POLICY VARIATIONS AMONG STATES
AND IMPACT ON THE CHOICE OF CHILD CARE SERVICE
By
Woo Jong Kim

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Social Work - Doctor of Philosophy
2014

ABSTRACT
POLICY VARIATIONS AMONG STATES
AND IMPACT ON THE CHOICE OF CHILD CARE SERVICE
By
Woo Jong Kim
High quality child care services have positive effects on children’s developmental and
long term educational outcomes, as well as promote their parents’ stable employment. Center
based care, which involves taking care of children in a formal or established child care center,
has been found to have a higher quality compared to other types of care service, such as family
home care and kinship care. States have used child care subsidy policies and quality regulations
to support the choice of center based care among low income working families, but states’ child
care policy choices and the influences on the choice of different types of care have not been
examined.
This study conducted two different analyses to examine the variations in states’ child
care policy decisions and the influences of child care subsidy and quality regulation rules on
working families’ choice of care arrangement among various options of child care services.
Cluster analyses showed significant variations in each policy rules across states. States
strategically formed child care policy in consideration with each state’s characteristics.
Multilevel analysis revealed that the choice of center based care was significantly associated
with higher child-staff ratios, higher copayment rates, and higher levels of tiered reimbursement
rates. When investigating policy influences on specific types of care compared to counterpart by
choosing specific populations, the choice of center based care rather than family home care had
significant associations with a higher child-staff ratio in center based care and less training
requirements for family home care among families using formal care. The choice of family home

care versus informal care such as grandparents or relative care was significantly associated with
higher family home care capacity and more training requirement for family home care. This
study reveals that both child care subsidy policy and regulation significantly vary across states
and have substantial roles in the choice of care service among working families. Given budget
limitations and widening state authority, these findings support social workers in the policy area
to effectively redesign child care policy to increase the use of better quality of service.

TABLE OF CONTENTS

LIST OF TABLES.................................................................................................................................................... vi
LIST OF FIGURES .................................................................................................................................................vii
I.

INTRODUCTION ........................................................................................................................................ 1

II.

BACKGROUND AND CONTEXT ............................................................................................................ 3
Child care service: variations in types of service and quality ................................................. 3
Choice of child care service: theory and factors affecting choice of services .................... 4
Child care policy in the US..................................................................................................................... 7
Background on policy changes ............................................................................................................ 8
Policy variations across states ............................................................................................................ 9
Child care subsidy rules ............................................................................................................ 9
Quality regulation ..................................................................................................................... 10
How the policy operates: theory and practice ........................................................................... 11
Previous research on state variations ........................................................................................... 15
Previous research on impact of policy on choice of care arrangement across states 17
Purposes of current study .................................................................................................................. 21
Research aims ......................................................................................................................................... 21

III.

METHODS ................................................................................................................................................. 23
Data and sample..................................................................................................................................... 23
Measures ............................................................................................................................................. 24
State policy and characteristics ........................................................................................... 24
Variables on children’s characteristics............................................................................. 29
Research design ..................................................................................................................................... 32
Analysis ..................................................................................................................................................... 33

IV.

RESULTS.................................................................................................................................................... 39
Research aim 1 ....................................................................................................................................... 39
Child care subsidy ..................................................................................................................... 39
Quality regulation ..................................................................................................................... 41
Research aim 2-A ................................................................................................................................... 44
Research aim 2-B ................................................................................................................................... 48

V.

DISCUSSION ............................................................................................................................................. 51
States’ variation among clusters...................................................................................................... 51
Different policy impacts on individuals versus overall populations ................................. 53
Significance of quality regulations on the choice of services ............................................... 54
Limitations ............................................................................................................................................... 56

VI.

CONCLUSION AND IMPLICATION ................................................................................................... 58
iv

REFERENCES ....................................................................................................................................................... 60

v

LIST OF TABLES

Table 1.

Means, Standard Deviations, and ranges of state policy rules and state
characteristics (n=50) ........................................................................................................... 25

Table 2.

Sample demographics ............................................................................................................ 30

Table 3.

Cross tabulation of child care subsidy and quality regulation cluster
memberships ............................................................................................................................ 43

Table 4.

Bivariate correlation coefficients among sample demographics.......................... 45

Table 5.

Result for multilevel analysis: choice of center vs. other types (n=2742) ........ 46

Table 6.

Results of multilevel analysis: policy impacts on choice of specific care .......... 49

vi

LIST OF FIGURES

Figure 1.

Research framework ............................................................................................................ 32

Figure 2.

Differences in child care subsidy policy across clusters ........................................ 39

Figure 3.

Differences in quality regulation policy rules across clusters ............................. 41

vii

I.

INTRODUCTION

High quality child care services have positive effects on children’s developmental and
longer term educational outcomes and parents’ stable employment (Burstein & Layzer, 2007;
Chaudry, 2004; Albers, Riksen-Walraven & de Weerth, 2010; Campbell, Pungello, MillerJohnson, Burchinal, & Ramey, 2001; Gormley, Gayer, Phillips, & Dawson, 2005; Vandell,
Belsky, Burchinal, Steinberg & Vandergrift, 2010). The quality of child care services is
substantially different depending on which types of services families use. Center based care,
which involves taking care of children in a formal or established child care center, has been
found to have higher quality compared to other types of care service, such as home based care
and kinship care (Gormley Gayer, Phillips, & Dawson, 2005; Loeb, Bridges, Bassok, Fuller, &
Rumberger, 2007; Magnuson, Ruhm, & Waldfogel, 2007).
The choice of care arrangement is affected by several different factors including family
characteristics, child care market, and policy interventions. However, low-income children are
generally less likely to participate in center-based settings than more affluent children because of
price and limited high quality care providers near their residences (Chaudry, 2004; Fuller, Kagan,
Caspary, & Gauthier, 2002; Layzer & Goodson, 2006; Lowe & Weisner, 2004; Meyers & Jordan,
2006). Under these circumstances, policy interventions, such as state child care subsidies and
government child care quality regulations, can play important roles in supporting low income
parents’ choices of better quality of care.
States have key roles in policy making for these child care policies. With the enactment
of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, states were
given increased flexibility in determining child care policy including funding schemes and state
1

policy rules. Increased state autonomy led to substantial variations in policy rules for child care
subsidies, such as eligibility limits, copayment rates, and reimbursement rates across states
(Schulman & Blank, 2011). States also have different regulations for the types of care allowable
for subsidies, as well as different rules and requirements for licensing (National Association of
Child Care Resource & Referral Agencies [NACCRRA], 2011). Thus, combinations of subsidy
policy and regulations may create unexpected policy impacts not only on targeted population but
also on overall child care market.
Despite states having critical roles in forming child care policy, which is a key policy
scheme in supporting low income families as well as helping families to make a transition into
the workforce, very little is known about how policy choices and subsequent practices may
differentially affect the choice of service among families. Prior research has paid little attention
to inter-state policy variation and the potential differences in outcomes for children and families.
Families residing in different states have different financial situations even if they equally use
center based care (Schulman & Blank, 2011). Similarly, depending on the strictness of quality
regulation, families may have wide variations in service quality and the availability of providers
across states. Therefore, understanding the complexity of policy choices and how states
implement child care policy has become both increasingly difficult and important.
This study will explore child care policy variations across states and investigate how
various policy rules shape child care environments, and as a consequence, the choice of care
arrangement among working families with young children.

2

II.

