proficient Parents and Parents
with Limited English Proficiency, by the Number of
Persons over the Age of 18 in the Household 59
Table 10: Percentages of English Proficient Parents and Parents
with Limited English Proficiency, by Parents' Work Status 59
Table 11: Percentages of English Proficient Parents and Parents
with Limited English Proficiency, by the Number of
Different Types of Child Care Used, among Those Using
Care 60
Table 12: Odds Ratios from Logistic Regression Models Used to
Estimate the Effects of Different Factors on the Likelihood
of Receiving Any Child Care, after Adjusting for Other
Characteristics 61
Table 13: Odds Ratios from Logistic Regression Models Used to
Estimate the Effects of Different Factors on the Likelihood
of Receiving Financial Assistance for Child Care, among
Those in Any Care, after Adjusting for Other
Characteristics 63
Table 14: Odds Ratios from Logistic Regression Models Used to
Estimate the Effects of Different Factors on the Likelihood
of Receiving Center-Based Care, among Those in Any
Care, after Adjusting for Other Characteristics 64
Table 15: Odds Ratios from Logistic Regression Models Used to
Estimate the Effects of Different Factors on the Likelihood
of Participating in Head Start, after Adjusting for Other
Characteristics 65
Figures
Figure 1: Size and Growth of Population of Individuals with
Limited English Proficiency, 1990-2000 8
Figure 2: Flow of Funds Under CCDF and Head Start 10
Figure 3: Relative Odds of Selected Outcomes for Children of
Parents with Limited English Proficiency Compared to
Children of Parents Proficient in English, for Hispanics
and Asians, after Adjusting for Selected Family
Characteristics 17
Figure 4: English and Chinese Versions of a Local Agency's
Child
Care Quality Brochure 28
Abbreviations
ACF Administration for Children and Families
CBRS Computer-Based Reporting System
CCB Child Care Bureau
CCDF Child Care and Development Fund
CRADLE Culturally Responsive and Aware Dual Language Education
ECLS-K Early Childhood Longitudinal Study,
Kindergarten Class of 1998-99
FACES Family and Child Experiences Survey
HHS Department of Health and Human Services
HSNRS Head Start National Reporting System
LEP limited English proficiency
NACCRRA National Association of Child Care Resource
and Referral Agencies
NCES National Center for Education Statistics
OCR Office for Civil Rights
PRISM Program Review Instrument for Systems Monitoring
TANF Temporary Assistance for Needy Families
This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.
United States Government Accountability Office Washington, DC 20548
August 17, 2006
The Honorable Charles E. Grassley Chairman Committee on Finance United
States Senate
The Honorable Max Baucus Ranking Minority Member Committee on Finance
United States Senate
The Honorable Christopher J. Dodd Ranking Minority Member Subcommittee on
Education and Early Childhood Development Committee on Health, Education,
Labor, and Pensions United States Senate
Children whose parents have limited English proficiency are at greater
risk of experiencing difficulties in school than children from
Englishspeaking households. Research suggests that quality early care
experiences can greatly improve the school readiness and future school
success of young children, particularly children at greatest risk of
failure.
U.S. Census Bureau data from 2000 indicate that more than 1.6 million
children age 5 and younger lived in households where no one aged 14 or
over reported English proficiency. Census data also show that these
children were more likely than other children to be from low-income
families. There is interest in how this population is faring in accessing
child care and early education programs that can ease children's
transition to school.
The U.S. Department of Health and Human Services (HHS) administers the two
largest federally funded programs that support early childhood activities.
HHS's Child Care Bureau (CCB) provides block grants to states through the
Child Care and Development Block Grant, commonly referred to as the Child
Care and Development Fund (CCDF), to subsidize child care for low-income
children while their parents work or participate in education or training
activities. HHS's Office of Head Start funds local grantees through its
Head Start program, a comprehensive program designed to foster healthy
child and family development and to help lowincome children achieve school
readiness. States receiving CCDF block grants and Head Start grantees have
limited funds for these programs and employ priorities and waiting lists
to ration services. In fiscal year 2006, CCDF provided approximately $4.9
billion in federal funds to states and territories. In fiscal year 2004
(the latest year for which service delivery data were available), states
and territories received about $4.7 billion in federal funds and served
approximately 1.74 million children in their CCDF programs. In fiscal year
2005, Head Start grantees received about $6.8 billion in federal funding
and served approximately 900,000 children. The majority of individuals
with limited English proficiency are immigrants-individuals not born in
the United States-although most children of immigrants were born in the
United States. Children must be
U.S. citizens or legal residents to receive CCDF subsidies, while a
child's immigration status is not a factor in determining eligibility for
Head Start. The parent's immigration status is not relevant for
determining eligibility for either program.
Organizations working on issues affecting children and parents with
limited English proficiency have raised concerns that these families may
have difficulties accessing federally funded child care and early
education programs. In this context, you asked us to answer the following
questions:
(1)
What is known about the participation of children whose parents
have limited English proficiency in child care and early education
programs funded through CCDF and Head Start? (2) What challenges
do these families face in accessing these programs? (3) What
assistance is provided by selected state and local entities to
facilitate access for these families?
(4)
What actions has HHS taken to ensure that these families can
access CCDF child care subsidies and Head Start?
To address these issues, we used multiple data collection methodologies.
To determine the participation in federally funded child care and early
education programs by children of parents with limited English
proficiency, we reviewed HHS data from a survey of Head Start participants
and from a reporting system used by Head Start grantees. To assess the
reliability of these data, we interviewed relevant HHS officials and
contractors and reviewed documentation related to the procedures for
collecting and analyzing these data. We found the Head Start survey data
to be sufficiently reliable for the purposes of this report, and while we
did not independently verify the data available through the reporting
system, we found no evidence to suggest that they were unreliable. We also
requested information from all 50 states and the District of Columbia on
their collection of language data for CCDF subsidy recipients. To obtain
information on the child care and early education patterns of these
children that could not be obtained from HHS data, we analyzed national
survey data collected in 1998 as part of the Early Childhood Longitudinal
Study, Kindergarten Class of 1998-99 (ECLS-K), from parents of
kindergarten children about their children's experiences the year before.
Specifically, we used a logistic regression model to estimate the effect
of parents' English proficiency on children's child care and early
education patterns, while controlling for selected individual and family
characteristics such as race and parental education. ECLS-K, conducted by
the National Center for Education Statistics (NCES), was the most recent
national dataset that allowed us to examine child care and early education
experiences of children while considering the English proficiency of their
parents. We assessed the reliability of NCES data and found these data to
be sufficiently reliable for the purposes of this report. To understand
the challenges that parents with limited English proficiency face and what
state and local entities are doing to assist them, we visited five states
(Arkansas, California, Illinois, North Carolina, and Washington) and
contacted 11 counties across these states. We interviewed state and local
officials administering CCDF and Head Start as well as local child care
and early education providers. We selected our site visit locations on the
basis of the size and growth of their population with limited English
proficiency, the presence of any initiatives focused on individuals with
limited English proficiency, and their geographic location. We also
conducted 12 focus groups in California, North Carolina, and Washington
with mothers who spoke Spanish and Vietnamese, reported limited English
proficiency, and had children aged 5 or younger enrolled in child care who
likely qualified for CCDF subsidies based on their family's income and
parental work and education activities. Six focus groups were conducted
with mothers whose children received a government child care subsidy, and
six focus groups were conducted with mothers whose children were eligible
for but did not receive the subsidy. To determine what HHS is doing to
ensure access to its programs, we interviewed HHS officials from the
Office for Civil Rights (OCR), the Office of Head Start, and CCB, and
reviewed relevant documents, legislation, guidance, and other federal
resources related to language access. Appendix I contains more information
about our scope and methodology. Appendix II contains information on the
regression analysis of ECLS-K data that we conducted. We conducted our
work between July 2005 and June 2006 in accordance with generally accepted
government auditing standards.
Results in Brief
The most recent national survey data showed that in 1998, children of
kindergarten age whose parents had limited English proficiency were less
likely than other children to have received financial assistance from a
social service or welfare agency for child care or to participate in Head
Page 3 GAO-06-807 Child Care and Early Childhood Education
Start in the year before kindergarten, after controlling for selected
individual and family characteristics such as race and parental education.
Eighty-eight percent of these children were Hispanic, and the results
differed between them and Asian children. However, these data could not be
used to assess their likelihood of enrollment in CCDF programs because the
survey questions did not ask for the specific agency providing financial
assistance. Further, CCB did not have information on the total enrollment
in CCDF programs of children of parents with limited English proficiency
because it did not require states to collect and report any language data
from parents of children receiving federal subsidies, such as their
primary language or English proficiency. We found that 13 states collected
some language data from parents whose children received CCDF subsidies,
primarily to determine the need for interpreters or translated forms.
However, these data had limitations that reduced their usefulness in
assessing participation in CCDF programs by children of parents with
limited English proficiency. For example, 5 states made the collection of
language data by caseworkers optional, and state officials told us they
could not guarantee that the information was consistently collected. The
Office of Head Start collected some language data on the language spoken
by Head Start participants, which showed that about 13 percent of parents
of the approximately 900,000 children enrolled in Head Start in 2003
reported speaking English "not well" or "not at all."
Focus group participants, state and local child care officials, and
advocates told us that parents with limited English proficiency faced
multiple challenges in accessing federally funded child care and early
education programs for their children. Analysis of data from focus groups
with mothers whose children were eligible for federal child care subsidies
revealed that some of them were not aware of the programs. Parents also
faced challenges during the application process, according to focus group
participants and state and local officials interviewed. For example, some
of them faced obstacles due to a lack of bilingual staff or translated
applications, especially for languages other than Spanish. Additionally,
parents reported difficulties communicating with their children's
Englishspeaking providers. Officials reported shortages of providers with
the language ability to serve families with limited English proficiency.
Parents' immigration status also presented indirect challenges to the
participation of children in federally funded child care and early
education programs. For example, local officials and community advocates
told us that some parents with limited English proficiency may be
reluctant to apply for fear of exposing undocumented immigrant members in
the household. Finally, some parents with limited English proficiency
experienced challenges that were common to low-income families generally.
For example, difficulty finding care at nontraditional hours, lack of
transportation, and the limited number of subsidized child care slots
available affected the ability of these parents to access programs.
The majority of state and local agencies and providers that we interviewed
on our site visits took some steps to assist parents with limited English
proficiency, but officials reported challenges in serving these parents.
In all counties that we visited, agencies offered some form of oral
language assistance, although the scope of this assistance varied and
parents continued to experience challenges when accessing services. For
example, agencies in 5 of the 11 counties visited had staff that could
speak several languages; agencies in the remaining counties had
Spanish-speaking staff, although in one case, the staff were not
specifically assigned to work with program applicants and had other
responsibilities. Most agencies also made available some written language
assistance, such as translated applications, although the scope of the
translations varied as well. For example, local agencies in one state used
applications that the state had translated into eight languages, while
agencies in 2 other states had statetranslated applications only in
Spanish. The majority of agencies and providers also disseminated
information in other languages to raise awareness of their programs and
services. Several state and local agency officials told us that they did
not extensively disseminate information about their programs because their
programs were already operating at full capacity or had substantial
waiting lists. Agencies in the majority of locations that we visited had
initiatives to increase the supply of providers able to communicate
effectively with parents. For example, one local agency we visited, which
provided child care information to parents and worked with child care
providers in the community, offered training and other guidance to Somali-
and Russian-speaking women interested in opening family child care homes.
State and local officials cited several challenges in serving parents with
limited English proficiency, including difficulties hiring qualified
bilingual staff and the expense of translating materials into multiple
languages. Some officials that we interviewed expressed the need for
additional information on cost-effective strategies to serve parents with
limited English proficiency, and several officials said it would be
helpful to learn about provider training in use elsewhere.
HHS provided a variety of assistance to grantees on serving children of
parents with limited English proficiency, but gaps remained in its program
review efforts. HHS's Office for Civil Rights conducted outreach to states
to help them implement guidance on access to HHS programs by individuals
with limited English proficiency and offered technical assistance in
identifying appropriate language access strategies. The Office of Head
Start provided assistance to increase awareness of the Head Start program
and to help grantees better serve children of parents with limited English
proficiency. The Office also reviewed grantees' assessments of child care
and early education resources in their communities relative to the needs
of their communities' Head Start-eligible children. In addition, the
Office conducted formal monitoring reviews of grantees' compliance with
Head Start performance standards, including standards specific to
providing language access to children and parents with limited English
proficiency. An Office of Head Start official, however, told us that the
office could not ensure that its review teams consistently reviewed
grantee compliance with these standards, and in our prior work we found
that no mechanism existed to ensure consistency in the monitoring process.
CCB provided a variety of assistance to help states and child care
providers offer language access to individuals with limited English
proficiency, such as translating brochures. CCB officials told us that
because CCDF is a block grant, CCB's oversight of CCDF is limited to
reviewing states' CCDF plans and investigating complaints. However, CCB
does not require states to report in their CCDF plans how they will
provide language access for individuals with limited English proficiency
or have a mechanism for ensuring that eligible children of parents with
limited English proficiency are not inadvertently excluded from receiving
CCDF assistance because of their parents' citizenship or immigration
status.
To help agencies plan for and provide language assistance to parents with
limited English proficiency who may want to access federally funded child
care and early education programs for their children, we recommend that
HHS work with states to help them explore cost-effective strategies for
collecting data on CCDF subsidy recipients' language preference or English
proficiency. Once these data are available, HHS may consider collecting
information on existing cost-effective ways that agencies could use to
provide language assistance and to recruit providers who speak other
languages, as well as disseminating this information in the locations
where the data show the greatest need. To provide opportunities for
eligible children to receive federal child care subsidies regardless of
their parents' English proficiency, we recommend that HHS develop and
implement specific strategies to review whether and how states provide
access to CCDF programs for these families. These strategies include the
revision of the CCDF plan template to require states to report on how
access will be provided and a systematic review of states' eligibility
criteria to ensure that states comply with HHS policies related to
participation of children whose parents have limited English proficiency.
In its comments on a draft of this report, HHS's Administration for
Children and Families (ACF) generally agreed with our recommendation to
help states explore strategies for collecting data on CCDF subsidy
recipients' language, and provided additional information on its plans and
actions toward implementation of this recommendation. ACF also agreed to
examine the feasibility of using the CCDF plan template to ask states to
report on how they provide access to parents with limited English
proficiency seeking CCDF subsidies for their children. However, ACF did
not address our recommendation that it systematically review states'
program eligibility criteria to ensure that states do not inadvertently
exclude otherwise eligible children of parents with limited English
proficiency from CCDF participation. In addition, ACF submitted detailed
comments on certain aspects of this report, including comments related to
our analysis of ECLS-K data.