BACKGROUND AND CONTEXT

Child care service: variations in types of service and quality
There are a variety of options available to families who seek a non-parental child care
arrangement for their children. Child care services in the US are provided by various providers in
diverse settings, such as center based care, family home care, babysitters, and relative care.
Regulations for care differ based on the type and location of care. Based on where children are
cared, they can be classified into two categories: center based arrangements, which occur in
organized facilities, and home based arrangements, which occur in the child's or provider's home,
including relatives, babysitters, and family care providers. These two categories do not only
differ in terms of setting and providers’ characteristics, but also in the degree of regulation,
standards of care, quality of care, and cost of service. Child care centers are subject to heavier
regulations for licensing than are home based care facilities in terms of levels of child-staff ratios,
certain levels of education attained, and training requirement for center staff (National
Association of Child Care Resource & Referral Agencies [NACCRRA], 2011). Also, continuous
inspections are followed in center based facilities after licensing. Home based care providers are
not consistently subject to licensing procedures, but there are variations within them. Family
home providers who care for a group of unrelated children in their own homes are subject to
licensing unlike other types of home based care. Even though family home providers may be
exempted if providers care for a sufficiently small number of children (NACCRRA, 2011), they
are objects of state’s certification in many states. In this respect, center based care and family
home care can be categorized as formal care while other types of home based care including
relatives and babysitters can be categorized as informal care.
3

A significant body of research has demonstrated the differences of quality within services,
and shown an especially higher level of quality for center based care. Quality differences among
child care types have been found to be particularly stark in low income communities (Fuller,
Kagan, Loeb, & Chang, 2004; Li-Grining & Coley, 2006). Participating in center-based child
care improves children's later school outcomes and the effects tend to be particularly large for
children from low-income families (Gormley, Gayer, Phillips, & Dawson, 2005; Loeb, Bridges,
Bassok, Fuller, & Rumberger, 2007; Magnuson, Ruhm, & Waldfogel, 2007). Thus, it is even
more important to understand factors that influence higher quality childcare options for families.
Choice of child care service: theory and factors affecting choice of services
Basic labor supply theory suggests that the quality and cost of alternative arrangements
for non-parental care may affect the values of mother’s time at home and in employment (Blau,
Ferber, & Winkler, 2002). This theory assumes that working mothers seek to maximize
satisfaction with regard to allocation of time between direct care giving and other labor and with
regard to the quality of care received by their children (Meyers & Jordan, 2006). According to
this theory, if working mothers have enough purchasing power, they would buy the better quality
of childcare service. Empirical research found that the effect of price on the choice of care
arrangements is sensitive especially to the choice of center based care. Demand for formal child
care, such as center based care or licensed care providers, is more sensitive to price than is
demand for informal care, such as babysitter and relatives, and a reduction in price leads to a
substitution of formal care for informal arrangements (Michalopoulos & Robins, 2002; Powell,
2002).
However, in reality, the choice of type of care in which children end up is a result of the
dynamic and inter-related family characteristics, market function, and policy forces that
4

influence selection (Blau, 2003). Previous studies have demonstrated factors affecting parents’
child care choices in terms of family characteristics, community context, and policy effects. The
need for non-parental care is closely related to mother’ employment and associated with factors
related to mothers’ work status, such as working schedules and working hours (Connelly &
Kimmel, 2003; Henly, Ananat, & Danziger, 2006; Kimmel & Powell, 2006). Empirical studies
have also identified other family characteristics that were strongly associated with choices of
care arrangement. The presence of a spouse/partner or other adult relative in the household
reduces the likelihood that a mother will use market and formal care arrangements during hours
of work, presumably by providing alternative sources of care (Huston, Chang & Gennetian,
2002). Children with more family members tend to receive less formal care than those in smaller
families; when children from larger families are in non-parental care, it is more likely to be
informal care (Huston, Chang, & Gennetian, 2002; Liang, Fuller, & Singer, 2000).
Characteristics of the child also have significant impacts on the choice of care
arrangements. The age of a child has shown strong association with the type of child care parents
choose. When a child is either younger (infant or toddler) or at school age, informal and homebased care arrangements are more often used, while formal center based care is used more often
during the preschool years starting around age three (Burstein and Layzer, 2007; Early &
Burchinal, 2001; Huston, Chang & Gennetian, 2002; Tout, Zaslow, Papillo, & Vandivere, 2001).
In addition, gender of a child has shown significant association with the types of child care
parents choose. Girls are significantly more likely to be in regular, non-relative care than boys
are (Hiedemann, Joesch, & Rose, 2004).
Child care arrangements also vary with parents’ socio-demographic factors including
education, income level, and race/ethnicity. Studies found that these socio-demographic factors

5

affected parental preferences and beliefs. Children of mothers with high levels of education are
more likely to be in center care than in family day care or relative care (Early & Burchinal, 2001).
Low income, racial minority and immigrant parents were particularly likely to use relative
caregivers (Brown-Lyons, Robertson, & Layzer, 2001; Lowe & Weisner, 2004; Meyers, Heintze,
& Wolf, 2002). National data have consistently shown that Hispanic families use proportionately
more relative care and more informal care for their children than either White or African
American families. For African American families, center care use was most prevalent and did
not vary by family income (Capizzano, Adams, & Jason, 2006; Liang, Fuller, & Singer, 2000;
Shlay, 2010). Even though there are variations in the choice of child care by race, it may be hard
to conclude that the high use of informal care among Hispanic families and high use of organized
care among African American families indicate preferences for these types of child care
arrangements. Research argues that differences in child care use by race and ethnicity may not be
due to different preferences for particular types of care, but for bundles of care characteristics
people may believe are associated with particular care situations, such as market constraints and
lack of information (Shlay, 2010).
Even though individual factors and family characteristics shape the choice of child care
service, this decision is limited by the availability of various options within their neighborhoods
(Adams, Rohacek, & Snyder 2008; Bromer and Henly 2009). The influence of availability is
especially large for low income families (Fuller, Kagan, Caspray, & Gauthier, 2002; Tang, Coley,
& Votruba-Drzal, 2012). Low income communities tend to have fewer regulated child care
providers than higher income communities; many children living in low income communities
have limited service options, and it can thus shape child care preferences and use (Chaudry, 2004;
Layzer & Goodson, 2006; Meyers & Jordan, 2006). In addition, limited financial resources make

6

low income working families less likely to choose center based care (Chaudry, 2004; Fuller,
Kagan, Caspary, & Gauthier, 2002; Lowe & Weisner, 2004). As a result, low income children
are less likely to participate in center-based settings than more affluent children (Capizzano &
Adams, 2003). In this situation, policy intervention can have important roles in supporting the
use of high quality services for working families.
Child care policy in the US
The Child Care and Development Fund (CCDF) is the primary federal program
specifically devoted to providing families with child care subsidy and supports to improve
quality in current US child care system (Administration for Children & Families, 2012). As the
federal government’s largest child care program, it serves about 1.6 million children per month
on average by spending a total of $8.8 billion in child care support for working families in 2011
(United States Department of Health and Human Services [US-DHHS], 2012a; US-DHHS,
2013). To pursue two policy goals- facilitating employment and promoting children’s health
development, CCDF is allocated into two main policy schemes: providing child care subsidies
and improving overall quality of care. Even though the CCDF has two different goals, the
current child care system actually focuses on increasing employment rather than children’s
development by emphasizing child care subsidy provision more than service quality
improvement. Most of CCDF allocation is spent on child care subsidies and only small portions
are allotted to quality improvement. There is a federal requirement on child care expenditures for
quality activities, but the minimum requirement is only 4% of CCDF (Administration for
Children & Families, 2012).

7

Background on policy changes
The emphasis on the child care subsidy is partly because of development history of
CCDF. The increase of child care funding occurred with the welfare reform in 1996. The
Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA)
fundamentally changed federal child care assistance programs for low- income families. The
legislation eliminated federal child care entitlements and consolidated several different major
sources of federal child care subsidies for low- income children into a single block grant to states,
the Child Care and Development Fund (Long, Kirby, Kurka, & Boots, 1998). This change made
two significant differences in the child care assistance from the past. First, the overall funding for
child care assistance dramatically increased after the reform due to the mandatory work related
activities for welfare recipients. It became increasingly important to provide childcare subsidies
so that families could access to child care services that were previously unaffordable and given
that the lack of services was a barrier to moving into jobs.
Second, with the changes in the funding system from matching funds to a single block
grant for each state, the Child Care Development Fund devolved authority to state governments
(Rigby, Kagan, & Brooks-Gunn, 2004). This change allowed states to decide how much they
would spend of public expenditures on child care assistance, depending on the level of need
which is related to child poverty, available fiscal resources, and political climate (Douglas &
Flores, 1998; Levy & Michel, 2002). Depending on the states’ decision, the expenditures for
child care assistance vary across states. In addition, by allowing states to transfer the TANF
block grant to CCDF and utilize it for child care services (US-DHHS, 2012b), the variations
became substantial. In 2011, eight states spent more than 30 percent of their TANF funds on
child care, including direct spending and transferred funds, while 21 states spent less than 10