Background
Population Changes
The population of individuals with limited English proficiency in the
United States has grown dramatically in recent years. The 2000 Census
shows that the number of people reporting that they do not speak English
well or very well grew by 65 percent, from 6.7 million in 1990 to almost
11 million in 2000. The data also show that while growth in the population
of individuals with limited English proficiency continues in states along
the border, such as California and Texas, it is most rapid in other
states. (See fig. 1.)
Figure 1: Size and Growth of Population of Individuals with Limited English
Proficiency, 1990-2000
Sources: GAO analysis of U.S. Census data. Copyright Corel Corp. All
rights reserved (map).
Note: In our analyses of Census data, we categorized the population of
individuals reporting that they do not speak English well or very well as
those with limited English proficiency.
Page 8 GAO-06-807 Child Care and Early Childhood Education
As figure 1 shows, for example, the number of individuals who did not
speak English well or very well increased by more than 300 percent between
1990 and 2000 in North Carolina and Georgia, and by more than 200 percent
in states such as Nebraska, Arkansas, and South Carolina. In 2000, 14
percent of children age 5 and younger in households below the federal
poverty level lived in linguistically isolated households. 1
HHS Child Care and Early Education Programs
The two largest sources of federal support for child care and early
education are CCDF and Head Start. CCB administers CCDF and the Office of
Head Start 2 administers Head Start. Both entities are housed within ACF.
CCB provides block grants to states through CCDF to subsidize child care
expenses of eligible families. In contrast, the Office of Head Start
awards grants for the operation of Head Start programs directly to local
public or private organizations, school systems, or Indian tribes. The
flow of funds under CCDF and Head Start is shown in figure 2.
1
The U.S. Census defines a "linguistically isolated household" as one in
which no person aged 14 or over reported either speaking only English at
home or speaking English very well.
2
In June 2006, the Head Start Bureau was officially renamed the Office of
Head Start.
Page 9 GAO-06-807 Child Care and Early Childhood Education
Figure 2: Flow of Funds Under CCDF and Head Start
a
Source: GAO analysisand Art Explosion images.
a
These are examples of providers that parents may choose with CCDF
vouchers. Parents may choose any other legally operating provider
authorized by the state.
CCDF is used to subsidize the child care expenses of low-income families
with children under age 13 and to improve the overall quality and supply
of child care. The goals of the program are to (1) allow each state
maximum flexibility in developing child care programs and policies; (2)
promote parental choice to empower working parents to make their own
decisions on the child care that best suits their family's needs; (3)
encourage states to provide consumer education information to help parents
make informed choices about child care; (4) assist states to provide child
care to parents trying to achieve independence from public assistance; and
(5) assist states in implementing the health, safety, licensing, and
registration standards established in state regulations. The parent whose
child receives child care assistance may either enroll the child directly
with a provider who has a grant or contract from the state for the
provision of child care services or receive a certificate to enroll the
child with a provider of the parent's choosing. Parents may choose from
any child care legally offered in the state, which could include care
provided in child care centers, family child care homes, or by relatives
or nonrelatives in the child or provider's home. CCDF is a combination of
discretionary and mandatory funds. In federal fiscal year 2006, CCDF
provided about $4.9 billion in federal funds to states and territories. In
fiscal year 2004 (the latest year for which data were available), the
program served approximately 1.74 million children with federal funding of
about $4.7 billion. In addition, federal CCDF funds are supplemented with
state contributions, and HHS officials reported that total federal and
state expenditures for CCDF amounted to almost $9.4 billion in fiscal year
2004.
Congress gave states considerable flexibility in administering and
implementing their CCDF programs. States are required to submit biennial
plans to CCB describing their CCDF activities. States determine income
eligibility thresholds up to a federal maximum of 85 percent of the state
median income. In their CCDF plans for federal fiscal years 2004 and 2005,
almost all states reported setting lower income eligibility limits, with
only 5 states at the federal maximum of 85 percent. 3
Because CCDF is a nonentitlement program-one with limited funding and not
necessarily intended to cover all eligible persons-states are not required
to provide child care subsidies to all families whose incomes fall below
the state-determined eligibility threshold, and states may establish
National Child Care Information Center, "Trends in State Eligibility
Policies: A CCDF Issue Brief," Vienna, Virginia, July 2004.
Page 11 GAO-06-807 Child Care and Early Childhood Education
priorities for serving eligible families, such as prioritizing families
receiving Temporary Assistance for Needy Families (TANF), in order to
support their work efforts. States can augment their CCDF funds with other
funding sources, such as TANF, to increase funding available for
subsidies. States spent $1.4 billion in federal TANF funds directly on
child care in fiscal year 2004. 4 States may also transfer up to 30
percent of their TANF block grants into their CCDF programs. In fiscal
year 2004, the latest year for which data were available, $1.9 billion in
TANF funds was transferred to CCDF. Funds transferred from TANF to CCDF
must be spent in accordance with CCDF rules. This is significant partly
because the effect of the child's or the parent's citizenship or
immigration status on the child's eligibility differs depending on the
program. For example, parents' immigration status may affect their
eligibility for child care assistance under TANF, 5 whereas only the
immigration status of the child matters for determination of eligibility
for subsidies from CCDF. Although legislation authorizing CCDF did not
specify the effect of citizenship or immigration status on program
eligibility, HHS's guidance to state agencies indicated that states should
consider only the citizenship and immigration status of the child when
determining the child's eligibility for federal child care assistance. 6
Therefore, children who are citizens or legal residents are eligible for
CCDF subsidies regardless of their parents' citizenship or immigration
status.
States are also required to dedicate at least 4 percent of their CCDF
allotments to activities to provide comprehensive consumer education to
4
Partly as a condition of receiving federal funds, states also used their
own funds for this purpose. According to HHS, this brought total federal
and state child care expenditures under TANF to about $3.4 billion in
fiscal year 2004.
5
The Personal Responsibility and Work Opportunity Reconciliation Act,
(PRWORA, Pub.
L. No. 104-193) restricts access by some legal immigrants to certain
programs and denies access by illegal immigrants to many government-funded
programs. States can decide the eligibility for TANF of most of the
qualified aliens who arrived in this country prior to August 22, 1996.
Most of the qualified aliens who entered the United States on or after
August 22, 1996 are barred from receiving TANF the first 5 years after
their entry, although some states choose to provide their own state-funded
public assistance to such immigrants.
6
The guidance states that "for implementing the verification requirements
mandated by title IV of PRWORA, only the citizenship and immigration
status of the child, who is the primary beneficiary of the child care
benefit, is relevant for eligibility purposes." U.S. Department of Health
and Human Services, Administration on Children, Youth, and Families. Log
No. ACYF-PI-CC-98-08, November 25, 1998.
parents and to improve the quality and availability of child care. 7
States may use some of this quality set-aside to fund child care resource
and referral services that are available in every state and most
communities in the United States. These agencies provide information to
parents on finding and paying for quality child care, offer training to
child care providers, and frequently engage in efforts to analyze and
report on child care supply and demand in their communities. Often,
resource and referral agencies also manage the CCDF subsidy program or are
part of local organizations that administer the subsidy in the community.
Head Start offers child development programs to low-income children
through age 5 and their families. The overall goal of Head Start is to
promote the school readiness and healthy development of young children in
low-income families. In addition to providing classroom programs for the
children, Head Start grantees provide or arrange for a variety of
services, including medical, dental, mental health, nutritional, and
social services. Children in families with incomes below the federal
poverty level ($20,000 for a family of four in 2006) 8 are eligible for
available Head Start programs regardless of their or their parents'
immigration status. Head Start grantees must adhere to certain performance
standards, including standards related to providing language access in
Head Start programs. The Office of Head Start reviews the performance of
Head Start grantees on these standards using a structured guide known as
the Program Review Instrument for Systems Monitoring (PRISM). In fiscal
year 2005, Head Start was funded at $6.8 billion and served 906,993
children.
Ensuring Meaningful Program Access for Persons with Limited English
Proficiency
HHS has responsibility for monitoring grantees' compliance with program
requirements. Through its Office for Civil Rights (OCR), HHS also oversees
compliance with Title VI of the Civil Rights Act of 1964, 9 which states
that no person shall "on the ground of race, color, or national origin, be
excluded from participation in, be denied the benefits of, or be subjected
7
In addition to the minimum 4 percent quality set-aside, annual
appropriations have provided funding for child care quality activities.
HHS officials noted, for example, that the agency's 2006 fiscal year
appropriation provided approximately $270 million for quality improvement
activities, including nearly $100 million to improve the quality of infant
and toddler care and approximately $10 million for child care research and
evaluation initiatives.
8
Federal Register, Vol. 71, No. 15, January 24, 2006, pp. 3848-3849. Alaska
and Hawaii have higher guidelines.
9
42 U.S.C. S: 2000d et. seq.
to discrimination under any program or activity receiving federal
financial assistance." HHS has issued regulations to recipients of HHS
funds on implementing the provisions of Title VI, including requiring an
assurance in every application for federal financial assistance that the
program will be operated in compliance with all requirements imposed under
HHS's Title VI regulations.
Moreover, Executive Order 13166, issued in 2000, required federal agencies
to prepare a plan and issue guidance to their funding recipients on
providing meaningful access to individuals who, as a result of national
origin, are limited in their English proficiency. In August 2003, HHS
published revised guidance pursuant to Executive Order 13166. The guidance
states that Title VI and its implementing regulations require that
grantees take reasonable steps to ensure meaningful access for individuals
with limited English proficiency, and the guidance is intended to assist
grantees in fulfilling their responsibilities to ensure meaningful access
to HHS programs and activities by these individuals. Under the guidance,
grant recipients are to determine the extent of their obligation to
provide language assistance services by considering four factors: (1) the
number or proportion of individuals with limited English proficiency
eligible to be served or likely to be encountered by the program or
grantee; (2) the frequency with which these individuals come in contact
with the program;
(3) the nature and importance of the program, activity, or service
provided by the program to people's lives; and (4) the resources available
to recipients of federal funds and costs of language assistance. The
guidance states that grantees have two main ways to provide language
assistance services: oral interpretation, either in person or via
telephone, and written translation. Finally, the guidance lays out
elements of an effective plan of language assistance for persons with
limited English proficiency.
Monitoring compliance with Title VI and providing technical assistance are
functions of HHS's OCR. OCR enforces Title VI as it applies to agencies'
responsibilities to ensure access for individuals with limited English
proficiency. The mechanisms available to OCR for ensuring that agencies
comply with their obligations to provide access include complaint
investigations, compliance reviews, efforts to secure voluntary
compliance, and technical assistance.
Children of Parents with Limited English Proficiency Were Less Likely than
Other Children to Participate in Subsidized Programs
The most recent national survey data showed that in 1998 children of
parents with limited English proficiency, 88 percent of whom were
Hispanic, were less likely than other children to receive financial
assistance from a social service or welfare agency for child care or to
participate in Head Start in the year before kindergarten, after
controlling for selected individual and family characteristics. However,
these data could not be used to assess their likelihood of enrollment in
CCDF programs because the survey questions did not ask for the specific
agency providing financial assistance. Further, CCB did not have
information on the total enrollment in CCDF programs of children of
parents with limited English proficiency because it did not require states
to collect and report any language data from parents of children receiving
federal subsidies, such as their primary language or English proficiency.
The Office of Head Start collected some data on the language spoken by
Head Start participants, which showed that about 13 percent of parents of
the approximately 900,000 children enrolled in Head Start in 2003 reported
speaking English "not well" or "not at all."
Children of Parents with Limited English Proficiency Were Less Likely to
Receive Financial Assistance for Child Care or to Participate in Head
Start
National survey data from ECLS-K showed that in 1998, kindergarten
children of parents with limited English proficiency who were in
nonparental child care in the previous year were less likely 10 than other
children in child care to receive financial assistance from a social
service or welfare agency for that care, after controlling for selected
individual and family characteristics. 11 However, parents' limited
English proficiency had a different effect for Hispanics than for Asians
in the dataset. 12 Specifically, as shown in figure 3, Hispanic children
of parents with limited English proficiency (who represented 88 percent of
all children in the
10
All differences reported were statistically significant at the 95 percent
confidence level unless otherwise noted. This means that if no difference
actually existed in the population, we would only expect to find a
difference as large as the one found in the ECLS-K sample less than 5
percent of the time.
11
The characteristics we controlled for in the analysis of the receipt of
financial assistance for child care were race, household income and
parental education, the number of individuals over 18 in the household,
the presence of a parent who was not working, whether care was provided in
a center-based facility, whether the child was in multiple types of care,
and the child's participation in Head Start. In our analysis, we treated
receipt of center-based care and Head Start participation as two distinct
outcomes.
12
Our analysis was limited to Hispanics and Asians because the numbers of
parents with limited English proficiency in other racial or ethnic
categories in the survey were too small to allow the same analysis.
dataset whose parents had limited English proficiency) were less likely
than children of Hispanic parents proficient in English to receive
financial assistance for their care. Among Asians, who constituted about 8
percent of all children of parents with limited English proficiency, we
did not find a statistically significant difference in the receipt of
financial assistance for child care between children of parents with
limited English proficiency and other children. These results, however,
cannot be used to draw conclusions about enrollment in CCDF programs by
children of parents with limited English proficiency because the survey
questions referred to assistance from a social service or welfare agency
generally and did not ask specifically whether assistance came from CCDF.
13 Also, while ECLS-K data are representative of the experiences of
children in the year prior to entering kindergarten, they cannot be
extrapolated to children of all ages. (See app. II for discussion of the
methodology we used to analyze ECLS-K data and the results of our
analyses.)
We also examined differences in the likelihood of being in any type of
nonparental child care in general and in center-based care in particular
in the year before kindergarten. We found that, among Hispanics, children
of parents with limited English proficiency were less likely to have been
in nonparental child care than other children. We did not find a
significant difference in the use of nonparental child care among Asians,
nor did we find a significant difference in the use of center-based care
for either Hispanics or Asians.
Page 16 GAO-06-807 Child Care and Early Childhood Education
Figure 3: Relative Odds of Selected Outcomes for Children of Parents with
Limited English Proficiency Compared to Children of Parents Proficient in
English, for Hispanics and Asians, after Adjusting for Selected Family
Characteristics
Source: GAO analysis of ECLS-K data.
Note: The numbers in this figure show how the odds of having a certain
outcome compare among children whose parents have limited English
proficiency and other children of the same race. For example, among
Hispanics, children of parents with limited English proficiency were less
than half as likely (0.44 times) to receive financial assistance for their
care than other children.
a
Indicates that children of parents of that racial or ethnic group and with
limited English proficiency had statistically significantly different odds
(at the 95 percent level) of having that outcome compared to children of
parents proficient in English of the same race.
Our analysis of ECLS-K data also indicated that after controlling for
selected individual and family characteristics, 14 children of parents
with
The characteristics we controlled for in the analysis of the use of Head
Start were race, household income and parental education, number of
individuals over 18 in the household, and the presence of a parent who was
not working.
Page 17 GAO-06-807 Child Care and Early Childhood Education
limited English proficiency were less likely to participate in Head Start
in the year before kindergarten. Again, this result did not hold
consistently across racial and ethnic groups. Specifically, as shown in
figure 3, children of Hispanic parents with limited English proficiency
were less likely than children of Hispanic parents proficient in English
to participate in Head Start in the year before kindergarten. In contrast,
children of Asian parents with limited English proficiency were more
likely than children of Asian parents proficient in English to participate
in Head Start.
While 1998 ECLS-K data showed that children of parents with limited
English proficiency were less likely than other children to receive
financial assistance for child care and to participate in Head Start in
the year before kindergarten, it cannot be concluded from these data alone
that the differences are due to language barriers in access to programs.
Other factors, such as the availability of child care and early education
programs in the areas in which members of different language groups reside
or access to support networks that provide information about available
programs may also explain this result. In addition, since the time of the
survey, HHS has taken steps to increase the participation of minorities
and children of parents of parents with limited English proficiency, such
as translating CCDF program brochures and undertaking initiatives to raise
awareness of the Head Start program in the Spanish-speaking community.
Furthermore, HHS officials reported substantial increases in federal and
state child care funding since ECLS-K data had been collected, suggesting
that these increases may have increased program access for parents of
children with limited English proficiency. 15 However, neither CCB nor the
Office of Head Start has more recent information on whether children whose
parents had limited English proficiency are more likely to access
financial assistance for child care and Head Start relative to children
whose parents are proficient in English. 16 ECLS-K was the most recent
national dataset that allowed us to examine the receipt of financial
assistance for child care and the participation in Head Start by children
of parents with limited English proficiency in relation to the participation
of similar children whose parents are proficient in English.
15
As mentioned earlier, ECLS-K collected information on the receipt of
financial assistance for child care generally (rather than the receipt of
CCDF subsidies specifically). Therefore, while the ECLS-K data show that
children of parents with limited English proficiency were less likely to
receive financial assistance for child care, these data cannot be used to
comment on the accessibility of a specific program such as CCDF.
16
According to an Office of Head Start official, there has been an increase
in the number of linguistically and culturally diverse children and
families served by Head Start in recent years. However, this increase
could result from the increase of the population of such children and
families in the United States generally.
CCB Does Not Collect Language Data on Children Receiving CCDF Subsidies,
and the Data in the 13 States that Collect Them Have Limitations
While CCB requires that states submit a variety of demographic information
in monthly or quarterly reports, such as information on the race and
ethnicity of CCDF subsidy recipients, it collects no information on the
language spoken by or the English proficiency of parents whose children
receive CCDF subsidies. CCB officials told us that they had no plans to
collect language data for those receiving CCDF subsidies because they
generally collect only information specifically listed in the legislation
authorizing CCDF. A CCB official with responsibility for the demographic
data collected from states and officials from 1 state we visited told us
that requiring states to provide language data would create difficulties
for states, such as developing ways to identify individuals with limited
English proficiency. Despite the potential difficulties, various state and
local officials in states that do not collect this information, including
the official who cited potential difficulties collecting the data, told us
that having such data would help them evaluate program performance.
While data on the receipt of CCDF subsidies were not available nationally,
13 states collected some language data from parents whose children receive
CCDF subsidies. 17 The specific type of data collected and the manner in
which these data were collected varied among these 13 states, preventing
comparisons among them on the extent to which state CCDF programs were
serving children of parents with limited English proficiency. Officials in
10 of the 13 states that collected language data told us that their states
used the data either to provide translated forms or interpreters to
clients during the application process or for planning or program
evaluation purposes, such as identifying areas with significant increases
in the number of individuals with limited English proficiency and to
determine the need for bilingual staff. State data, however, had
limitations that decreased their usefulness in assessing participation in
CCDF programs by children of parents with limited English proficiency. For
example, 5 states made the collection of language data by caseworkers
optional, and officials in another 5 states told us that despite requiring
caseworkers to collect the language data, compliance with the data
requirements could not always be guaranteed. Officials in 8 of the 13
states that collected language data told us that they could benefit from
having more information on the collection or use of language data or from
learning how other states collect or use them.17
Two of the 5 states that we visited reported collecting their own language
data from clients. In addition, we surveyed the remaining 45 states and
the District of Columbia, and 11 of the 41 states responding to our e-mail
requests for information reported collecting these data.
Page 19 GAO-06-807 Child Care and Early Childhood Education
Head Start Data Indicate That about One-Eighth of Participating Children
Have Parents with Limited English Proficiency
The Office of Head Start collected some language data from the
approximately 900,000 children enrolled in Head Start and their parents
from two sources. First, the Office of Head Start interviewed parents
through its Family and Child Experiences Survey (FACES), a series of
longitudinal surveys of nationally representative samples of children in
Head Start. Based on the 2003 parent interviews administered, FACES data
showed that about 20 percent of parents of 3- and 4-year-old children in
Head Start 18 reported that a language other than English was most
frequently spoken at home, and about 13 percent of parents reported that
they spoke English "not well" or "not at all." 19 Second, the Office of
Head Start collected demographic information on all 4- and 5-year-old
children in Head Start 20 through its National Reporting System (HSNRS),
including information on the child's primary language. These data showed
that about one-quarter of children enrolled in Head Start in Spring 2005
had a primary language other than English. 21
18
According to HHS, 3- and 4-year-olds constituted 87 percent of children
enrolled in Head Start during the 2002-2003 program year. (See
http://www.acf.hhs.gov/programs/hsb/research/2004.htm.)
19
While this number appears similar to the percentage of children aged 0 to
5 in low-income families that were in linguistically isolated households
as reported in Census 2000 (14 percent), the two cannot be directly
compared because they were collected in different years and because the
definition of limited English proficiency we used in analyzing the
information from FACES is different from the Census definition of a
linguistically isolated household.
20
According to HHS, 4- and 5-year-olds constituted 58 percent of children
enrolled in Head Start during the 2002-2003 program year. (See
http://www.acf.hhs.gov/programs/hsb/research/2004.htm.)
21
The Office of Head Start also surveys grantees annually through its
Program Information Report (PIR). The PIR asks grantees to report a
variety of demographic information about children enrolled in their
programs, including the primary language of the family at home, but not
their need for language assistance. However, our prior work identified
limitations of PIR data. See GAO, Head Start: Better Data and Processes
Needed to Monitor Underenrollment, GAO-04-17 (Washington, D.C.: Dec. 4,
2003), and GAO, Head Start: Comprehensive Approach to Identifying and
Addressing Risks Could Help Prevent Grantee Financial Management
Weaknesses, GAO-05-176 (Washington, D.C.: Feb. 28, 2005.)
Parents with Limited English Proficiency Faced Multiple Challenges That May
Have Limited Their Children's Participation in Federally Funded Child Care and
Early Education Programs
Results from our focus groups, which were composed of mothers with limited
English proficiency whose children were eligible for federal child care
subsidies, revealed that some participants were unaware of the various
federal child care and early education programs that may be available to
them. Parents with limited English proficiency also faced challenges in
the process of applying for programs and financial assistance, such as
lack of interpreters and translated materials. They also encountered
difficulties communicating with English-speaking child care providers.
Some of the challenges to program access that these parents faced were the
same challenges that many low-income families face, including difficulty
finding care at nontraditional hours, lack of transportation, and the
limited availability of subsidized child care slots.
Lack of Program Awareness and Challenges during the Application Process
May Have Limited Program Participation
Many parents with limited English proficiency were unaware of child care
assistance available to them. All six of the focus groups with
Spanishspeaking and Vietnamese-speaking mothers who were eligible but not
receiving subsidies revealed that the majority were unaware of the
assistance available. In addition, the mothers that we interviewed in
Arkansas and focus group participants in North Carolina also told of
misunderstandings and myths that some parents had regarding the
consequences of participating in government-funded programs. For example,
they had heard rumors that if they applied for child care assistance,
their child might one day be drafted into the armed forces to repay the
assistance they received.
Shortages of bilingual staff also presented challenges to parents with
limited English proficiency applying for subsidies for their children.
State and local officials and providers that we interviewed identified the
availability of bilingual staff as a factor that played a role in the
ability of parents with limited English proficiency to apply for the
subsidies. For example, subsidy administration officials in one rural
county told us that they sometimes had to ask clients to come back because
no staff were available to assist them in their language. In three of the
four focus groups with Spanish-speaking mothers with subsidies, those who
generally found the subsidy application process to be easy cited the
availability of bilingual case workers as a factor in allowing them to
apply for assistance successfully.
In addition to shortages of bilingual staff, the lack of available
translated materials also presented challenges to parents with limited
English proficiency. Some programs did not have application forms
translated into other languages, and local officials and parents expressed
concerns about the quality of existing translated materials, saying that
they were often translated by volunteers and that no quality checks were
done. For example, one community group representative told us that
volunteers had translated the Spanish forms that the local subsidy
administration office used and that no quality controls had been applied,
resulting in materials of such poor quality that she advised parents not
to request the Spanish version of the application.
These challenges may be more acute for individuals with limited English
proficiency who speak languages other than Spanish. Local officials in
three states reported that there were limited services available in
languages other than Spanish. For example, local officials in Washington
said that services to smaller, more diverse populations, such as African,
Asian, and East Indian language speakers, were more limited. In North
Carolina and California, local officials also reported that services for
populations such as the Hmong were more limited than for English or
Spanish speakers.
Finally, although immigration status has no impact on Head Start
eligibility and only the immigration status of the child is relevant to
the determination of eligibility for CCDF subsidies, it nonetheless
created indirect challenges for some children of parents with limited
English proficiency. Local officials and community advocates told us that
citizen children of parents with limited English proficiency might not
participate in federal child care and early education programs because of
fear within the family of exposing undocumented immigrant members in the
household. Several officials told us that some of these families were
reluctant to provide personal information and were inhibited from applying
because of fear about how their personal information might be used. In one
case, we discovered a state that improperly required a declaration of
satisfactory immigration status for every member of the household in order
to apply for federally funded child care subsidies, thereby potentially
excluding some children who are U.S. citizens and otherwise eligible for
subsidies. Officials in two states also told us that many parents with
limited English proficiency were paid in cash, making it difficult to
verify their income for eligibility purposes.
Parents with Limited English Proficiency Had Difficulties Communicating
with Providers
Parents reported difficulties communicating with their children's
providers, and officials reported shortages of providers with the language
ability to serve families with limited English proficiency. For example,
officials at one local resource and referral agency that we visited in the
county with the most Spanish speakers in the state told us that providers
in the county did not have the capacity to meet the needs of families with
limited English proficiency. Spanish-speaking mothers that we interviewed
during a site visit to another state complained that some programs
advertise themselves as bilingual when in reality they are not. Parents in
focus groups also expressed concern about their ability to communicate
with their child care providers. Local officials in one urban area that we
visited said that among the primary challenges faced by families with
limited English proficiency was the effect of the language barrier on the
parents' ability to communicate with their child care providers. They
stated this also made it difficult to ensure the same level of
parentprovider interaction for families with limited English proficiency
as for other families. For example, one provider with no bilingual staff
said that she had a child with a disability in her center whose parents
were limited in their English proficiency, making it difficult for staff
to communicate with the parents about the child's needs. These
communication difficulties had consequences in the classroom as well. For
example, one Head Start provider reported instances of therapists and
educators who were not trained to work with Hispanic families inaccurately
assessing the needs of children with language or cultural differences.
Low-Income Parents with Limited English Proficiency Faced Some of the Same
Challenges to Program Access as Other Low-Income Families
Low-income parents with limited English proficiency faced some of the same
challenges when attempting to access child care and early childhood
education programs as other low-income families. Across all states
visited, state and local officials as well as providers said that many
low-income families, especially families with limited English proficiency,
work nontraditional hours and have difficulty finding care that meets
their needs. For example, a resource and referral agency official in one
rural community said that the first shift at a local employer begins at
5:30 a.m., while most providers do not offer care before 6:00 a.m., and
employees working second and third shifts face even more difficulty
finding child care. Lack of transportation, especially in rural
communities, also restricts the child care options available to low-income
families. Officials said that it can be especially difficult for families
with limited English proficiency to navigate public transportation or call
transit agencies for assistance. In a previous report, we found that lack
of English skills reduced individuals' ability to access public
transportation systems. 22
Parents in some communities also faced shortages of child care and child
care subsidies, especially for infants and toddlers. Officials with
resource and referral agencies and local subsidy administration offices in
6 of the 11 counties that we contacted said that there were shortages of
infant care in their communities. In addition, because funding for CCDF
subsidies was limited, not all states provided subsidies to all families
who applied and met eligibility criteria. Our prior work showed that 20
states did not serve all families who met state-determined eligibility
criteria, 23 and three of the five states that we visited (Arkansas,
California, and North Carolina) had waiting lists for CCDF subsidies. In
five of the eight focus groups with Spanish-speaking mothers (including
both those receiving and not receiving subsidies), participants identified
waiting lists as one of the difficulties they faced when seeking
assistance for child care. In the two other states that we visited
(Illinois and Washington), state officials said that although they did not
maintain waiting lists, they spent all of the funds available to them for
CCDF subsidies. To manage demand for the limited financial assistance
available for child care, states took steps such as giving priority to
certain groups. For example, in the three states we visited that
maintained waiting lists, two (Arkansas and North Carolina) set priorities
for eligible families, such as preferences for families on or coming off
of TANF. In the third, California, families on or transitioning off of
TANF were provided child care assistance through a guaranteed funding
stream, while funding for other low-income families was capped. Officials
in California told us that this system made it extremely difficult for
low-income families that were not in the TANF system to receive subsidized
child care. While prioritization of TANF families would affect all
low-income families, it may have additional implications for some children
of parents with limited English proficiency. Census 2000 data show that 82
percent of individuals with limited English proficiency are foreign-born,
and since immigration status is a factor in TANF eligibility, children of
immigrants who do not qualify for TANF would be less likely to receive
CCDF subsidies in those states that give priority to TANF families. In
2005, we found that 17 of 20 states not covering all applicants who
otherwise met the eligibility criteria gave TANF families priority for
CCDF funds, 24 consistent with CCDF's goal of providing child care to
parents trying to become independent of public assistance.
See GAO, Transportation Services: Better Dissemination and Oversight of
DOT's Guidance Could Lead to Improved Access for Limited
English-Proficient Populations, GAO-06-52 (Washington, D.C.: Nov. 2,
2005).