8

percent of TANF funds on child care, with ten spending less than five percent (Schott, Pavetti, &
Finch, 2012).
These two changes comprised one of the critical policy schemes in welfare system given
that the child care services were necessary in welfare reform to transition low income families
from welfare to work. In addition, those changes in funding systems led to changes in governing
local child care systems by enabling states to decide the policy rules based on their fiscal
capability and child care situations. Less federal oversight and decentralized funding schemes led
to variations across states in child care environments and families’ choices were subject to the
resulting changes in local childcare.
Policy variations across states
State level decision making became the key determinant of how much child care support
would be available and how it would be distributed. The block grant program gave states greater
authority in designing their child care subsidy programs to better meet the states' child care needs
and objectives. States can determine child care policy rules including child care subsidy policy
rules quality regulation rules.
Child care subsidy rules
The current child care subsidy system provides child care services through certificates,
vouchers, or grants and contracts with providers, to low income working families
(Administration for Children & Families, 2012). Parents can select a child care provider among
providers in the market, who satisfy any applicable state and local requirements, including basic
health and safety requirements. State can adjust the amount and quality of child care available,
and how and to whom it is distributed using policy levers, including income eligibility limits,
family copayment, and reimbursement rates. Those policy levers can affect parental choice and
9

access to quality child care. Federal rules describe that families’ earning incomes up to 85% of
the state medium income (SMI) are eligible for child care subsidies, but states can lower the
income limits to reduce the number of families to be served. Families are supposed to contribute
to the cost of child care services as the family copayment is based on the income and family size,
but this requirement may be waived for those at or below the federal poverty level based on
states’ decision. While some states waive the copayment to families under the poverty level or
under special circumstances, such as a single parent family, other states impose the same
copayment to those families. In addition, states can set the reimbursement rates which are the
payment rates to providers for child care services by conducting a regular market rate survey on
child care services. Since states are suggested to set the reimbursement rate equal to 75th
percentile of market rates, this is not mandatory and thus, reimbursement rates vary across states.
Additionally, some states have the tiered reimbursement system which applies different
reimbursement rates among providers depending on their assessed quality of services for the
purpose of increasing the number of quality providers (Schulman and Blank, 2011).
Quality regulation
Increased state authority for funding and choice of policy rules also affects regulation
policy. CCDF improves the quality of care and helps programs meet high standards that is linked
to children’s healthy development and learning by supporting child care licensing and regulation,
and quality improvements systems. CCDF also provides support for child care workers to attain
more training and education. Based on those purposes, states are required to use at least four
percent of CCDF on quality activities, such as consumer and parent education and increase of
availability in high quality service (Administration for Children & Families, 2012).

10

Child care is regulated by the states in order to reduce the risk of harm to children from
exposure to potential harms, such as communicable illnesses, infectious diseases, and injury
(Alkon, To, Mackie, Wolff, & Bernzweig, 2010). Most states regulate structure and program
features such as child to staff ratio in the classroom, the education and training of the staff and
director, building and equipment safety, sanitation, fire safety, staff background checks,
discipline practices, food safety, and immunization (NACCRRA, 2011). Those regulations
dramatically vary in terms of the degree of stringency and types of providers (NACCRRA, 2011).
While some states regulate only specific types of care, usually center based care, other states
regulate not only center based care but also family home care. Even though family home care is
regulated, the degree of stringency and comprehensiveness in regulatory factors for family home
care, such as teacher’s education and health and safety, are less than those for center based care
(NACCRRA, 2011).
How the policy operates: theory and practice
Subsidies and regulation are the primary tools available to governments to impact a
mixed-delivery, market-based, service industry such as child care (Rigby, Ryan, & Brooks-Gunn,
2007). By increasing purchasing power and raising quality of service and its availability,
subsidies and regulation may shape more favorable child care environments, which offer high
quality and various options of child care services for low income families. However, capturing
the impact of policy on actual choices of child care arrangement is not simple. Increased state
autonomy in child care policy makes it more difficult to examine policy effects on choices of
care arrangements. Substantial variations in child care policy across states limit generalizability
of findings made in previous studies on impacts of any particular child care policy. In addition,
even though subsidies and regulation have direct policy purposes, they often operate in multi11

dimensional ways. As Morris (1999) argued that only a well-designed mix of subsidies and
regulation policy has the potential to improve access to the quality child care, measuring policy
impacts require careful examinations.
Even though child care subsidies are designed to help low income families be self
sufficient and to support the use of quality child care service, traditional economic theory
suggests that subsidies have not only direct effects on the use of quality services but also indirect
effects on the child care market (Rigby, Ryan, & Brooks-Gunn, 2007). If parents have greater
purchasing power, the demand for child care should intensify, and this would increase the supply
of providers, competition among providers, and eventually promote higher quality of care across
child care settings (Gennetian, Crosby, Huston, & Lowe, 2004; Wolfe, 2002). More specifically,
a supply of providers would grow naturally in a particular geographic area if families demanded
a particular type of arrangement (Fuller & Strath, 2001). In fact, states utilize policy rules to
adjust demand and supply of child care services. States may reduce or increase the population
they serve by raising or lowering income eligibility thresholds. Copayments can be also used to
adjust the number of families served by the child care subsidy system. Higher levels of
copayments may enable states to have more funding available and widen the distribution of
services over the eligible population (Levy & Michel, 2002). However, if the copayment rates
are too high, eligible families will be unable to afford it and be more likely to switch to lower
quality of service or unpaid sources of child care. Previous research has argued that low
copayment rates and high value of subsidies were associated with receiving child care subsidies,
which have been found to significant impacts on the use of center based care (Grobe, Weber,
Davis, 2008; Long, Kirby, Kurka, & Waters, 1998). Reimbursement rates affect the supply and
accessibility to needy families to subsidized childcare, and quality of services by making more

12

center care available for children (Levy & Michel, 2002). Tiered reimbursement system may also
have similar impacts on the market. Larger gaps between highest and lowest reimbursement rates
among providers may induce the market entry of service providers and encourage service
providers to increase their service quality. Overall, if states provide child care subsidies with
higher values (higher reimbursement rates) to more people (high eligibility rates) with low levels
of copayment, their use of center based care would increase a supply of quality services as a
result.
However, this theoretical argument may not be true in the child care market. Most parents
use child care that is nearby and easily accessible. As a result, the market for child care is not
uniform across a country. Instead, it is composed of many smaller local markets each serving an
area (Blau & Mocan, 2002). Critics also suggest that subsidy programs encourage families to
increase their child care consumption, but do not provide incentives for the use of high-quality
arrangements (Blau & Currie, 2006). Parents would purchase low-quality care since they cannot
distinguish higher quality care because of the lack of information. In this case, subsidies may
neither support the choice of high quality care nor the increase of service quality.
Although subsidies may not directly increase the choice of high quality service in some
cases, quality regulation may deal with this problem. A key rationale for child care regulation is
to reduce information asymmetries between child care providers and consumers (Blau, 2007).
Because parents lack reliable information about the quality of care offered by different providers,
they must rely upon government regulators to specify and to enforce a minimum quality level.
Previous research revealed that regulation does have positive impacts on increasing overall
quality of service (Helburn, 1995; Vandell, 2004). More training hours were found to increase
the quality of family home care (Raikes, Raikes, & Wilcox, 2005) and a stringent (low) child-

13

staff ratio has been found to be more beneficial for children (Vandell & Wolfe, 2000). Training
requirements and child-staff ratios had significant effects on quality of service and choice of
service type among families (Rigby, Ryan, & Brooks-Gunn, 2007).
However, the degree of stringency may have unexpected impacts on the availability of
quality service. By increasing operating costs, more stringent regulation in center based care may
have unintended consequences, such as increasing service price and reducing supply of quality
care, as well as increasing choice of alternative service with lower quality, such as family day
care and parental care (Blau, 2007; Gormley, 1999; Hotz & Xiao, 2011). These consequences are
more prevalent in low income communities. Hotz and Xiao (2011) found that the enforcement of
regulations on child care centers reduced the number of these types of establishments in lower
income communities. Conversely, similar regulations in higher income areas increased the
quality of services provided (Hotz & Xiao, 2011).
Taken together, if subsidies are generous enough for choice of high quality services and
stringent regulation that is likely to improve overall child care quality, the combination of two
policy levers may increase overall use of quality child care service. However, in reality, the
impact of policy interactions on the choice of child care service may not operate as shown in
theory. For example, beneficiaries might be willing to access higher quality service providers
with more generous subsidies, while at the same time providers may be reluctant to enter the
child care service market because of higher operating cost caused by stringent regulation. Also,
even though there are many quality providers in the market, higher copayment rates and lower
reimbursement rates, which are percentages of market prices the subsidies covers, may actually
increase the burden of service users and lead to choosing cheaper services with lower quality. In
either case, families may not be able to access the high quality care service.