23See GAO-05-667.
Page 24 GAO-06-807 Child Care and Early Childhood Education
Selected Agencies Took Some Steps to Assist Parents with Limited English
Proficiency but Reported Challenges in Serving Them
The majority of state and local agencies and providers that we visited
took some steps to assist parents with limited English proficiency in
accessing child care and early education programs for their children. Most
agencies provided some oral and written language assistance, although the
scope of the assistance varied. Most agencies also implemented initiatives
to increase the supply of providers able to communicate effectively with
parents. Officials told us that they faced several challenges in providing
services to parents with limited English proficiency. Some state and local
officials indicated that additional information on cost-effective
strategies used by others to serve this population would facilitate their
efforts to provide access.
Selected State and Local Agencies Offered Language Assistance
The majority of the agencies that we visited had taken some steps to
provide oral and written language assistance, such as interpreters and
translated materials, to parents with limited English proficiency. In all
11 counties that we contacted, the local offices administering CCDF
subsidies and providing resource and referral services offered some oral
language assistance to clients with limited English proficiency although
the scope of the assistance varied. In 5 of these counties, agencies had
staff that could speak several languages, a fact that officials said
reflected the community they served. In the other 6 counties, agency staff
had bilingual capacity for Spanish only, but officials said the vast
majority of the individuals with limited English proficiency they served
were Spanishspeaking. Although the subsidy administration office in one of
these 6 counties had bilingual Spanish-speaking staff, these staff were
not specifically assigned to work with individuals applying for CCDF
subsidies but were clerical workers with other responsibilities. In most
counties visited, child care and Head Start centers had bilingual staff to
help parents with limited English proficiency enroll their children in the
programs. For example, an official in one child care center that we
visited where the majority of the families spoke Spanish said that all
staff responsible for enrolling families in the program spoke Spanish.
24
See GAO, Child Care: Additional Information Is Needed on Working Families
Receiving Subsidies, GAO-05-667 (Washington, D.C.: June 29, 2005).
Page 25 GAO-06-807 Child Care and Early Childhood Education
Several agencies that we visited also used telephone interpretation
services to provide oral assistance to clients with limited English
proficiency. 25 For example, the subsidy administration offices that we
visited in Washington primarily used a state-contracted telephone language
line that connected agency staff with bilingual telephone operators who
could offer interpreting assistance in a language spoken by the client. In
an effort to help local agencies serve clients with limited English
proficiency in a cost-effective manner, North Carolina was in the process
of entering into a contract for a language line that would allow local
social service agencies, including those administering CCDF subsidies, to
provide oral language assistance to clients if bilingual staff were not
available on-site. A state official told us that once the contract is
awarded, the state will make the service available to all local social
service agencies at a reduced cost.
Several agencies also coordinated with one another to share resources for
offering oral language assistance. For example, to help interpret for
their Russian-speaking clients, a resource and referral agency in
California with language capacity in Cantonese and Mandarin coordinated
with staff at another nearby resource and referral agency that had
language capacity in Russian. Subsidy administration officials in one
rural county that we visited told us that the local hospital had a
contract for the language line and they coordinated with the hospital to
make use of that service. However, we did not find efforts to coordinate
language assistance strategies among agencies in some locations visited,
and agency officials in a few locations said that they could not always
provide oral language assistance to clients with limited English
proficiency on their own.
The majority of agencies that we visited provided written language
assistance, such as translated subsidy application forms. Seven of the 11
subsidy administration offices contacted had subsidy applications
translated into Spanish. Local agencies in Washington, California, and
Illinois had applications that had been translated by the state.
Washington required its application for the child care subsidy to be
translated into
Through a contract with organizations providing telephone interpretation
services, agency staff typically can dial a telephone number provided by
the organization and request to be connected to a professional interpreter
speaking a particular language. The interpreter, proficient in both
English and another language, listens to the conversation between the
staff and the client with limited English proficiency, analyzes the
meaning of the message, and conveys the meaning to each side.
Page 26 GAO-06-807 Child Care and Early Childhood Education
eight languages, 26 while California and Illinois made applications
available in Spanish and gave local agencies the option of translating
materials into other languages. Arkansas and North Carolina had no
translated applications at the time of our visits, although officials in
North Carolina said that the state was in the process of translating the
subsidy application into Spanish. 27 All of the resource and referral
agencies that we visited translated materials into Spanish, such as
brochures containing information on how to receive child care assistance
and what to look for when choosing a provider. A few resource and referral
agencies also made efforts to translate written information into other
languages. For example, as shown in figure 4, one agency translated a
brochure on child care quality into Chinese. However, some state and local
officials told us that their offices lacked the resources to translate
materials into other languages.
26
Applications in Washington were available in Spanish, Chinese, Korean,
Russian, Cambodian, Vietnamese, Somali, and Laotian.
27
In June 2006, a North Carolina official told us that the translation of
the CCDF subsidy application into Spanish has been completed. Local
agencies currently have access to the translated document, and the North
Carolina Department of Health and Human Services is in the process of
making both the English and the Spanish versions of the document available
electronically for their use.
Figure 4: English and Chinese Versions of a Local Agency's Child Care
Quality Brochure
The majority of local agency officials and providers that we interviewed
told us that they relied on agency staff and volunteers to translate
materials. For example, officials from a Head Start program told us that
their staff had translated materials about the program into Spanish,
Hmong, and Laotian. Officials at another Head Start program told us that
they relied on bilingual staff, parents of children enrolled in the
program, and Spanish-speaking volunteers from the community health clinic
to translate the materials. Some agency officials told us that they also
used outside contractors or other resources, such as commercially
available translation software, to translate materials. Community group
representatives expressed concerns about the quality of translations done
by the local agencies, particularly in instances when volunteers or
translation software had been used.
Most local agencies and providers that we interviewed said that they
disseminated translated information to raise awareness of their programs
and services among parents with limited English proficiency. Agencies and
providers employed various mechanisms to disseminate information,
including using print and radio media and direct distribution of
informational materials in the communities where many families with
limited English proficiency reside. For example, some resource and
referral agencies and providers said that they advertised their programs
and services on Spanish-language television and radio stations, and a few
agencies had placed advertisements in the Yellow Pages. Most of them also
reported distributing information in various locations in the community,
such as churches, neighborhood markets, and laundromats.
Despite these agencies' various outreach efforts, mothers in focus groups,
many of whom were unaware of the available assistance, said that there was
a need for greater information dissemination in their communities.
Spanish- and Vietnamese-speaking mothers in all 12 focus groups indicated
that disseminating information in their language would help them learn
about child care assistance and child care and early education programs
for their children. At the same time, focus groups with Spanishspeaking
mothers in California who were already receiving the subsidies revealed
their ambivalence about increased advertising of certain child care
programs because some of these programs already had waiting lists. Some
state and local officials also acknowledged that they did little or no
advertising because their programs were already operating at full capacity
or had substantial waiting lists.
Selected Agencies Took Steps to Increase the Supply of Providers Able to
Communicate Effectively with Parents
Agencies in the majority of locations that we visited had initiatives to
increase the supply of providers who spoke other languages or to offer
training in other languages to existing providers. Some agencies had come
up with initiatives that focused on helping individuals speaking other
languages to enter the child care field. For example, one resource and
referral agency that we visited offered the classes required for obtaining
a child care license in Spanish, and another one offered them in
Cantonese. A resource and referral agency that we visited in an urban
county developed a program to help Somali- and Russian-speaking women in
the community obtain the training necessary to become licensed family
child care home providers. In four of the five states that we visited,
officials told us that selected community colleges participated in efforts
to increase provider capacity to serve children of parents with limited
English proficiency. For example, a community college in Illinois offered
early childhood education classes in Spanish, while a community college in
California coordinated with a local resource and referral agency to offer
these classes in Cantonese. However, some officials said that such efforts
were insufficient, and in one state visited, an official from a university
early childhood education program said that she was not aware of any
efforts in the state to offer classes in other languages.
Many agencies that we visited also provided training to existing child
care providers who had limited English proficiency. For example, local
referral agencies in Illinois included bilingual individuals in the
technical assistance teams available to assist providers in improving the
quality of care.
Three of the five states that we visited used CCDF quality funds for
various provider initiatives related to language, such as offering
training to providers on working with families that had limited English
proficiency or translating materials into other languages. 28 For example,
Arkansas used quality funds for training and technical assistance to help
providers understand cultural issues that families with limited English
proficiency face. California used these funds to offer training to
providers throughout the state on working with children who speak other
languages. Officials in North Carolina said that while they did not have
any projects funded with CCDF quality funds that directly related to
serving children of parents with limited English proficiency, they had
used some of the funds to translate materials on child care health and
safety practices into Spanish. Two of the states visited-Washington and
Illinois-did not use CCDF funds directly on initiatives related to serving
children of parents with limited English proficiency or providers working
with them. However, both states used the funds to support other
initiatives, such as the work of resource and referral agencies, which
included outreach to parents with limited English proficiency in some of
their efforts.
State and local officials told us that despite efforts made, there was a
shortage in some locations of training opportunities for providers who
speak other languages. For example, officials across states and counties
that we visited cited examples of child care providers with limited
English proficiency who had attended training, such as training required
for licensing, although they could not fully understand the course content
because it was not available in their languages. An official we
interviewed told us that this could affect the quality of child care they
would offer to children because the training covered critical issues, such
as health and safety procedures.
States are required to describe in their CCDF plans how CCDF quality funds
will be used, but are not required to use them for initiatives focused on
providers serving children of parents with limited English proficiency.
Page 30 GAO-06-807 Child Care and Early Childhood Education
Officials in Selected State and Local Agencies Reported Challenges in
Providing Services to Parents with Limited English Proficiency
State and local agency officials, providers, and community college
representatives reported several challenges associated with providing oral
language assistance to parents with limited English proficiency applying
for child care and early education programs for their children. Officials
told us they faced challenges providing oral language assistance because
of the difficulties that agencies had hiring qualified bilingual staff.
Even when qualified bilingual individuals were found, officials said that
these individuals were in very high demand and agencies could not always
compete with other organizations interested in hiring them. For example,
some child care and Head Start providers told us that they are losing
qualified bilingual staff to school districts that offer higher salaries.
Rural areas especially experienced difficulties hiring bilingual staff
because their pool of qualified candidates was smaller than in the cities
or virtually nonexistent. A few officials said that the lack of reliable
transportation in rural areas makes it difficult to recruit staff from the
cities. For example, a resource and referral agency official in one rural
area that we visited told us that her office's bilingual staff had quit
because they had difficulty getting to work. Officials also cited
difficulties with finding professional interpreters and with the expense
associated with hiring them when agencies lacked bilingual staff of their
own to offer oral language assistance to clients.
Agency officials also reported challenges providing written language
assistance to parents with limited English proficiency. They said that
translating materials into other languages was expensive, particularly for
agencies that served clients from several different language groups and
had to translate materials into multiple languages. Local agencies
frequently relied on their own staff to translate the materials, but a few
officials said that this posed a burden on staff with other full-time
responsibilities. At the same time, state and local officials said that
contracting out for translations was expensive. Although state officials
acknowledged the expense associated with translating materials into other
languages, some states left local agencies to shoulder the burden of
translating documents on their own. For example, state officials in
California told us that the expense prevented the state from translating
applications into languages other than Spanish, but local agencies had
absorbed the cost of translating applications themselves in order to meet
the needs of program applicants who spoke other languages.
In addition, officials said that providing language assistance or training
in other languages was not always cost-effective because of the relatively
small number of individuals that would benefit from such efforts. For
example, one resource and referral agency official told us that the cost
of ordering materials in Spanish was higher than the cost of ordering the
same materials in English because the materials had to be purchased in
smaller orders, thereby increasing their cost. Some officials said that
while they were able to offer language assistance to larger language
groups in the area, such as Spanish speakers, they chose not to expand
their assistance to include other language groups because of the small
number of individuals that would benefit from it.
Despite challenges faced, agency officials that we interviewed expressed
the need for effective and affordable ways to provide services to
individuals with limited English proficiency. Officials in three states
visited told us that they would benefit from having additional information
on costeffective strategies to serve parents with limited English
proficiency. Several officials also told us that it would be helpful for
them to learn more about the professional development opportunities for
providers offered at other locations. For example, officials in Illinois
said that the state's current capacity for provider training in Chinese
was limited and that they would like to learn more about any curricula
developed in other states with larger Asian populations.
HHS Provided Assistance to Grantees on Serving Children of Parents with
Limited English Proficiency, but Gaps Remain in Its Program Review Efforts
HHS issued general guidance, translated materials, and provided technical
assistance to grantees on serving children of parents with limited English
proficiency, but gaps remain in its program review efforts. The Office of
Head Start has provided assistance to increase awareness of the Head Start
program among families with limited English proficiency and has monitored
local programs' efforts to provide access to these families by reviewing
grantees' biennial assessments of need in the communities they serve and
by conducting formal monitoring reviews of grantees. However, an Office of
Head Start official told us that the office could not ensure that its
review teams consistently reviewed grantee compliance with program
standards related to language access, and in our prior work we found that
no mechanism existed to ensure consistency in the monitoring process. CCB
provided assistance to help programs serve children whose parents have
limited English proficiency, as well as reviewed states' CCDF plans and
investigated complaints. However, CCB had no mechanism for reviewing how
access to CCDF subsidies was provided for children of
parents with limited English proficiency or for ensuring that these
children were not inadvertently excluded from the subsidies as a result of
state eligibility criteria that were inconsistent with CCB's program
eligibility guidance.
HHS Issued General Guidance to Grantees on Providing Access to Federal
Programs for Individuals with Limited English Proficiency
In 2003, consistent with Executive Order 13166, HHS issued guidance to
federal financial assistance recipients regarding the Title VI prohibition
against national origin discrimination as it affects individuals with
limited English proficiency. The guidance was intended to help recipients
of HHS funds, such as agencies administering CCDF subsidies and Head Start
programs, provide meaningful access for individuals with limited English
proficiency. The guidance, however, applied to all HHS programs and did
not refer specifically to child care or early education.
HHS' OCR provided outreach to potential beneficiaries of HHS programs and
offered technical assistance to grantees to help them comply with the
guidance. For example, OCR officials told us that they disseminated
information about serving individuals with limited English proficiency at
Hispanic health fairs, through recorded public service announcements and
interviews on Spanish-language media, and by giving presentations before
community service organizations. They also said that they provided
grantees with technical assistance in identifying appropriate language
access strategies. Regional OCR officials told us that their offices
served as a resource for local social service agencies, directing them to
less costly language access strategies, such as sharing interpreter
services, and providing information on available resources and practices.