14

Previous research on state variations
Studies describing state level variations in individual child care subsidy policies and
quality regulation efforts have also shown substantial insight regarding how individual policy
choices affect availability and affordability of quality child care services. However, most of these
studies did not specifically discuss how different policy levers may interact, thereby shaping
characteristics of state child care environment. Schulman and Blank (2011) documented
significant variations in six individual subsidy policies including income eligibility, copayment
rates, reimbursement rates, waiting list, payment rates for types of care providers, and subsidy
priority. For example, in some states, such as Maine and Mississippi, subsidy eligibility was
given to families whose income was over 75% of state median income, whereas other states
allow only families with income equal or less than 40% of state median income to receive child
care subsidies (Schulman & Blank, 2011). Additionally, differences in eligibility rules may
influence financial burden for subsidy users. For example, a family with one child with an
income at 100% of the federal poverty line requires no copayment in Arkansas, but in North
Dakota the copayment amounts to 17% of family income or $258 (Schulman & Blank, 2011).
Discrepancies in quality regulation across states were also found in prior research. The
National Association of Child Care Resource & Referral Agencies, which is a network of
nonprofit, state sponsored organizations, describes variations in child care licensing and
regulation across states in their reports from 2005 (NACCRRA, 2011). For example, while some
states set staff- to- child ratios as low as 1:4 for children at age 18 months, other states set it for
1:8 or 1:9 for these children. In addition, center based care is inspected for quality monitoring for
more than four times a year in some states, while they are inspected less than once per year in
other states (NACCRRA, 2011). There are state variations in not only states’ quality regulations,

15

but also efforts to improve quality. Using data collected by the survey of child care
administrators, Pittard and colleagues (2006) found that states had diverse activities for quality
improvement. For example, Illinois provided wage supplements to licensed facilities to promote
job stability of child care workers and diminish disruption in relationships between children and
caregivers, while West Virginia focused on improving quality of family home providers and
helped them meet health and safety requirement by providing small grants (Pittard, Zaslow,
Lavelle, & Porter, 2006).
Even if research did not directly investigate differences in child care policies, it is notable
that research on policies targeting low income families with children has examined gaps in social
programs among states. Based on the argument that the devolution from the federal government
to state governments in 1990s increased the differences of state on social policy, Meyers and her
colleagues compared state policy packages based on 11 social programs, including child care
assistance, for low income families and children (Meyers, Gornick, & Peck, 2001). Using cluster
analysis, results suggested that based on variations across multiple policy dimensions, states
were categorized into five distinctive policy approaches; they ranged from the most minimal
provisions, which could be called a conservative approach emphasizing private responsibility, to
the generous provisions which combined direct assistances with employment supports and
enforced family responsibility. Also, comparisons of policy changes between before (1994) and
after welfare reform (1998) suggested that states made substantial changes in policies supporting
low income families with children after the 1996 welfare reforms, but the magnitude and
direction of these changes remained consistent with the state clusters identified in 1994.
Even though this study did not focus on the child care policy solely, the result and
analytic method in the previous study provide insight for the current study. As a critical policy

16

scheme of welfare reform, child care policy may have similar patterns to those policies. In
addition, analyzing policies as a package rather than dealing with them individually seems
applicable to the current study. The cluster of child care policy levers states use with great
flexibility determines the distributions of child care assistance among low income working
families and affects the shape and content of public child care overall (Levy, 2000). Cluster
analysis using a variety of child care policy rules will be relevant to enhance understanding of
child care policy choices and the collective interactions of policy choices in each state.
Previous research on impact of policy on choice of care arrangement across states
Previous studies have examined the impact of child care policy on the choice of high
quality care. However, there is a dearth of research which has examined impacts of both policies
on the choice of child care service among low income families. As child care subsidies have
expanded after the welfare reform, a large body of research has examined the relationship
between subsidies and the use of quality services (Brooks, 2002; Coley, Li-Grining, & ChaseLansdale, 2006; Crosby, Gennetian, & Huston, 2005; Ertas & Shields, 2012; Henly, Ananat, &
Danziger, 2006; Ryan, Johnson, Rigby & Brooks-Gunn, 2011; Tekin, 2005; Weinraub, Shlay,
Harmon, & Tran, 2005). However, majority of studies have examined the impact of “subsidy
receipt” only, not considering effects of “subsidy policy” which are various across states. In this
respect, even though they commonly found that subsidy recipients were more likely to choose
center based (versus home based care) or licensed care (versus unlicensed care) than
demographically similar parents who did not receive subsidies (Brooks, 2002; Coley, Li-Grining,
& Chase-Lansdale, 2006; Crosby, Gennetian, & Huston, 2005; Henly, Ananat, & Danziger, 2006;
Tekin, 2005; Weinraub, Shlay, Harmon, & Tran, 2005), the impacts of subsidy on the choice of

17

center based care across states may not be same. Current policy variations within subsidy policy
rules may differentiate the impact of subsidies measured by previous studies.
There are a few studies that considered policy variations and tried to reflect them in their
analysis. The most common methodological approach to do so was to control variations by
including the variable indicating the state of residence in the analytic model. Very few research
studies examined policy variations in their analysis. Using data from the National Survey of
American Families, Blau and Tekin (2007) examined the effect of subsidy receipt on
employment, school attendance, unemployment, and welfare participation. They used two- stage
least –squares estimates that treat subsidy receipt as exogenous determined by state subsidy
system, such as eligibility, subsidy reimbursement rate, child care subsidy expenditure per capita,
unemployment rate, median income, and the child poverty rate. Interestingly, however, the
preliminary analysis showed that all state related factors were not significant. Another study
which used same data examined the effect of subsidy receipt on joint employment and choice of
child care arrangements. Tekin (2005) used the same analytic method, two- stage model, in the
study on impact of actual subsidy receipt of single mothers on their joint employment and child
care mode decisions. He included two variables as instrumental variables, the percent of eligible
children served by child care subsidies in the state and the average amount of CCDF funds spent
per child in the state. Results showed that fiscal efforts and fiscal capacity expressed those
variables had significant effects on subsidy receipt, as well as employment and use of center
based care. These two studies are good methodological examples to reflect policy variations in
the model. They also revealed state’s fiscal variations had significant impact on subsidy receipt.
However, these studies only revealed how subsidy receipt was affected by state characteristics
and did not examine how different policy levers affects families with and without subsidies.

18

Therefore, they did not examine policy variations and their impact on the choice of care
arrangements for families across states.
There are studies which examined how quality regulation influences the choice of child
care services. These studies commonly concluded that more stringent regulation decreased the
availability of service and the choice of child care service (Gormley, 1999; Hofferth & Chaplin,
1998; Hotz & Xiao, 2011). This is because the stringency of regulation is closely related to the
increased operating cost and service price in most studies. For example, in the study using data
from the National Child Care Survey in 1990, Hofferth and Chaplin (1998) examined the
impacts of several state regulations including training requirement, inspection, child/staff ratio
on the choice of center based care. Based on findings of negative impacts of training
requirements on the choice of center based care, they argued that without increasing government
assistance to pay for training or educating parents to increase their willingness to pay for it,
regulation would reduce the use of quality service. In this respect, if consumers would have
enough purchasing power to buy better quality of service and governments would invest enough
to increase the quality of services, the result would be much different. Another study also
confirmed this assumption with more detailed findings. Unlike other studies which tested
impacts of regulation on choice of service across states, Blau (2007) used variations over time in
regulations within states and also compared effects of regulations between two states which had
different levels of regulations. The results showed that tougher regulation had significant positive
impacts on choice of center based care, but the estimated effects were small. Based on these
findings, he argued that regulations were not self- enforcing in general and other approaches to
improving the quality of child care, such as quality certification and vouchers that are tied to the
quality of care are needed. When considering subsidy reimbursement rates that are tied to quality

19

of providers in several states, his argument should be reflected in the analysis on impacts of child
care policy on the choice of child care arrangements.
To the best of my knowledge, there is only one study which investigated how both
subsidies and regulations shape the choice service between center and non-center care. Using
data from the Child Care Supplement (CCS) to the Fragile Families and Child Wellbeing Study,
Rigby and her colleagues examined effects of subsidies and regulations on the choice of service
among children who were 3 years old in 14 cities (Rigby, Ryan, & Brooks-Gunn, 2007). In the
analysis, impacts of state policy rules, such as subsidy eligibility threshold, subsidy spending,
child to staff ratios and training requirement in center and family care, were tested on choice of
center care and licensed care. They found that both subsidy spending and child to staff ratio had
significant influences on the choice of center based child care. While subsidy spending had
positive association, child to staff ratios in both center based care and family child care had
negative effects on the choice of center care. Even though this study examined impacts of
subsidies and regulation considering state variations, there are several limitations. First, they did
not include several important policy rules in subsidy policy which may have substantial impacts
on shaping child care environment and using quality care, such as copayment rates and
reimbursement rates. Several states also use tiered reimbursement rates systems which make
subsidies cover higher percentages of market prices when families use better quality services.
These policy rules may be stronger incentives to select better quality of service. Furthermore,
this study did not consider the impacts of receiving subsidies on the choice of center based care,
even though subsidies are significantly influential on the use of center based care (Brooks, 2002;
Coley, Li-Grining, & Chase-Lansdale, 2006; Crosby, Gennetian, & Huston, 2005; Henly, Ananat,
& Danziger, 2006; Tekin, 2005; Weinraub, Shlay, Harmon, & Tran, 2005).

20

Purposes of current study
The purpose of this study is to better understand variations in child care policy across
states and how policy decisions and subsequent practices may influence choices of child care
services among families. Although policy variations across states make examining policy
impacts challenging, they also provide an uncommon opportunity to explore how states shape
different child care environments and how various policy rules differentially affect individuals’
child care choices. Adding to previous findings, this study examines variations in child care
policy decisions across states which shape state child care environments and investigate the
influence of both child care subsidies and quality regulations on working families’ choice of care
arrangement among various options of child care services.
Research aims
To isolate and understand the different components of complex policy problems, this
study has two different research aims in navigating policy variations across states and examining
policy impacts.
1. The research aim 1 explores substantial differences across states in subsidy policy and
regulation. More specifically, some states may have more generous subsidies and/or
more stringent regulations than other states. Some states may provide generous
subsidies to specific targets, such as families with income below the federal poverty
level, while other states may provide more restrictive subsidies to relatively broader
populations including both families whose income is below and above the poverty
level; some states may have more stringent regulations for governing over structural
features, such as child-staff ratios and required training hours, while other states
impose less stringent regulations.
21

2. The research aim 2 examines the impacts of state’s child care policy on the choice of

care service. The impacts of various state’s fiscal ability and efforts reflected in the
state expenditure and subsidy eligibility on subsidy receipt are also examined.
A. The research aim 2-A examines how policy levers influence the choice of center
based care.
: Generous child care subsidy policy such as higher eligibility rates and higher
reimbursement rates may have positive associations with the choice of center
based care; stricter regulation, such as lower child-staff ratio may have a negative
association with impacts on the choice of center based care while training
requirement and frequency of inspection may have a positive association with the
choice of center based care.
B. The research aim 2-B explores influences of child care policy on the choice of
various types of care.
: The stringency of regulation policy applied to family home care may be
negatively associated with the choice of center based care, while it may be
positively associated with the choice of family home care versus informal care,
such as care by grandparents or relatives.

22

III.

METHODS

Data and sample
Fifty states in the United States including the District of Columbia constituted the sample.
Idaho was not included because the Child Care Licensing Study in 2011, which was one of main
datasets used for this study did not include information for Idaho.
Given that there was no one single dataset including all information needed to analyze
effects of policy factors on individuals’ choice of service type, I used information from two
different datasets, Child Care and Development Fund (CCDF) Policies Database in 2011 and
Child Care Licensing Study in 2011. All data used in the analysis were publicly available from
the Interuniversity Consortium for Political and Social Research (ICPSR). Data from CCDF
Policies Database in 2011 was the main source for analysis in child care subsidy policy
variations. This data included details on subsidy policy, such as subsidy eligibility limits, family
copayment, and reimbursement rate (Giannarelli, Minton, Durham, & Department of Health and
Human Services, 2012). The Child Care Licensing Study in 2011 was used to examine
variations in quality regulation across states. It was gathered by the National Association for
Regulatory Administration (NARA) to collect information on states' licensing policies, including
staffing, monitoring, and enforcement of licensing regulations (Fischer & Martella, 2013). These
data sets were used to explore child care policy variations in the state level for the research aim 1
and used as independent variables for the research aim 2 to examine the impacts of child care
policy rules on individuals.
To investigate impacts of policy on choice of service types among children (research aim
2), the Survey of Income and Program Participation (SIPP) was used. The SIPP includes
comprehensive information about the income and program participation of individuals and
23

households in the United States, and about the principal determinants of income and program
participation (Westat and Mathematica Policy Research, 2001). The SIPP also produces nationallevel estimates for the U.S. resident population and subgroups. In this respect, the SIPP enables
this study’s evaluation of the effectiveness of federal, state, and local program. The SIPP collects
panel data on general core topics, such as income and program participation, every four months.
In addition, it periodically gathers information on specific topics which are called topical
modules such as child care, financial support, education and employment. This study used the
topical module of SIPP 2008 wave 8 which was collected from January to April in 2011 in order
to include information on child care arrangement for children between the ages of 0-5 in a
particular household. Family characteristics, such as income and family composition were added
from core topics of SIPP 2008 wave 8. To examine the child care policy impacts on the choice of
specific types of care services (research aim 2), children between the ages of 0 to 5 who used at
least one type of non-parental care and whose primary guardian was employed were selected.
The primary guardian in this data could be defined as a primary care-taking person. In this data,
more than 90% of children’s primary guardians were their mother. 2742 children in 2208
households were included in the sample of the study.
Measures
State policy and characteristics
State child care policies including child care subsidy policy and regulation were
measured using the CCDF Policies Database in 2011 and the Child Care Licensing Study in
2011. Child care subsidy policy included four policy rules that may substantially shape the
purchasing power of low income families and reflect states’ capacity to provide subsidies to low
income families: subsidy eligibility thresholds, family copayment rates, state reimbursement
24

rates, and tiered reimbursement rates based on quality. Those child care policy rules were key
elements in child care subsidies and have been used in previous research (Ha & Ybarra, 2013).
Regulation policy included three policy levers: child-staff ratios for center based care, training
requirements, and frequency of inspection. States applied different regulation rules to various
types of services. Regulations rules applied to center based care and family home care were
measured and included in this study. Policy measures were used to examine all hypotheses. In
addition, state’s characteristics which had been revealed to have significant impacts on subsidy
receipt were measured to examine the research aim 2. Table 1 provides descriptive statistics of
each measure.
Table 1. Means, Standard Deviations, and ranges of state policy rules and state characteristics
(n=50)
Mean
Eligibility threshold1
Copayment rate1
Reimbursement rate1
Incentive1
Number of children per staff (center based care) 2
Training hours per year (center based care) 2
Number of inspection per year (center based care) 2
Number of children per staff (family home care) 2
Training hours per year (family home care) 2
Number of inspection per year (family home care) 2
Capacity of center based care (licensed) 2
Capacity of family home care (licensed) 2
Proportion of child care subsidy eligible children3
Expenditure level: Expenditure per child age under 63

55.61
2.97
78.93
6.02
5.56
14.40
1.61
3.84
0.96
1.24
35.54
8.84
37.14
490.71

Standard
deviation
11.75
2.31
14.40
10.38
1.47
6.86
0.98
1.36
4.40
1.03
15.80
9.01
10.30
217.53

Range
35-80%
0-10.13%
43.88-100%
0-50%
3-9 children
0-40 hours
0.2-4 times
2-6 children
0-20 hours
0-4 times
8.12-72.11%
0-39.22%
21.11-58.03%
$207.51-1050.57

1

Source: Child Care and Development Fund (CCDF) Policies Database in 2011.
Source: Child Care Licensing Study in 2011.
3
Source: Author calculated based on Fiscal Year 2011 CCDF State Expenditure Data. & Fiscal Year 2011 CCDF
Data Tables (US-DHHS, 2013) and data from Estimates of child care eligibility and receipt for fiscal year
2009: ASPE Issue brief, December (Office of the Assistant Secretary for Planning and Evaluation, 2012).
2

Eligibility thresholds in each state indicated the maximum income of eligibility criteria.

In this study, the income eligibility threshold was measured by the proportion of state median

25

income (SMI) based on the initial eligibility income for a family of three to reflect the
differences in income levels across states. A higher income eligibility limit indicates a more
generous subsidy provision and a lower one indicates a restrictive subsidy provision. The mean
of eligibility thresholds was 55.61% of SMI (SD=11.75) with the range of 35~80%.
Families who receive subsidies are required to pay copayments based on a share of family

income as part of subsidy program participation. Copayment rates were measured by the
proportion of income for a family of three with $15,000 in annual earnings, which is the closest
to 100% of income at poverty guidelines by the Department of Health and Human Services for a
family of three in available data. A higher copayment rate indicates a restrictive copayment
provision while a lower one indicates a generous copayment provision. The mean of copayment
rates was 2.97% with range of 0 to 10.13% of income (SD=2.31).
States set reimbursement rates and copayment rates to cover the total cost based on a

market rate survey every two years. States’ reimbursement rates were measured by a
proportion of states’ reimbursement compared to 75th percentile of the market price as required
by the Federal government (Schulman and Blank, 2011). The reimbursement rates were different
based on the age of children and those for a child with age of 1 using center based care were
selected for this study. If reimbursement rates are too low to cover the market rates, an extra
financial burden for service users may influence them to choose cheaper services because
providers are reluctant to reduce rates. A higher reimbursement rate is more generous than a
lower one. The mean of reimbursement rates was 78.9% with range of 43.9% to 100% of 75th
percentile of the market price (SD=14.4)1.