OCR also participated in the Federal Interagency Working Group on Limited
English Proficiency that developed, among other things, a Web site devoted
to serving persons with limited English proficiency (www.lep.gov). The Web
site serves as a clearinghouse, providing information, tools, and
technical assistance regarding limited English proficiency and language
services for federal agencies, recipients of federal funds, users of
federally assisted programs, and other interested parties. It makes
available a range of guidance and information on offering language
assistance through mechanisms such as interpreter services and translated
materials for clients with limited English proficiency in the areas of
health care, the courts, and transportation. However, it does not include
specific information on providing language assistance in child care and
early education programs. In addition, CCB and Office of Head Start
officials and officials from several HHS regional offices told us that
they were unaware of the Web site.
OCR is required to investigate all complaints of alleged discrimination,
including lack of access to programs for individuals with limited English
proficiency. OCR officials told us that Title VI violations in child care
were rare. They said that when infractions do occur, they try to reach a
voluntary compliance agreement with the state and conduct follow-up to
ensure that the state takes corrective action to comply with the terms of
the agreement. For example, North Carolina entered into a voluntary
compliance agreement with OCR and implemented a corrective action plan for
providing access for program applicants with limited English proficiency.
A state official told us that the state was in the process of translating
the subsidy application into Spanish as a result of this agreement.
The Office of Head Start Provided Assistance to Increase Awareness of Head
Start and to Improve Service Delivery and Conducted Limited Monitoring of
Language Access in Head Start Programs
The Office of Head Start has provided a variety of assistance to increase
awareness of the Head Start program among families with limited English
proficiency. The office has twice hosted a National Head Start Hispanic
Institute, the goals of which included improving outreach to Hispanic
communities, developing methods to effectively serve Hispanic children and
families, and helping ensure positive outcomes in language and literacy
development for English-language learners. A Head Start official told us
that the needs of other language groups needed to be addressed as well,
and that the Office of Head Start was considering how to replicate the
institute for groups that speak other languages. According to officials,
the Office of Head Start has several other initiatives to reach parents
with limited English proficiency, such as placing public service
announcements on Spanish-language media and distributing a brochure in
Spanish informing families potentially eligible for Head Start of the
benefits of enrolling their children in Head Start.
The Office of Head Start has also provided assistance to grantees to
better serve children of parents with limited English proficiency.
Recently, the office conducted a national language needs assessment of
second language and dual language acquisition to identify culturally
responsive, research-based strategies to improve outcomes for children and
families. It also developed a Culturally Responsive and Aware Dual
Language Education (CRADLE) training initiative that is designed to
support grantees in their efforts to find best practices for language
acquisition for the birth-to-3-year-old population. In addition, through
its English Language Learners Focus Group, the Office of Head Start
created materials for grantees working with second language learners,
including Spanish speakers who constitute the majority of children in Head
Start whose parents have limited English proficiency.
The Office of Head Start monitors grantees' efforts to provide access for
individuals with limited English proficiency by reviewing their biennial
community assessments and conducting formal on-site monitoring reviews.
Head Start programs are required to conduct a community assessment at
least once every 3 years, and the Office of Head Start regional officials
review these assessments for demographic disparities between program
participants and the population of the community to be served. For
example, programs with assessments showing large numbers or proportions of
language groups in the community that are not reflected in the enrollees
or the classroom teachers may be found out of compliance with meeting
local needs. Head Start programs are also monitored by the Office of Head
Start once every 3 years through the PRISM process. Head Start programs
are required to adhere to program performance standards that define the
services that programs are to provide to children and their families, and
on-site PRISM review teams monitor Head Start grantees' adherence to the
standards. Several of the standards directly address interactions with
children and parents with limited English proficiency. For example, one
performance standard requires communications with parents to be carried
out in the parent's primary or preferred language or through an
interpreter. 29 Another performance standard directs programs in which the
majority of children speak the same language to have at least one
classroom staff member or home visitor who speaks that language. 30 The
contractor responsible for assigning bilingual reviewers to PRISM review
teams told us that about 17 percent of reviewers were bilingual and that
review teams requesting a Spanish-speaking bilingual individual had one
assigned 70 percent of the time.
A Head Start official with responsibility for the PRISM process told us
that given the vast number of regulations, however, it was impossible to
ensure that all of them were consistently reviewed in the course of a
1-week review. In our previous work, we reported that ACF had no process
in place to ensure that its reviewers consistently followed the standards
while conducting on-site PRISM reviews. 31 We recommended that ACF develop
an approach that can be applied uniformly across all of its
regional offices to assess the results of the PRISM reviews and implement
a quality assurance process to ensure that the framework for conducting
on-site reviews was implemented as designed. HHS agreed with our
recommendation, and Head Start officials indicated that the Office of Head
Start was developing new PRISM protocols and training reviewers to add
more uniformity to how grantees are assessed. In addition, the Office of
Head Start recently announced plans to conduct follow-up reviews of
grantees monitored through the PRISM system in an effort to ensure that
PRISM review teams did not miss grantee deficiencies, such as in providing
assistance to children and parents with limited English proficiency.
29
45 CFR S:1304.51(c)(2)
30
45 CFR S:1304.52(g)(2). In fiscal year 2004, the Office of Head Start
found three programs in noncompliance with this performance standard.
31GAO. Head Start: Comprehensive Approach to Identifying and Addressing
Risks Could Help Prevent Grantee Financial Management Weaknesses,
GAO-05-176 (Washington, D.C.: Feb. 28, 2005).
Page 35 GAO-06-807 Child Care and Early Childhood Education
CCB Provided Assistance to Help CCDF Programs Serve Children Whose Parents
Have Limited English Proficiency but Had No Mechanism for Reviewing
Agencies' Provision of Access
CCB provided assistance to raise program awareness among parents with
limited English proficiency whose children may be eligible for CCDF
subsidies. Officials told us that CCB had translated a number of its
consumer education materials into Spanish, including the CCDF program
brochure and public service announcements informing parents where and how
to locate child care. In a targeted effort to reach Hispanic families and
providers, CCB also translated into Spanish a brochure outlining what
providers should know about child care assistance for families. CCB,
through a cooperative agreement with the National Association of Child
Care Resource and Referral Agencies (NACCRRA), provides educational
information to parents through the Child Care Aware Web site
(www.childcareaware.org). In addition, NACCRRA has translated consumer
education publications into Spanish, including a publication on paying for
child care, which it made available through its Web site to resource and
referral agencies nationwide. CCB officials told us that they were also
looking into translating these publications into Chinese. CCB also
sponsors a National Child Care Information Center Web site
(www.nccic.org), which offers information on a wide range of child care
issues, including a number of documents that relate to serving children
from families with limited English proficiency.
CCB officials told us that they provided opportunities for agencies and
providers to share information, including information on serving children
of parents with limited English proficiency. For example, CCB convened
meetings of state CCDF administrators that, while not focusing
specifically on issues of limited English proficiency, covered topics such
as meeting the needs of diverse groups of children and parents. In
addition, CCB maintains an online forum for states to pose questions and
share ideas, which has been used to discuss such issues as converting
print materials into Spanish. CCB also offers child care providers online
access to training modules, practical strategies for serving children and
families, and interactive online chats in English and Spanish through the
Center on Social and Emotional Foundations for Early Learning Web site
(www.csefel.uiuc.edu).
While it has made efforts to assist states with serving the needs of
children whose parents have limited English proficiency, CCB has no
mechanism for reviewing how agencies provide access to CCDF subsidies for
eligible children of parents with limited English proficiency or ensuring
that these children are not inadvertently excluded as a result of state
CCDF eligibility criteria that are inconsistent with agency guidance. CCB
officials told us that CCDF is a block grant and CCB receives no funding
specifically for supporting monitoring activities. As a result, CCB's
oversight of CCDF is limited to reviewing states' CCDF plans and
investigating complaints. CCB, however, does not require states to include
assurances in their CCDF plans that state agencies are providing access to
CCDF subsidies for children of parents with limited English proficiency.
Regional officials told us that they had complaint processes in place and
would either review complaints or refer them to OCR, but said that they
were unaware of any complaints regarding restricted access for individuals
with limited English proficiency. Officials in one region told us that
states appeared to understand the CCDF program eligibility criteria.
Officials in another region told us that while they interacted with states
through phone calls and occasional on-site visits, these contacts
primarily focused on the provision of technical assistance. Thus, these
interactions were not a systematic review of how states determine
eligibility for federal child care assistance.
On our site visit to Arkansas, we found that the state had eligibility
requirements that appeared to violate CCB guidance. Specifically, although
guidance to state agencies administering CCDF clarified that only the
citizenship and immigration status of a child was relevant when
determining the child's eligibility for federal child care assistance,
applicants for child care assistance in Arkansas had to submit a
declaration that the applicant (typically a parent applying to receive
assistance for the child) and all the other members of the household were
U.S. citizens, nationals, or legal residents. In addition, the state's
policy manual for the administration of CCDF services indicated that the
state would deny any applications for child care assistance that were
submitted by parents or custodians who were neither citizens nor lawfully
admitted residents. These requirements have the potential of precluding
children who otherwise met the eligibility criteria from receiving federal
financial assistance on the basis of their parents' citizenship or
immigration status.
CCB officials told us that they were unaware of the situation until we
brought it to their attention and that they were in the process of
discussing with state officials how to resolve it. They further noted that
they would investigate formal complaints brought to their attention, which
would include complaints about states requesting unnecessary information
on their child care subsidy applications and adversely affecting
individuals with limited English proficiency. However, officials indicated
that they had received no such complaints from affected parties.
Conclusions
Access to high-quality child care and early education programs helps
promote healthy development of children and can provide an important
support for parents as they pursue employment or education to secure the
family's economic well-being and avoid public assistance. The resources
available for nonentitlement child care and early education programs, such
as CCDF subsidies and Head Start, are limited and not intended to cover
everyone who meets eligibility criteria and is in need of assistance.
Consequently, agencies have to make choices about who they will cover with
the limited funds, employing strategies such as prioritization of certain
groups of applicants or waiting lists. At the same time, federal, state,
and local entities play important roles in ensuring that parents' language
ability does not preclude children from being considered for coverage
under these programs.
These roles are becoming especially important as the demographics of many
communities are changing rapidly and localities across the country are
seeing increased numbers of individuals with limited English proficiency.
While state and local agencies are making efforts to address the needs of
this growing population, they experience difficulties offering language
assistance to parents seeking to access programs for their children and
recruiting new providers with the language ability to serve these
families. However, without reliable data on who is enrolled in their
programs, state and local officials may have difficulty determining the
extent to which parents with limited English proficiency have access to
these programs for their children and whether services need to be adjusted
to accommodate changes in the population served.
Although Congress provided states with flexibility in administering their
CCDF program grants, HHS is responsible for ensuring that states adhere to
the conditions of their grants and that they take reasonable steps to
ensure access to individuals with limited English proficiency. Yet, HHS's
existing methods for reviewing how CCDF funds are used by grantees do not
systematically assess how access for parents with limited English
proficiency is provided or identify state or local policies that may
adversely affect these parents' ability to access programs for their
children. HHS responds to complaints of any alleged discrimination or
agency actions that adversely affect the ability of eligible children to
access programs and services. However, HHS may lack the tools to ensure
equal access for children whose parents have limited English proficiency
if the parents do not bring complaints for reasons such as language
difficulties, unfamiliarity with how the complaint process works, or fear
about approaching government agencies. Without a mechanism to
systematically review access to CCDF-funded programs for these families,
HHS cannot provide all eligible children with the same opportunity to
participate in programs that would benefit them and their families and
possibly enhance their households' self-sufficiency.
Recommendations for Executive Action
To help state and local agencies plan for language assistance and assess
whether they provide meaningful access to eligible children, regardless of
their parents' English ability, we recommend that CCB work with states to
help them explore cost-effective strategies for collecting data on CCDF
subsidy recipients' language preference or English proficiency and
comparing these data with available information on community demographics.
Once these data are available, HHS may consider collecting information on
existing cost-effective ways for agencies to provide language assistance
and to recruit providers who speak other languages, as well as
disseminating this information in the locations where the data show the
greatest need.
To provide opportunities to parents with limited English proficiency to
access federal child care subsidies for their children, we recommend that
HHS develop and implement specific steps to review whether and how states
provide access to CCDF programs for eligible children of parents with
limited English proficiency, as well as provide information to help states
evaluate their progress in this area. Specifically, HHS should
o revise the CCDF plan template to require states to report on how they
will provide meaningful access to parents with limited English
proficiency seeking CCDF subsidies for their children, and
o systematically review states' program eligibility criteria for CCDF
subsidies to ensure that states comply with HHS policies related to
participation by children of parents with limited English proficiency.
Agency Comments and Our Evaluation
ACF provided written comments on a draft of this report, which are
reproduced in appendix III. In its letter, ACF agreed with most aspects of
our recommendations and provided information on its actions or plans that
would support their implementation. In addition, ACF provided a number of
technical comments that we incorporated as appropriate.
In response to our recommendation that HHS work with states to help them
explore cost-effective ways of collecting data on the primary language of
CCDF subsidy recipients, ACF provided some additional information on
actions it has taken to help states in this area. For example, it stated
that in July 2006, CCB launched a technical assistance initiative that
will, among other things, disseminate information to states on effective
strategies to assist families with subsidy access, including families
experiencing language barriers.
Regarding our second recommendation, that HHS develop a mechanism to
review how states provide access to CCDF subsidies for children of parents
with limited English proficiency, ACF indicated that it will examine the
feasibility of using the CCDF plan template to ask states to report on
their efforts to promote access to these families. However, ACF did not
address our recommendation that HHS systematically review states'
eligibility criteria for CCDF subsidies to ensure that states comply with
HHS policies related to participation by children whose parents have
limited English proficiency.
ACF also submitted detailed comments related to our analysis of national
survey data collected in 1998 as part of ECLS-K. ACF noted that ECLS-K
data only provide information on children in the year before kindergarten
and that the analysis omits other variables that may explain our findings,
such as preferences for certain types of care within ethnic communities
and parents' immigration status. Our report discusses these data
limitations, and as is the case with any statistical model, some of the
factors with the potential to affect the outcomes we examined could not be
included because the data measuring them were not collected. It is partly
for that reason that we employed multiple methodologies in addressing our
research objectives, including site visits and focus groups.
ACF noted that the data represent child care and early education patterns
for 1997 and that subsequent policy changes or increases in federal and
state child care funding, may have narrowed the gap in program
participation among different groups of children. However, we found that
some of the policy changes ACF cited were not consistently implemented and
ACF provided no more current data that would allow us to ascertain the
effects of these changes. As such, ECLS-K remained the most recent
national dataset that allowed us to compare children of parents with
limited English proficiency and similar children whose parents are
proficient in English with respect to their receipt of financial
assistance for child care from a social service or welfare agency and
their participation in Head Start.
As arranged with your office, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days
after the date of this letter. At that time, we will send copies of this
report to the Secretary of HHS, relevant congressional committees, and
other interested parties. We will also make copies available to others
upon request. In addition, the report will be made available at no charge
on GAO's Web site at http://www.gao.gov.
If you or your staff have any questions about this report, please contact
me at (202) 512-7215. Other contacts and major contributors are listed in
appendix IV.
Marnie S. Shaul, Director Education, Workforce, and Income Security Issues
Appendix I: Scope and Methodology
In conducting our work, we employed multiple methodologies, including a
review of available data on participation of children in child care and
early education programs, state and county site visits, focus groups with
mothers who have limited English proficiency, interviews with federal
officials and national experts, and a review of available legislation,
guidance, and other federal resources. We performed our work in accordance
with generally accepted government auditing standards between July 2005
and June 2006.
Analysis of National Program Participation Data and State Data Inquiries
To obtain information on the participation of children whose parents have
limited English proficiency in child care and early education programs
funded through the Child Care and Development Fund (CCDF) and Head Start,
we obtained and reviewed the most recent program participation data from
the U.S. Department of Health and Human Services (HHS), surveyed states
about their data on CCDF subsidy recipients, and analyzed national survey
data available through the Early Childhood Longitudinal Study,
Kindergarten Class of 1998-99 (ECLS-K). The relevant characteristics of
data sources we examined are shown in table 1.
Appendix I: Scope and Methodology
Table 1: Characteristics of Data Sources Examined
Ability to use the data Reasons why data cannot
to estimate participation be used to estimate
Availability of data on rates by children of participation rates by Source
of Programs Scope of data the language of parents with limited children of
parents with data covered collection program participants English
proficiency limited English proficiency
CCB CCDF Enrollment or
databases sample of program participants (depending on state)
None No CCB does not collect data related to language in its monthly or
annual reports from states.
State CCDF Enrollment or Varies by state
databases sample of program participants (depending on state)
No Approximately one-quarter of states collect data; their data have many
limitations, and states collect data differently.
NRS Head Start Enrollment, for all Child speaks language No NRS data are
collected only 4- and 5-year-old other than English at from children
enrolled in Head children in Head home, child's primary Start, so the
participation rate Start language, and child's in the overall population
is
English proficiency as unknown.
determined by local Data are not available on the staff parents' English
proficiency.
FACES Head Start Sample of parents Parent selfwith 3 and 4-year-assessment
of old children in language ability and Head Start language spoken at
home
No FACES data are collected only on children enrolled in Head Start, so
the overall participation rate is unknown.
ECLS-K Child Care Sample of children Parent self-No (for CCDF programs)
The survey questions did not assistance and in kindergarten assessment of
Yes (for Head Start ask for the source of child care Head Start language
ability and participation in the year financial assistance.
language spoken at before kindergarten)
home
Source: GAO analysis of HHS program participation data, ECLS-K, and
telephone interviews with state officials.
We reviewed CCDF program participation data collected by CCB in the
reports that states are required to submit on CCDF subsidy recipients but
found that these reports did not contain any data related to language from
CCDF subsidy recipients or their families. CCB officials confirmed that
they do not currently collect any language data, since such data
collection was not listed in the CCDF authorizing legislation. 32
32
CCB collects data from states on the race and ethnicity of subsidy
recipients, but these do not allow for identification of CCDF recipients
speaking other languages.
Page 43 GAO-06-807 Child Care and Early Childhood Education Appendix I:
Scope and Methodology
We reviewed language data for Head Start participants available from the
Office of Head Start through the Head Start National Reporting System
(HSNRS) and the Head Start Family and Child Experiences Survey (FACES).
HSNRS, implemented in August 2003, is the nationwide skills test of over
400,000 children aged 4 and 5 in Head Start, intended to provide
information on how well Head Start grantees are helping children progress.
The Computer-Based Reporting System (CBRS) was developed for HSNRS to
allow local Head Start staff to enter descriptive information about their
programs, including the demographic characteristics of children assessed
by HSNRS. We requested and reviewed HSNRS demographic data from spring
2005 that provided information on the primary language of children in Head
Start. FACES is a series of longitudinal surveys of nationally
representative samples of children in Head Start. We requested and
reviewed fall 2003 FACES data, which included about 2,400 parent
interviews that provided information on the languages spoken at home by
Head Start families, parents' self-reported English proficiency, and the
availability of Head Start staff to communicate with children and parents
in their preferred language.
To assess the reliability of Head Start data, we interviewed relevant HHS
officials and officials from Westat, a private research corporation
administering and analyzing HSNRS and FACES under a contract with the
Office of Head Start. In addition, we reviewed relevant documentation and
examined the logs of the computer code used to generate the data provided
to us. Because HSNRS data were collected only for 4- and 5-year-old
children in Head Start, they cannot be used to generalize about all
children in Head Start. 33 The HSNRS data were entered into CBRS by the
staff of local Head Start programs. While we did not independently verify
these data, we did not find any evidence to suggest that they were
unreliable. As part of FACES, interviews were held directly with parents
of children in Head Start. While Spanish interviewers were available,
parents with limited English skills who spoke other languages were
required to provide their own interpreter. Parents unable to participate
in an interview in English or Spanish or provide their own interpreters
could not be included in the survey. According to a Westat official,
however, only three interviews could not be conducted because of the lack
of an interpreter. We determined that FACES data were sufficiently
reliable for the purposes of this report.
The Head Start program also serves children who are 3 years old.
Appendix I: Scope and Methodology
Because the available agency data did not allow us to determine the total
participation of children of parents with limited English proficiency in
federal child care and early education programs, we also analyzed survey
data provided by NCES from ECLS-K, a national longitudinal study focusing
on following children's early education and school experiences from
kindergarten through 12th grade. We used data from the fall 1998 base year
survey of approximately 18,000 parents with children in kindergarten.
ECLS-K was the most recent national dataset that allowed us to compare
child care, financial assistance for child care, and Head Start usage
rates among children with parents who had limited English proficiency and
children whose parents were proficient in English. 34 Among other topics,
ECLS-K asked parents about their English proficiency, the languages spoken
at home, their child's use of child care in the year before kindergarten,
any financial assistance from a social service or welfare agency, and the
child's use of Head Start. 35 The survey did not ask for the specific
social service or welfare agency providing financial assistance for child
care, so we were unable to make estimates about the use of CCDF subsidies
from this dataset. NCES had bilingual interviewers available to conduct
the survey in Spanish, Chinese, Hmong, and Lakota if the respondent was
not able to speak English and no English-speaking member of the household
was available. Slightly more than 7 percent of the interviewers were
conducted in a language other than English. More information about our
analysis of ECLS-K data can be found in appendix II.
To assess the reliability of ECLS-K data, we reviewed relevant information
about the survey, including the user manual, data dictionary, and steps
taken to ensure the quality of these data, and performed electronic
testing to detect obvious errors in completeness and reasonableness. We
determined that the ECLS-K data were sufficiently reliable for the
purposes of this report.
34
NCES started following a new cohort of children, starting at birth, in
2001. However, the data on their experiences in the year before
kindergarten are not expected to be available until 2008.
35
NCES attempted to verify enrollment for children whose parents reported
that they were in Head Start. While only about half of the enrollments
could be confirmed, NCES concluded that "families with unconfirmed reports
of Head Start participation by their children had demographic
characteristics similar to those of families with confirmed participation
in Head Start. This lends support to the notion that a substantial
proportion of these children had indeed attended Head Start programs, even
though their attendance could not be verified." (Source: "User's Manual
for the ECLS-K Base Year Restricted-use Head Start Data Files and
Electronic Codebook," NCES 2001-025.")
We also contacted child care administrators in all 50 states and the
District of Columbia to determine whether any states collected their own
data on the language of CCDF subsidy recipients. We discussed data
collection with officials in 5 states in the course of our site visits and
contacted officials in the remaining 45 states and the District of
Columbia by e-mail. Of those contacted by e-mail, 40 states and the
District of Columbia responded. Overall, 12 states and the District of
Columbia collected some language data from parents whose children received
CCDF subsidies. We then followed up with officials in the District of
Columbia and all 12 states that reported collecting the data on the
language of CCDF subsidy recipients to ask questions about the type of
data collected, the methods by which the data were collected, the
challenges states faced in collecting the data, and the purposes for which
the data were used. We did not ask states to submit their data to us
because we determined that the differences in states' data collection
approaches and the limitations of state data would preclude us from
aggregating state data to produce national estimates of CCDF subsidy use
among children of parents who speak other languages.
Site Visits
To obtain information on the challenges that parents with limited English
proficiency face in accessing CCDF subsidies and Head Start and the
assistance provided to these families by state and local entities, we
visited 5 states-Arkansas, California, Illinois, North Carolina, and
Washington. We selected these states on the basis of the size and growth
of their population of individuals with limited English proficiency as
determined by our analysis of 1990 and 2000 data from the U.S. Census
Bureau, the states' geographic location, and the presence of initiatives
focused on individuals with limited English proficiency as determined by
our review of CCDF plans that states are required to submit to CCB every 2
years. We visited 10 counties across these states, as well as contacted
officials in 1 county by telephone. We selected counties with substantial
numbers of individuals with limited English proficiency or that have
experienced a significant growth in this population based on the analysis
of 1990 and 2000 U.S. Census data. (See table 2.) In choosing counties, we
also considered the proportion of residents living in urban and rural
parts of the county to obtain information on the experiences of families
in both urban and rural areas.
Appendix I: Scope and Methodology
Table 2: Selected Characteristics of Site Visit Counties
Individuals with limited English proficiency (LEP)
Percentage Percentage
change
Individuals Individuals change in in
with with proportion
of
Number of LEP as a Number of LEP as a number of individuals
with LEP
percentage percentage as a
individuals of individuals of individuals percentage
of
with LEP total with LEP total with LEP population
population population
County (1990) (1990) (2000) (2000) (1990-2000) (1990-2000)
Washington
County,
Arkansas 483 0.5 4,925 3.4 919.7 637.3
Fresno County,
California 66,070 10.9 86,776 11.8 31.3 8.6
Los Angeles
County,
California 1,153,956 14.2 1,395,347 15.9 20.9 11.9
San Francisco
County,
California 86,228 12.5 99,659 13.4 15.6 6.7
Cook County,
Illinois 247,814 5.2 392,663 7.9 58.5 50.1
Winnebago
County,
Illinois 2,510 1.1 6,208 2.4 147.3 123.3
Durham County,
North Carolina 1,330 0.8 8,886 4.3 568.1 442.6
Sampson
County,
North Carolina 377 0.9 2,618 4.7 594.4 451.2
King County,
Washington 27,329 1.9 63,004 3.9 130.5 98.2
Yakima County,
Washington 10,916 6.3 20,686 10.2 89.5 60.4
Chatham
County,
North Carolina a 297 0.8 2,243 4.8 655.2 488.2
Source: GAO analysis of data from the U.S. Census.
a
We contacted officials in this county by telephone.
On each site visit, we interviewed various stakeholders in the child care
and early education field at the state and local levels, including
officials responsible for administering CCDF subsidies, representatives of
child care resource and referral agencies, Head Start officials, and child
care and early education providers, as well as officials from community
organizations and advocacy groups working with individuals who have
limited English proficiency.
Focus Groups
To obtain information on the challenges that parents with limited English
proficiency face when accessing child care subsidies for their children,
we conducted 12 focus groups with mothers who had limited English
proficiency in California, Washington, and North Carolina. We selected
these locations in order to include both states with historically large
populations of individuals with limited English proficiency (California
and Washington) and a state experiencing a more recent growth in this
population (North Carolina)-based on our analysis of data from the U.S.
Census. GAO contracted with Aguirre International, a firm specializing in
applied research with hard-to-reach populations, to recruit focus group
participants through community-based organizations, arrange facilities for
focus groups in locations familiar and accessible to the participants,
provide transportation to and from child care during the focus groups,
moderate the group discussions, and translate focus group transcripts.
Focus groups were conducted from January 2006 to March 2006.
Consistent with focus group data collection practices, our design involved
multiple groups with certain homogeneous characteristics. All focus groups
were conducted with mothers of children aged 5 or younger enrolled in
child care. These mothers also had limited English proficiency as
self-reported by potential participants during the focus group recruitment
process and were eligible for CCDF subsidies as determined by family's
income and parental work and education activities. The focus groups varied
by primary language spoken and whether or not participants' children were
receiving government child care subsidies. 36 Eight of the 12 focus groups
were conducted in Spanish and 4 in Vietnamese. We chose to conduct focus
groups in Spanish and Vietnamese because these two languages were among
the most prevalent languages, other than English, spoken in the states of
interest. According to 2000
36
During the focus group recruitment process, mothers were asked a series of
questions to ensure that it was likely that they were receiving CCDF
subsidies or eligible for them (depending upon whether they were selected
for the subsidized or unsubsidized groups), and to screen out participants
in other similar local programs such as state preschool programs or local
subsidy programs. However, states may use multiple sources to fund their
child care assistance programs, and participants may not know whether the
source of their assistance is federal or state funds. Therefore, it is
possible that some of the government subsidies received by participants
were not funded entirely or at all by CCDF even though the recipients met
the criteria for eligibility.
Appendix I: Scope and Methodology
Census data, Spanish was the language most commonly spoken among these
households in the states we visited. In Washington, Vietnamese was the
most commonly spoken language after Spanish, and in California, Vietnamese
was the second most commonly spoken language after Spanish. We did not
conduct focus groups in Vietnamese in North Carolina because of the
limited number of individuals who spoke languages other than English or
Spanish in the state. Six of the focus groups consisted of mothers with
young children (ages 0-5) who were enrolled in child care and received a
government subsidy for that care; the other 6 groups consisted of mothers
with young children (ages 0-5) who were enrolled in child care and did not
receive a government subsidy for that care, but whose children likely
qualified for subsidies based upon their family's income and employment or
education activities. Table 3 describes the characteristics of the group
at each location and lists locations and dates for each focus group
conducted. The number of participants in each focus group ranged from 6 to
13.
Table 3: Composition of Focus Groups
Language Location Date
Subsidized Spanish Yakima, Wash. January 31, 2006
Spanish Siler City, N.C. February 4, 2006
Spanish Siler City, N.C. March 2, 2006
Spanish San Jose, Calif. February 8, 2006
Vietnamese San Jose, Calif. February 11, 2006
Vietnamese Seattle, Wash. February 20, 2006
Unsubsidized Spanish Yakima, Wash. February 20, 2006
Spanish Pittsboro, N.C. February 4, 2006
Spanish Pittsboro, N.C. February 18, 2006
Spanish Oakland, Calif. February 7, 2006
Vietnamese San Jose, Calif. February 9, 2006
Vietnamese Seattle, Wash. February 20, 2006
Source: GAO analysis of focus group transcripts.