1

There were 12 states which did not allow providers to charge parents the difference between the market rate
and the reimbursement rate in 2011 (Schulman and Blank, 2011). Since a full market price causes unaffordable
copayments that parents living in these may face, this study treated those 12 states the same as the other 38

26

To increase the use of high quality center based care, some states (n==31) use tiered
reimbursement rates which differentiate the reimbursement rates among child care providers
according to the quality standard. By doing so, it can encourage child care providers to improve
the quality of their care services (Schulman and Blank, 2011). Since information on quality
standards of all providers is publicly available, this information can help parents to identify and
select providers with better quality service among center based care. In addition, higher
reimbursement rates may decrease families’ actual copayment rates as discussed above. The
incentive to provide quality care was measured by calculating differences in reimbursement
rates between child care providers in the highest and the lowest tiers. Higher incentives indicate
public support to choose high quality service and lower incentives indicate less support to choose
quality service. The incentive to provide quality care ranged from 0% to 50% of market rates
with a mean of 6.02% (SD=10.38).
State child care licensing policy has regulations related to the maximum child-staff ratio
which one staff member can take care of according to the age of children and types of child care
services as a child-staff ratio is associated with quality child care. A low child-staff ratio has
been found to be more beneficial for children (Vandell & Wolfe, 2000). Furthermore, the
number of children one staff member takes care of is directly related to the number of staff
members the facility hires and, as a result, operating cost of facility. The child-staff ratio was
measured by the maximum number of children per one staff member in both center based care
and family home care with children at age of 18 months. Lower child-staff ratios indicate more
stringent quality regulations and higher child-staff ratios indicate less stringent regulation. The
number of children per staff in center based care ranged from 3 to 9 across states with a mean of
states which allowed providers to charge extras to parents to examine the generosity/ restrictiveness of
reimbursement rates across states.

27

5.6 children per staff (SD=1.5). Those in family home care were measured differently. Family
home care does not have different class levels based on age due to the smaller number of
children cared for. The number of children per staff for family home care was measured as
number of children under the age of 2 per staff. It ranged from 2~6 with a mean of 3.84 children
per staff (SD=1.36).
States require staff members in licensed child care to have trainings for a specific amount
of time every one or two years. Training requirements help maintain child care quality by
training staff members, but may impose additional cost to providers by hiring other staff for
replacement during training. Training requirements were measured by the number of hours of
trainings per year which lead teachers should have. A large number of training hours indicates
more stringent quality regulation. The number of hours required for training for lead teachers in
center based care ranged from 0 to 40 hours across states with a mean of 14.40 hours (SD=6.86).
Those for family home care providers ranged from 0 to 20 with a means of 10.96 (SD=4.40).
States visit child care providers to check whether they meet environment requirements,
such as safety and structural requirements. Findings of non-compliance with licensing
regulations during inspections were reported and can be used for service users to evaluate and
choose service providers in some states. Frequency of inspection was measured by the number
of required inspection for center-based care per year the state administrators conduct. More
frequent inspections indicate more stringent quality regulation. The number of inspection per
year for center based care ranged from less than one (0.2 times) to four times across states with a
mean of 1.6 (SD=0.98). Those for family home care ranged from 0 to 4 per year with a mean of
1.24 (SD=1.03).

28

State characteristics were included to examine the research aim 2 in addition to state
policy. Since current research focuses on state level differences, availability of center based care
and family home care in each state were added in the model. They were measured as proportions
of children under the age of 6 served by licensed care service of each type of provider. The
capacity of licensed center based care ranged from 8.12 to 72.11% with a mean of 35.54
(SD=15.80); the capacity of licensed home based care ranged from 0 to 39.22% with a mean of
8.84 (SD=9.01). In addition, previous research found that subsidy receipt has been affected by
state’s characteristics including the amount of CCDF expenditures and proportions of children
who are eligible for child care subsidies among all children (Tekin, 2005). These state level
characteristics were also included as explanatory variables for subsidy receipt in the analysis for
research aim 2. Child care expenditure (CCDF) levels were defined as the expenditure per child
under the age of 6 adjusted by the market rate of center base care service across states, since
there were variations in price of child care services. It ranged from 207.51 to 1050.57 dollars
with a mean of 490.71 (SD=217.53). The proportions of eligible children among children under
the age of 6 were accessed by the estimates of eligible children in 2009 divided by populations of
children with age 0 to 5 in 2011, since there was no more recent data available (Office of the
Assistant Secretary for Planning and Evaluation, 2012). It ranged from 21.11% to 58.03% with a
mean of 37.14 (SD=10.30).
Variables on children’s characteristics
The choice of care arrangements was measured using SIPP to examine the research aim 2.
Characteristics of children, parents (primary guardian), and family were included in the analysis.

29

Table 2. Sample demographics
Unweighted N

Weighted %

Child’s characteristics

2742

100

Primary child care types
Center based care
Family care (home based formal care, non-relative care)
Informal non-parental care (ex. relative, grandparents)
Receiving child care subsidy
Child is a girl
Primary guardian’s characteristics

1295
449
998
117
1329
2208

47.2
16.4
36.4
4.0
48.4
100

1380
321
337
170

60.7
15.3
16.2
7.8

1434
247
527

65.0
11.0
24.0

153
820
920
315
365
1592
1565
1642
M
32.8
4.21
1.24

6.6
36.6
42.1
14.7
16.0
73.1
71.8
74.7
SD
7.07
1.44
0.48

Race/ethnicity
White
African American
Hispanic
Others
Marital status
Currently married
Previously married
Never married
Education
Less than high school
High school graduation
College/associated college graduation
More than college graduation
Household income: in poverty
Regular working schedule
Fulltime employment
N of children under 6 is 1
Age of primary guardian
# of family members (in households)
N of children under age 6 (in households)
Source: 2008 the Survey of Income and Program Participation (SIPP) Wave 8

Primary care arrangement. The primary child care arrangements among children were
used as dependent variables in this study. Since one child can use multiple types of service, the
primary care arrangement was defined as the child care service types at which children spend the
greatest amount of time of all child care services they used. To investigate impacts of child care
30

policy applied to different types of service providers, dependent variables were dichotomously
coded: choice of center based care versus not (all other types); choice of center based care versus
other types of formal care (family home care or non- relative care); family home care versus
informal child care (relative, grandparents, siblings).
Control variables. Children’s ages (aged 0 up to 2 years versus aged 3 years up to 5
years), gender, and receipt of child care subsidies, and family characteristics such as income
level (below the poverty guideline versus not), the number of children aged 0-5 (more than 1
versus one child), and number of people in the households were included in the model. Primary
care takers’ age, education level (less than high school, high school graduation, college
graduation, more than college), work schedule (regular versus irregular) and work status (full
time versus part time), and marital status (married including both cohabiting and living apart,
previously married, never married) were also controlled.
Table 2 describes characteristics of the study sample. Among 2742 children, 1295 (47.2%)
used center based care while 16.4% (n=449) used family care. In this sample, 998 children were
cared by relative, grandparents, or siblings (informal care, 36.4%). Among children in the sample,
4% of children (n=117) received a child care subsidy and 48.4% of children were female. There
were 2208 primary guardians in the data and they were included in households. Primary
guardians, primary care taking people of children in each household, were 33 years old on
average (SD=7.07). Almost two thirds (60.7%, n=1380) were White, 15.3% African American
(n=321), 16.2% (n=337) Hispanic origins, and 7.8% (n=170) “other”. The majority of primary
guardians was currently married (n=1434, 65.0%), 11% were previously married (n=247), and
24% of primary guardians were never married (n=527). The study sample had higher education
levels compared to general population in the original dataset. People who had graduated from

31

college or have an associate college were the majority in the study sample (42.1%, n=920) and
the next largest group was those who had a high school diploma (36.6%, n=820). 14.7% of the
study sample (n=315) had education more than college graduation while 6.6% had education less
than high school graduation (n=153). All primary guardians in the sample were employed and
almost three quarters of primary guardians had regular working schedule (6 am to 6 pm, 73.1%,
n=1592). Fulltime employers (working more than 35 hours per week) were more than 70% of the
study sample (71.8%, n=1565). The number of people in households were 4.21 on average
(SD=1.44). The majority of households had one child under the age of 6 (74.7%, n=1642) and on
average there were 1.24 children with age of 0 to 5 in each household (M=1.24, SD=0.48). The
monthly income of 16% of sample households was below the poverty line (n=365).
Research design
Figure 1. Research framework

State level
- State child care policy
- Expenditure on child care
- % of subsidy eligible children
- Availability of quality service

Individual level
Family and child
Characteristics

Subsidy
Receipt

Choice of service

32

Figure 1 is the research framework describing how state policy and state characteristics
can shape the choice of care arrangements. Previous studies have found that individual families’
choices of service are affected by their family and children’s characteristics, as well as by
subsidy receipt. However, their choices depend on where they reside, since it will be influenced
by state’s child care environments. As mentioned in previous research, subsidy receipt is affected
by state’s fiscal capacity and fiscal efforts. State level child care policy rules and availability of
quality services also influence the child care environments. In the Figure 1, the small oval
indicates the choices of individual families occurred at individual levels. The small oval is
surrounded by the large oval which indicates state’s child care environments varying across
states. The research aim1 focuses on state level variations in the large oval and the research aim
2 analyzes how state level factors in the large oval influence the individuals including their
choice of service and the subsidy receipt in small oval.
Analysis
Two analyses were conducted to correspond to two separate research aims. In order to
examine the research aim 1 which explores the variations in states’ child care policy choices, I
conducted cluster analyses based on states’ child care policy choices in child care subsidies and
quality regulation for center based care. There were several reasons to choose the cluster analysis
for the research aim 1. The purpose of this analysis was to explore how state policy varies. Since
the list of variables used in this study have been never explored in previous research, the analysis
which did not assume a priori what dimensions might exist in the data was required in this study.
Instead, the theoretical model is incorporated in this analysis through the selection of specific
policy factors and their characteristics. Also, rather than investigating what were the factors
which made differences across states, this study focused on the variations in the combination of
33

state child care policy levers which influence the distributions of child care assistance and affect
the content of public child care and state child care environments. Cluster analysis is a
nonparametric multivariate statistical procedure that simultaneously analyzes variation across
multiple variables to organize observations into relatively homogeneous groups (Aldenderfer and
Blashfield, 1984). Cluster analysis parsimoniously reduces complex data by using variation that
exists within the data. Finally, cluster analysis provides some flexibility to interpret results by
making it possible to choose most relevant cluster memberships (number of clusters) after
reviewing results of cluster memberships. This flexibility makes it possible for researchers to
determine the most relevant cluster memberships based on the theory and previous findings.
Based on those advantages, cluster analysis was chosen for this study.
Given this theoretically derived set of measures, cluster analysis categorizes observations
on the basis of variation across multiple policy levers. Two separate cluster analyses were
conducted on two different child care policies using hierarchical cluster analysis: child care
subsidy policy included four policy rules including subsidy eligibility thresholds, copayment rate,
reimbursement rate, and incentive to provide quality care; and quality regulation included three
policy levers, specifically the number of children per staff, training requirement (hours per year),
and frequency of inspection. Hierarchical cluster analysis was run using Ward’s method and
squared Euclidean distances in order to maximize cluster homogeneity and minimize withincluster errors. During the course of running hierarchical cluster analysis, all variables were
standardized since the unit of measure in each variable was different. Chi-squares and ANOVAs
were run to determine the differences between the clusters and to describe initial cluster profiles.
The cluster solution was chosen based on the findings on previous research and differences in
scores across clusters. Analyses were conducted using SPSS.

34

A multilevel logistic regression model was used for research aim 2 which examined the
associations between the choice of particular types of child care and various state policies rules
used in the first analysis. This analytic method was appropriate for the research design where
data is organized at more than one level. In the current study, the units of analysis were
individuals (choice of services) and they were nested in the groups at which the key predictors
(state policy and characteristics) were measured (Raudenbush & Bryk, 2002). In the analysis for
the research aim 2, dependent variables were the choice of specific types of care service and
independent variables are state level child care policy levers. Basically, the child level and the
state level were the main levels in this analysis since this analysis examined the impacts of statelevel variables on the individual children’s choices. However, since there were households which
had more than one child under the age of six, including one more level reflecting household
characteristics in the analytic model had to be statistically considered. The comparison between
the two-level model and the three-level model was conducted to find the better model fit: the
two- level model did not differentiate children and household variables and included both in
level 1 and had state variables in level 2; the three-level model included children variables in the
level 1, households variables in level 2, and state variables in level 3. The result showed that the
three-level model had significantly better fit for this study design (χ2=164.88107, df=1, p<.001).
The current study design assumed that families were randomly distributed across states.
In other words, the assumption was made that families did not consider the state child care policy
when they decided where to live. This assumption enabled the random distribution of families
and children across states. Even though families were randomly distributed across states, there
might be relationships among the choice of service, employment, income levels, and where to
live. It is natural that people living in states where more jobs were available might have higher

35

employment rates and use center based care more than those living in states where there were
less chances to have a job. However, that dependence is treated in the multilevel model design.
In addition, the relationships among variables were reviewed before analysis. The possible
problem of multicollinearity was considered through reviewing correlation coefficients among
variables and conducting multicollinearity diagnostics in SPSS. The results are discussed in the
next section.
The random intercept and slope model were used to examine the impacts of state child
care policy and characteristics on intercept of level 1 (child care environments) and the slope of
subsidy receipt for the study sample. The intercept was estimated based on the argument that the
combination of state child care policy levers shapes state’ child care support for working families
and, as a consequences, child care environments (Levy, 2000). The state level included
individual state policy variables, including subsidy eligibility thresholds, copayment rate,
reimbursement rate, incentive to provide quality care, child to staff ratio, training requirement
(hours per year), and frequency of inspection, to estimate the intercept. The slope of child care
subsidy receipt was estimated based on the previous findings that subsidy receipt was affected by
the states’ fiscal characteristics, such as proportions of eligible children among children age
between 0 to 5, which were determined by eligibility thresholds, and child care expenditure per
child age between 0 to 5 (Blau & Tekin, 2007; Tekin, 2005). Since those characteristics may
differentiate the probability of receiving child care subsidies among families across states, these
two variables were included in the model.
Children’s demographic information as well as subsidy receipt and household
characteristics were entered at the level 1 and level 2 respectively because these vary at the
individual level, as described in Figure 1. Through the level 1 and level 2, the mechanism of
36

choosing child care arrangement among all working families within each states, affected by child
care environments, was analyzed. Individual weight and household weight were included in each
level of analysis. All demographic variables with continuous values such as primary guardian’s
age and the number of family members at the level 2 were group-mean-centered. The level 3
examined the impacts of states’ characteristics and variations which are formed by child care
policy rules. Variables indicating state characteristics and state child care policy were entered at
the level 3 to predict the choice of care arrangements and actual subsidy receipt among sample
across states. The research aim 2-A was examined using all study samples. To examine Research
aim 2-B, populations who used specific types of services were selected to estimate the impacts of
child care policy rules on 1) the choice of center based care versus family home care and 2) the
choice of family home care versus informal care services.
Level 1:
														Log	 ൤

P
൨ = 	β0 + β1ሺVector	of	child′ s	characteristicsሻ + 	β2ሺsubisdy	receiptሻ + R
1	 − P

Level 2:
														β0 = 	γ00 + γ01ሺVector	of	Family	Demographicsሻ + e
Level 3:
							γ00 = 	δ01ሺeligibility	thresholdsሻ + δ02ሺcopayment	ratesሻ
+ 	δ03ሺincentive	to	provide	cneter	careሻ + δ04	ሺreimbursement	rates	ሻ
+ δ05ሺchild	to	staff	ratio	for	center	based	careሻ + 	δ06ሺtraining	requirementሻ
+ 	δ07ሺfrequency	of	inspectionሻ + 	δ08ሺcapacity	of	child	care	servicesሻ + 	ε
	β2 = 	δ01ሺstate	child	care	expenditureሻ
+ δ02ሺproportion	of	subsidy	eligible	child	among	children	age	under	6	ሻ + ε

37

In multilevel logistic regression model, intra-class correlation coefficients cannot be
calculated in the traditional way because the individual level variance and the state level variance
are not directly comparable: Whereas the state level residual variance is on the logistic scale, the
individual level residual variance is on the probability scale (Merlo et al., 2004). As an
alternative, it was calculated by the equation below, suggested by Snijders and Bosker (1999):
τଶ଴
ρ୍ =
ଶ 	
τଶ଴ + π ൗ3
τ02 indicates a variance component for higher level and pi (≈3.14) squared divided by 3 is
used as a substitute for a variance components for level 1. Multilevel logistic regression analyses
were conducted by HLM software using penalized quasi-likelihood estimation.