To help the moderator lead the discussions, GAO developed a guide that
included open-ended questions related to mothers' experiences finding
appropriate child care and attempting to access financial assistance to
help pay for the care. Discussions were held in a structured manner and
followed the moderator guide.
Focus groups involve structured small group discussions designed to gain
in-depth information about specific issues that cannot easily be obtained
from single or serial interviews. Methodologically, they are not designed
to provide results generalizable to a larger population or provide
statistically representative samples or reliable quantitative estimates.
They represent the responses only of the mothers who participated in our
12 groups. The population of individuals with limited English proficiency
in the United States consists of many cultural backgrounds and languages
in addition to Spanish and Vietnamese, and those and other factors may
influence the experience and attitudes of parents with limited English
proficiency regarding child care. Therefore, the experiences of other
mothers may be different from those of focus group participants. In
addition, while the composition of the groups was designed to include
different states, languages, and subsidy participation status, the groups
were not random samples of mothers with limited English proficiency.
Other Methodology
To assess HHS's efforts to ensure access to its programs for parents with
limited English proficiency, we interviewed HHS officials, reviewed
documents and guidance produced by HHS for state and local grantees, and
analyzed relevant legislation. We interviewed officials from CCB, the
Office of Head Start, HHS's Office for Civil Rights, and the five HHS
regional offices that covered the states that we visited. 37 We also
reviewed informational materials produced by HHS to facilitate access to
programs for individuals with limited English proficiency and online
resources pertaining to language access that were available through HHS's
and the Department of Justice's Web sites. Additionally, we analyzed
relevant legislation, federal regulations, and reports from research
organizations.
Finally, to obtain information pertaining to our research objectives, we
interviewed officials from various national organizations working on
issues related to early child care and education, as well as organizations
advocating on behalf of individuals with limited English proficiency.
37
The regional offices that we contacted were: Region IV (North Carolina);
Region V (Illinois); Region VI (Arkansas); Region IX (California); and
Region X (Washington).
Appendix II: Analyses of the Effects of Limited English Proficiency on Child
Care and Early Education Patterns
We analyzed national survey data collected in 1998 as part of the Early
Childhood Longitudinal Study, Kindergarten Class of 1998-99 (ECLS-K) from
parents of kindergarten children about their children's experiences in the
year before kindergarten. To conduct our analyses, we used logistic
regression models to estimate the "net effects" of the parent's limited
English proficiency on children's child care and early education patterns.
We defined parents as having limited proficiency in English if the parent
participating in the interview reported that a language other than English
was spoken at home, and if the respondent him or herself reported speaking
English either "not very well" or "not well at all." We made this decision
because we surmised that speaking is one of the main channels through
which information about child care is communicated. Additionally, we made
the decision to focus on the English language ability of the parent
participating in the interview on the assumption that the respondent
participating in the survey about his or her child would have a primary
role in child care decisions.
We considered the effect of the parent's limited English proficiency on
four outcomes. First, we looked at the effect it had on the likelihood of
their child receiving any type of nonparental child care in the year
before the child was in kindergarten, regardless of whether the care was
provided in a child care center (including a prekindergarten program) or
by relatives or nonrelatives in some other setting. Second, we looked at
the effect that limited English proficiency had on the likelihood of
receiving financial assistance from a social service or a welfare agency
to help pay for child care among those who did receive child care. Third,
we looked at the effect that limited English proficiency had on the
likelihood that the child care provided was in a center-based facility
(rather than care provided by relatives or nonrelatives) because it has
been suggested that children whose parents have limited English
proficiency may be less likely to receive center-based care than other
children. Fourth, and finally, we considered whether limited English
proficiency affected the likelihood of participating in Head Start.
By "net effects," we mean the effects of limited English proficiency that
operate after we control for other factors that affect these different
outcomes and that are related to limited English proficiency. The most
obvious among these other factors is race or ethnicity. That is, the
probability of using any nonparental care, receiving financial assistance
for child care, having center-based care rather than some other form of
care, and participating in Head Start are different among racial and
ethnic groups, and English proficiency is vastly different for some
groups, particularly Hispanics and Asians, than for whites and other
races. As
such, after looking at the difference between children of parents with
limited English proficiency and other children on these outcomes, we used
multivariate logistic regression models to re-estimate this difference
when controlling for the effect of characteristics such as the child's
race or ethnicity. The other characteristics we controlled included
household income (because of its effect on eligibility for some child care
assistance programs and Head Start) and parental education(because
previous studies have shown it to have an effect on participation in child
care and early education programs). We also controlled for the number of
persons over 18 in the household and whether the parent or parents in the
house were employed because these can affect the availability of care
givers in the home and determine the need for child care and child care
assistance outside of the home. Another reason why we controlled for
parental employment status is that it is one of the factors considered for
CCDF eligibility. When we looked at the likelihood of receiving any care
or receiving that care in a center-based facility, as well as at the
likelihood of receiving financial assistance for care received, we
controlled for whether the family participated in Head Start, since we
surmised this may affect whether additional child care was needed.
Additionally, because we thought that being in multiple types of child
care may affect the likelihood of one of them being provided in a
center-based facility or being subsidized by an outside source, we also
controlled for whether the child received multiple types of child care
when we looked at the likelihood of a child being in center-based care or
receiving financial assistance for child care. Finally, when we looked at
whether financial assistance was received for the care, we controlled for
whether the care was provided in a center-based facility on the assumption
that the cost of care may be higher when it is provided in a formal
center-based setting. Additionally, other factors, such as family
preferences for a certain type of care and parents' immigration status, as
well as changes in the CCDF program and child care policies within a
particular state of residence may affect child care and early education
patterns of children. We partially mitigated the potential effect of
preferences for certain types of care on the receipt of financial
assistance for child care by controlling for whether or not the child was
in center-based care. However, we could not include all factors that may
have had an effect on the outcomes in the analysis because the ECLS-K did
not collect the data to measure them.
An understanding of how to interpret the results of these multivariate
logistic regression models is facilitated by first considering tables 4
and 5, which estimate the effects of limited English proficiency, and race
or ethnicity, on the first two of these four outcomes. Tables 4 and 5
estimate how English proficiency and race or ethnicity are related to
receiving any
Appendix II: Analyses of the Effects of Limited English Proficiency on
Child Care and Early Education Patterns
nonparental child care and to receiving financial assistance for child
care (among those who received any nonparental child care). It is
important to note that these estimates are unadjusted for other
characteristics that are related to these outcomes, such as education,
income, and employment status. The top section of tables 4 and 5 shows the
effect of parents' limited English proficiency on the two outcomes, the
middle section shows the effect of the child's race or ethnicity, and the
bottom section shows the joint effect of the two, or the effect of limited
English proficiency within each racial or ethnic category. We show these
effects in each section of the tables by first providing percentages of
children of parents with limited English proficiency and other children
having a certain outcome. We then calculate odds and odds ratios for the
likelihood of children within each of the two groups having these
outcomes. Odds and odds ratios are the measures used to describe effects
that underlie the logistic regression models we later employ to estimate
net effects of limited English proficiency while controlling for other
factors.
Consider table 4, which provides percentages, odds, and odds ratios
related to the differences in receiving any type of child care across
children that differ by their parents' English proficiency, their race or
ethnicity, and both. We see in the top section that while approximately 75
percent of children whose parents are English proficient received some
form of child care in the year preceding kindergarten, the same is true of
only 46 percent of children whose parents have limited English
proficiency. These percentages are derived from weighted data in our
sample that take account of the fact that we are working with a sample
that is not a simple random sample (where all individuals have an equal
chance of being selected), but one in which children in some groups,
namely Asians and Pacific Islanders, were oversampled. They are based,
however, on the unweighted number of cases in our sample of 18,033
respondents (16,784 of them with parents proficient in English and 1,249
with parents with limited English proficiency), given in the third column
of the table. The difference in these two percentages is sizable, and
statistically significant, and would lead us to conclude that children of
parents with limited English proficiency are less likely to receive
nonparental care of any form.
Table 4: Differences in the Percentages and Odds of Receiving Any
Nonparental Care, by Parents' English Proficiency Status, Race or
Ethnicity, and Both, among Preschool-Aged Children
No Yes N Odds on Odds Parents' English proficiency status (%) (%)
(unweighted) yes: no ratio
English proficient 25.2 74.816,784 2.97
Limited English proficient 54.4 45.6 1,249 0.84 0.28*
Total 27.3 72.7 18,033
Race or ethnicity
White 21.878.2 10,262 3.59
Black 29.5 70.5 2,638 2.39 0.67*
Hispanic 39.7 60.33,205 1.52 0.42*
Asian 29.6 70.4 979 2.380.66*
Other 33.9 66.1 989 1.95 0.54*
Total 27.3 72.7 18,073
Race or ethnicity Parents' English proficiency status
White English proficient 21.7 78.310,204 3.62
Limited English proficient 38.1 61.9 34 1.62 0.45*
Black English proficient 29.2 70.82,611 2.43
Limited English proficient 60.839.2 10 0.65 0.27*
Hispanic English proficient 32.4 67.6 2,217 2.09
Limited English proficient 55.6 44.4 974 0.80 0.38*
Asian English proficient 25.4 74.6 753 2.94
Limited English proficient 46.0 54.0 222 1.17 0.40*
Other English proficient 33.9 66.1 976 1.95
Limited English proficient 33.7 66.38 1.97 1.01
Total 27.2 72.8 18,009
Source: GAO analysis of ECLS-K data.
* Denotes significance at the 95 percent level.
An alternative way to look at this difference is by calculating the odds
of receiving child care, which is the percentage of children who receive
child care divided by the percentage of children who do not. In the case
of children of parents that are English proficient, these odds are
74.8/25.2 = 2.97, which implies that in that group, approximately 3
families use child care for every family that does not (or that 300
families do for every 100 families that do not). In the case of children
of parents that are not English proficient, these odds are 45.6/54.4 =
0.84, which implies that for them, approximately 0.8 families use child
care for every family that does not (or that 80 families do for every 100
that do not). The ratio of
these two odds, or 0.84/2.97 = 0.28, tells us that the odds on receiving
any care are decidedly lower for children of parents with limited English
proficiency than for children of parents that are English proficient, by a
factor of 0.28.
The middle section of table 4 shows the differences in the percentages and
odds of children receiving child care across racial or ethnic categories.
The percentages of children receiving child care in the year before
kindergarten are lower for minority children than for whites, and these
differences are reflected in the odds as well. Among white children, about
3.6 children received child care for every child that did not, while among
blacks and Asians approximately 2.4 children received child care for every
child that did not. Among Hispanics, approximately 1.5 children received
child care for every child that did not. Where variables have more than
two categories, such as different categories of race and ethnicity, we
chose one category as the reference category and calculated odds ratios
that reflect how different each of the other categories is relative to
that one. In this case, whites were chosen as the reference category, and
the odds ratios of 0.67, 0.42, 0.66 and 0.54 indicate how much lower the
odds of receiving child care were for blacks, Hispanics, Asians, and other
races, respectively, than for whites.
The bottom section of table 4 shows the differences in the percentages of
children receiving any child care across the joint (or combined)
categories of parents' English proficiency and the child's race or
ethnicity. Here we have calculated odds for each of the joint categories,
and the odds ratios, which indicate how different the odds are across
English proficiency categories, within each category of race or ethnicity.
We can see that within most categories of race or ethnicity, children of
parents with limited English proficiency have lower odds of receiving any
child care than children of parents that are proficient in English, by
factors such as
0.38 for Hispanics and 0.40 for Asians. The odds ratios for whites,
blacks, and others were based on very small numbers of children of parents
with limited English proficiency. Of the 1,249 children of parents with
limited English proficiency, only 34, 10, and 8 children are white, black,
and other, respectively, and these numbers are too small for us to assess
whether and how much they differ from children of parents that are
proficient in English.
In sum, table 4 indicates that children of parents with limited English
proficiency were less likely to receive any child care than children of
parents proficient in English. Some of this is due to the fact that
children of parents with limited English proficiency tend to be Hispanic
and Asian,
Appendix II: Analyses of the Effects of Limited English Proficiency on
Child Care and Early Education Patterns
groups that are less likely than whites to receive child care. However,
not all of it is due to race or ethnicity differences, since among
Hispanics and Asians the children of parents with limited English
proficiency were less than half as likely as others within the same racial
or ethnic group to receive any child care.
Table 5 provides similar information with respect to the likelihood of
receiving financial assistance for child care, among those children that
received any care. Overall, children of parents with limited English
proficiency were less likely than those with parents proficient in English
to receive financial assistance (odds ratio = 0.60), though most racial or
ethnic minorities, except for Asians, were more likely than whites to
receive financial assistance when they received some type of care. That
is, while Hispanic children were twice as likely as white children to
receive financial assistance, and blacks and other races were
approximately four times as likely, Asians' odds of receiving financial
assistance were not statistically distinguishable from those of whites
(odds ratio = 0.70). Further, in the two groups-Hispanics and Asians-that
had sizable numbers of children of parents with limited English
proficiency, the effect of limited proficiency was different. Among
Hispanics, the odds of receiving financial assistance were lower for
children of parents with limited English proficiency than for children of
parents that were proficient in English (odds ratio = 0.46), while among
Asians the odds of receiving financial assistance were not statistically
distinguishable between children of parents with limited English
proficiency and children of parents that were proficient in English (odds
ratio = 1.95). Among the other groups, the numbers of children of parents
with limited English proficiency who received child care in the year prior
to kindergarten were too small for us to be able to reliably detect any
difference between them and others in the likelihood of receiving
financial assistance.
Table 5: Differences in the Percentages and Odds of Receiving Financial
Assistance for Child Care, among Those in Any Prekindergarten Care, by
Parents' English Proficiency Status, Race or Ethnicity, and Both
No Yes N Odds on Parents' English proficiency status (%) (%) (unweighted)
yes:no Odds ratio
English proficient 93.2 6.812,732 0.07
Limited English proficient 95.84.2 584 0.04 0.60**
Total 93.4 6.6 13,316
Race or ethnicity
White 95.9 4.1 8,173 0.04
Black 85.4 14.6 1,874 0.17 4.02*
Hispanic 92.5 7.5 1,976 0.081.90*
Asian 97.1 2.9 676 0.030.70
Other 85.1 14.9 637 0.17 4.11*
Total 93.4 6.6 13,336
Race or ethnicity Parents' English proficiency status
White English proficient 95.9 4.1 8,138 0.04
Limited English proficient 100.0 0.0 20 0.00 0.00
Black English proficient 85.4 14.6 1,861 0.17
Limited English proficient 73.1 26.9 4 0.37 2.15
Hispanic English proficient 91.5 8.5 1,533 0.09
Limited English proficient 95.9 4.1 435 0.04 0.46*
Asian English proficient 97.5 2.5 554 0.03
Limited English proficient 95.2 4.8120 0.05 1.95
Other English proficient 85.2 14.8629 0.17
Limited English proficient 100.0 0.0 5 0.00 0.00
Total 93.4 6.6 13,299
Source: GAO analysis of ECLS-K data.
* denotes significance at the 95 percent level.
** denotes significance at the 90 percent level.
The tables above showed the gross or unadjusted differences in receiving
child care and receiving financial assistance for child care between
children of parents with limited English proficiency and children of
parents proficient in English, and what those differences look like when
we control for or take account of race or ethnicity, the factor with which
parents' limited English proficiency is most closely associated. However,
limited English proficiency is associated with a number of other factors
that affect these two outcomes, as well as the other two outcomes that
were of interest to us, which were the likelihood of receiving
center-based
care (as opposed to care from relatives or nonrelatives in some other
setting) and the likelihood of participating in Head Start. Tables 6
through 9 show that the percentages of children that are Hispanic or
Asian, from lower-income families, have less educated parents, and have
three or more persons in the household over the age of 18 are higher among
children of parents with limited English proficiency than among other
children. Tables 10 and 11 show that the percentage of children that have
their parent (in single parent households) or both parents working and the
percentage of children that receive multiple types of care are lower among
children of parents with limited English proficiency than among other
children.
Table 6: Percentages of English Proficient Parents and Parents with Limited
English Proficiency, by Race or Ethnicity
Race or ethnicity
Total N White (%) Black (%) Hispanic (%) Asian (%) Other (%) (%)
(unweighted)
English proficient 61.5 16.814.2 2.5 5.0 100.0 16,784
Limited English proficient 2.6 1.1 87.7 8.30.4 100.0 1,249
Source: GAO analysis of ECLS-K data.
Table 7: Percentages of English Proficient Parents and Parents with Limited
English Proficiency, by Family Income
Income (percentage of the poverty level)
200 percent of N poverty level (%) poverty level (%) poverty level (%) Total (%)
(unweighted)
English Proficient 20.9 23.2 55.9 100.0 16,784
Limited English Proficient 55.4 32.0 12.6 100.0 1,249
Source: GAO analysis of ECLS-K data.
Table 8: Percentages of English Proficient Parents and Parents with
Limited English Proficiency, by Education
Highest education level of parent(s) in the household
High school Total N graduate (%) graduate (%) graduate (%) (%) (unweighted)
English 7.827.2 65.0 100.0 16,784 proficient
Limited English 47.4 29.9 22.7 100.0 1,249 proficient
Source: GAO analysis of ECLS-K data.
Table 9: Percentages of English Proficient Parents and Parents with
Limited English Proficiency, by the Number of Persons over the Age of 18
in the Household
Number of persons over 18
Total N 1 (%) 2 (%) 3+ (%) (%) (unweighted)
English proficient 16.1 72.1 11.8 100.0 16,782
Limited English 8.5 64.327.2 100.0 1,249 proficient
Source: GAO analysis of ECLS-K data.
Table 10: Percentages of English Proficient Parents and Parents with
Limited English Proficiency, by Parents' Work Status
Parents working
Not all All N working (%) working (%) Total (%) (unweighted)
English Proficient 33.4 66.6 100.0 16,550
Limited English 63.7 36.3 100.0 1,220 proficient
Source: GAO analysis of ECLS-K data.
Table 11: Percentages of English Proficient Parents and Parents with
Limited English Proficiency, by the Number of Different Types of Child
Care Used, among Those Using Care
Number of different types of child care used, among those using care
N One (%) Two or more (%) Total (%) (unweighted)
English proficient 72.327.7 100.0 12,732
Limited English 81.6 18.4 100.0 proficient
Source: GAO analysis of ECLS-K data.
Note: Each difference between families with limited English proficiency
and other families in tables 611, except for the category of high school
graduates, is significant at the 95 percent confidence level.
In tables 12 through 15 we show what the adjusted effect of parents'
limited English proficiency is on the likelihood of their child (1)
receiving any nonparental child care, (2) receiving financial assistance
for child care, (3) receiving center-based care, and (4) participating in
Head Start, when we estimate its effect using logistic regression models
to control for the effects of the other factors. In the first two columns
of each table, we show the unadjusted effect of parents' limited English
proficiency on each outcome across all racial/ethnic groups, and what the
adjusted effect looks like when we control for race or ethnicity and other
factors. In the third and fourth columns of each table, we show the
unadjusted and adjusted effect of parents' limited English proficiency for
Hispanics, and in the last two columns we show those same effects for
Asians. Separate analyses were done only for Hispanics and Asians because,
as table 6 shows, the percentage of children of other races whose parents
have limited English proficiency was very small. For the adjusted models,
we also show the effects of the other factors that we controlled for, such
as income and education, on the four outcomes. In the case of variables
that have multiple categories (such as race or ethnicity, income or
poverty status, education, and number of persons in the household over 18
years of age), the odds ratios indicate how much more or less likely the
categories of families indicated are to have each outcome than the
reference (or omitted) category. The reference category for race or
ethnicity is white, the reference category for poverty status is less than
100 percent of the federal poverty level, the reference category for
education is less than high school graduate, and the reference category
for the number of persons in the household over 18 is one.
Likelihood of receiving any nonparental care. Table 12 shows that before
adjusting for other factors, the effect of parents' limited English
proficiency on the likelihood of receiving any type of nonparental
childcare was negative and significant for all groups considered together,
and for Hispanics and Asians considered separately (odds ratios of 0.28,
0.38, and 0.40, respectively). After controlling for these other factors,
the differences between children of parents with limited English
proficiency and other parents in terms of their receipt of any type of
child care were smaller for all groups considered together and for
Hispanics (odds ratios of 0.77 and 0.75, respectively), but not
statistically significant among Asians (odds ratio of 0.85). While almost
all of the control variables attain statistical significance in the model
that included all racial and ethnic groups, the statistical significance
of individual control variables in the models including only Asian or
Hispanic children varies.
Table 12: Odds Ratios from Logistic Regression Models Used to Estimate the
Effects of Different Factors on the Likelihood of Receiving Any Child
Care, after Adjusting for Other Characteristics
All ps Hispanics Asians
grouUnadjusted Adjusted Unadjusted Adjusted Unadjusted Adjusted
Limited
English 0.28* 0.77* 0.38* 0.75* 0.40* 0.85
proficient
White Ref
Black 1.19**
Hispanic 0.87*
Asian 0.81*
Other 0.86
< 100 percent poverty Ref Ref Ref
100-200 percent poverty 1.09 1.15* 0.66*
> 200 percent poverty 2.19* 1.96* 1.88*
< High school graduate Ref Ref Ref
High school graduate 1.44* 1.53* 1.48
> High school graduate 2.26* 2.16* 1.79
1 Person over 18 Ref Ref Ref
2 Persons over 18 0.65* 0.72* 0.83
More than 2 persons over 18 0.77* 0.89* 0.81
All parent(s) work 3.16* 3.17* 3.24*
Head Start 0.38* 0.48* 0.22*
Source: GAO analysis of ECLS-K data.
* denotes significance at the 95 percent level.
** denotes significance at the 90 percent level. Note: "Ref" refers to
reference categories.
Page 61 GAO-06-807 Child Care and Early Childhood Education Appendix II:
Analyses of the Effects of Limited English Proficiency on Child Care and
Early Education Patterns
Likelihood of receiving financial assistance for child care. Table 13
shows that before adjusting for other characteristics, the odds ratios
estimating the effect of parents' limited English proficiency on the
likelihood of receiving financial assistance for child care were 0.60,
0.46, and 1.95 for all groups together, Hispanics, and Asians, although
the result for Asians was not statistically significant. While other
factors were significantly related to the likelihood of receiving
financial assistance for child care, controlling for their effects did not
markedly diminish the estimated difference between children of parents
with limited English proficiency and other children overall, or for
Hispanics or Asians. After other factors are taken into account, children
of parents with limited English proficiency were about half as likely as
others to receive financial assistance overall and among Hispanics (odds
ratios of 0.41 and 0.44, respectively), but among Asians the difference
was not statistically significant (odds ratio = 1.85).
Table 13: Odds Ratios from Logistic Regression Models Used to Estimate the
Effects of Different Factors on the Likelihood of Receiving Financial
Assistance for Child Care, among Those in Any Care, after Adjusting for
Other Characteristics
All groups Hispanics Asians Unadjusted Adjusted Unadjusted Adjusted
Unadjusted Adjusted
Limited English proficient 0.60** 0.41* 0.46* 0.44* 1.95 1.85
White Ref
Black 1.64*
Hispanic 1.36*
Asian 0.75
Other 2.73*
<100 percent poverty Ref Ref Ref
100-200 percent poverty 0.68* 0.64* 8.40*
>200 percent poverty 0.18* 0.25* 2.83
< High school graduate Ref Ref Ref
High school graduate 1.12 1.26 0.21
> High school graduate 0.92 1.23 0.37**
1 person over 18 Ref Ref Ref
2 persons over 18 0.42* 0.27* 0.09*
More than 2 persons over 18 0.51* 0.21* 0.09*
All parent(s) work 1.31* 1.13 1.66
Head Start 1.83* 1.95* 3.34
Center-based care 2.16* 1.63 1.64
Multiple types of care 1.46* 1.65* 2.20**
Source: GAO analysis of ECLS-K data.
* denotes significance at the 95 percent level.
** denotes significance at the 90 percent level. Note: "Ref" refers to
reference categories.
Likelihood of receiving center-based care. Table 14 shows that before
adjusting for other factors, the effect of parents' limited English
proficiency on the likelihood of receiving center-based child care among
those who received any type of child care was significant when all
racial/ethnic groups were considered together (odds ratio = 0.44), and
significant for Hispanics (odds ratio = 0.73) but not for Asians (odds
ratio = 0.92). None of the differences between children of parents with
limited English proficiency and other children were statistically
significant, however, after we controlled for other factors.
Table 14: Odds Ratios from Logistic Regression Models Used to Estimate the
Effects of Different Factors on the Likelihood of Receiving Center-Based
Care, among Those in Any Care, after Adjusting for Other Characteristics
All groups Hispanics Asians Unadjusted Adjusted Unadjusted Adjusted
Unadjusted Adjusted
Limited English proficient 0.44* 0.98 0.73* 0.83 0.92 1.10
White Ref
Black 0.98
Hispanic 0.61*
Asian 0.77*
Other 0.53*
<100 percent poverty Ref Ref Ref
100-200 percent poverty 0.98 1.05 0.89
>200 percent poverty 1.72* 1.37* 1.46
< High school graduate Ref Ref Ref
High school graduate 1.36* 1.17 0.59
> High school graduate 2.45* 1.94* 1.15
1 person over 18 Ref Ref Ref
2 persons over 18 1.06 0.95 0.76
More than 2 persons over 18 0.64* 0.78 0.23*
All parent(s) work 0.39* 0.33* 0.51*
Head Start 0.18* 0.26* 0.17*
Multiple types of care 9.81* 8.50* 17.66*
Source: GAO analysis of ECLS-K data.
* denotes significance at the 95 percent level.
** denotes significance at the 90 percent level. Note: "Ref" refers to
reference categories.
Likelihood of participating in Head Start. Table 15 shows that before
adjusting for other factors, children of parents with limited English
proficiency had higher odds of participating in Head Start when all
ethnic/racial groups were considered together (odds ratio = 1.39). The
same was true when Asians were considered separately (odds ratio = 3.81),
but no significant effect of parents' limited English proficiency was
found for Hispanics (odds ratio = 0.98). After controlling for other
characteristics, children of parents with limited English proficiency had
significantly lower odds of participating in Head Start when all
racial/ethnic groups were considered together (odds ratio = 0.67), and
when Hispanics were considered separately (odds ratio = 0.69), but
significantly higher odds among Asians (odds ratio = 1.90).
Appendix II: Analyses of the Effects of Limited English Proficiency on
Child Care and Early Education Patterns
Table 15: Odds Ratios from Logistic Regression Models Used to Estimate the
Effects of Different Factors on the Likelihood of Participating in Head
Start, after Adjusting for Other Characteristics
All s Hispanics Asians
groupUnadjusted Adjusted Unadjusted Adjusted Unadjusted Adjusted
Limited
English 1.39* 0.67* 0.98 0.69* 3.81* 1.90*
proficient
White Ref
Black 3.21*
Hispanic 1.75*
Asian 1.57*
Other 2.69*
<100 percent poverty Ref Ref Ref
100-200 percent poverty 0.66* 0.72* 0.69**
>200 percent poverty 0.18* 0.22* 0.18*
< High school graduate Ref Ref Ref
High school graduate 1.19* 1.43* 0.53**
> High school graduate 0.71* 0.90 0.46*
1 person over 18 Ref Ref Ref
2 persons over 18 0.82* 0.87 0.94
More than 2 persons over 18 0.74* 0.91 0.60
All parent(s) work 0.97 1.09 0.69**
Source: GAO analysis of ECLS-K data.
* denotes significance at the 95 percent level.
** denotes significance at the 90 percent level. Note: "Ref" refers to
reference categories.
Appendix III: Comments from the U.S. Department of Health and Human Services
Appendix IV: GAO Contact and Staff Acknowledgments
GAO Contacts
Marnie S. Shaul, (202) 512-7215, [email protected]
Cornelia M. Ashby, (202) 512-8403, [email protected]
Staff Acknowledgments
Betty Ward-Zukerman (Assistant Director) and Natalya Barden
(Analyst-in-Charge) managed all aspects of the assignment. Laurie Latuda,
Janet Mascia, Jonathan McMurray, and Ethan Wozniak made key contributions
to multiple aspects of the assignment. Alison Martin, Grant Mallie, Amanda
Miller, Anna Maria Ortiz, James Rebbe, and Douglas Sloane provided key
technical assistance.
(130511)
GAO's Mission
The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.
Obtaining Copies of GAO Reports and Testimony
The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site (www.gao.gov). Each weekday, GAO posts GAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To
have GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."
Order by Mail or Phone
The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:
U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548
To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061
To Report Fraud, Waste, and Abuse in Federal Programs
Contact:
Web site: www.gao.gov/fraudnet/fraudnet.htm
E-mail: [email protected]
Federal Programs Automated answering system: (800) 424-5454 or (202)
512-7470
Congressional Relations
Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Relations
Washington, D.C. 20548
Public Affairs
Paul Anderson, Managing Director, [email protected] (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548
PRINTED ON
RECYCLED PAPER
*** End of document. ***