38

IV.

RESULTS

Research aim 1
Child care subsidy
The five-cluster solution was chosen in the cluster analysis on child care subsidy policy
based on the previous studies finding the substantial impacts of low copayment rates and higher
reimbursement rates on subsidy receipt and the choice of center based care. Anova F test showed
there were significant differences in mean scores of eligibility rates (F=13.41), copayment rates
(F=6.42), reimbursement rates (F=17.71), and incentives (F=31.04) among groups at p=.001
level.
Figure 2. Differences in child care subsidy policy across clusters
3.00

(1) Generous
provision
(n=11)

Z
score

2.00

1.00

(2) Moderate
provision
(n=12)

0.00

(3) Restrictive
provision (n=5)
(4) Incentive to
quality care
(n=8)

-1.00

-2.00

Eligibility
rate

Copayment

Reimbursemen
t rate

Incentive

(5) Support for
low income
family (n=14)

Figure 2 illustrates standardized mean scores of child care subsidy policy measures and
the number of states included in each cluster. Five clusters were characterized as followed: states
with generous provision; those with moderate provision; those with restrictive provision; states
39

which provide generous incentive to high quality care; and states which provide support focusing
on extremely low income families. Three clusters among five were characterized according to
the level of subsidies provided: generous provision, moderate provision, and restrictive provision.
Two clusters were characterized by the focuses of child care subsidy policy: providing more
incentive to quality care; and providing more subsidies to extremely low income families.
States in cluster 1 can be described as generous subsidy provisions. 11 states (22%)
included in this cluster had the most generous eligibility rate and the second smallest copayment
rates: in other words, these states provided child care subsidies to families with higher income
and imposed less family copayment, compared to states in other clusters. Average
reimbursement rates in cluster 1 were also the highest across clusters. However, incentive to use
high quality care was very low. Overall, states in the ’generous provision’ cluster focused on
providing child care subsidies to more families by including those with higher income and
reducing financial burdens of subsidy recipients. Cluster 2, which comprised 12 states (24%),
was characterized by moderate provision for child care subsidies with no incentive to choose
quality service. States in cluster 2 had averagely moderate eligibility limits. Even though family
copayment rates were the highest, reimbursement rates were relatively high and did not seem to
impose additional financial burden to subsidy families. There was no incentive to choose quality
service. Overall, states in cluster 2 can be characterized by providing moderate child care
subsidies with no support for use of quality service. Cluster 3 was characterized by restrictive
provision for child care subsidies. 5 states (10%) in cluster 3 had the lowest eligibility level
across clusters in average. Even though copayment rates were moderate, close to overall mean,
reimbursement rates were the lowest. In this case, providers may require service users to pay
additional cost in addition to their copayment and actual family copayment rates may increase.

40

The incentive for choosing quality care was moderate. Overall, these states provided
provide child care
subsidies to only families with lower income and potentially imposed
impose higher copayment rates to
them by having low reimbursement rates. Cluster 4 was characterized by the strong incentive to
choose quality care. 8 states (16%) which had this cluster membership provided
provide child care
subsidies to families with relatively higher income than average of all states with moderate
copayment rates. Reimbursement
ement rates were low, but the incentive to choose high quality care
was the highest on average. Overall, states in this cluster focused on encouraging families to
choose better quality of center based care and supported providers to increase quality. States
State in
cluster 5 were characterized by having generous provision in child care subsidies to families with
low income.. 14 states (28%) in this cluster imposed the lowest family copayment, had second
highest reimbursement rates, and provided moderate the incentive
tive to choose better quality of
service. Only eligibility rates were relatively low across all clusters. Overall, these states
provided generous child care subsidies but focused on only families with lower
low income.
Quality regulation
Figure 3. Differences in quality regulation policy rules across clusters
3.00

(A) Loose
regulation (n=7)

Z score
2.00

(B) Less stringent
child-staff ratio
(n=6)
(C) Most stringent
child-staff ratio
(n=16)
(D) Stressing staff
training (n=16)

1.00

0.00

-1.00

-2.00

N of children
per staff

N of Training hours
41

Inspection
per year

(E) Frequent
inspection (n=5)

An additional cluster analysis was conducted to categorize states based on child care
quality regulation rules. The five-cluster solution was chosen in the cluster analysis on quality
regulation based on the findings of previous research that child to staff ratio had substantial
influence on the market entry of providers. The group differences among clusters were also
considered in the cluster choice. Anova F test showed there were significant differences in mean
scores of the number of children per staff (F=32.54), the number of training hours required
(F=14.28), and the number of inspection (F=22.21) among clusters at p=.001 level.
Figure 3 describes standardized mean scores of quality regulation measures and the
number of states included in each cluster. Cluster A included 7 states (14%). These states had the
second highest child to staff ratios among clusters. The average hours of training required by
lead teachers per year and the number of inspection states conduct annually were the smallest
across the clusters. Overall, by having less stringent regulation and conducting inspections least
frequently, these states imposed the least regulation on child care providers in terms of quality
control. Cluster B was characterized by lenient rules on the number of children per staff. 6 states
(12%) in cluster B allowed one staff member to take care of the largest number of children (8
children) across clusters on average. Hours of training required per year were small to moderate
and the average number of inspections conducted per year was second highest. Overall, these
states had less stringent regulation on the number of children per staff, but tried to regulate
quality through frequent monitoring and inspection. Unlike cluster B, cluster C had most
stringent regulation on the maximum number of children per staff. In 16 states included in the
cluster C (32%), one staff member could take care of up to 4 children aged 18 months on average,
which was half compared to the states in cluster B. However, training hours required were
relatively small and the number of inspection per year was also below the average of all states.
42

Overall, theses states seemed to focus on regulating quality of providers by imposing more
stringent rules on the number of children. Cluster D included 16 states (32%) which were
characterized by their emphasis on staff training. These states required the largest amount of
hours for annual training to teacher. The number of children one staff member takes care of and
the number of inspection conducted per year were moderate. Cluster E was characterized by the
frequent inspection. 5 states in this cluster conducted most frequent inspections on providers
across clusters. The maximum number of children per staff and required training hours per year
were moderate.
Table 3. Cross tabulation of child care subsidy and quality regulation cluster memberships

Quality regulation (QR)

Child care subsidy
policy (CCS)
(A) Loose
regulation
(B) Least stringent
child-staff ratio
(C) Most stringent
child-staff ratio
(D) Stressing
staff trainings
(E) Frequent
inspection
Total

1
Generous
provision

2
Moderate
provision

3
Restrictive
provision

4
Incentive to
quality care

5
Supporting
low income
family

Total

1

3

1

0

2

7

5

1

0

0

0

6

0

4

3

3

6

16

4

4

1

4

3

16

1

0

0

1

3

5

11

12

5

8

14

50

Table 3 shows the memberships of states in the combination of both clusters. States
providing more generous child care subsidies (Cluster 1) were likely to have less stringent
regulation rules than those with other cluster memberships. More than a half of those states (6
out of 11) in Cluster 1 either allowed one staff member to take care of more children (Cluster B)
or had loose regulations (Cluster A), while those with other cluster memberships were less likely
to have less stringent regulation memberships (Cluster A and B). There was no state which had
43

cluster memberships with stringent (low) child-staff ratios (Cluster C) among those in Cluster 1.
There are other states which had a tendency to reinforce quality of center care providers by using
both child care subsidy and regulation rules. All states in Cluster 4 which provides more
incentives to providers with high quality were included in either Cluster C, D, or E, which used
different regulation rules to improve or maintain quality of center care services and did not share
the cluster memberships with less stringent regulations (Cluster A and B). Also, most states
which imposed less stringent child-staff ratios (Cluster B) provided generous child care subsidies
to families. 5 out of 6 states in cluster B in quality regulation were included in cluster 1 in child
care subsidy policy. Furthermore, states which had more stringent child-staff ratios (Cluster C)
either provided relatively less generous subsidies or had specific policy targets. For example,
states in cluster C shared the cluster memberships of cluster 2 thru 5 quite evenly, but did not
have a cluster membership of cluster 1 which had the most generous subsidy provision.
Research aim 2-A

The research aim 2 focused on how various policy rules described in the previous
analysis were associated with the choice of child care services. First analysis for the research aim
2-A investigated the associations between the choice of center based care and child care subsidy
rules and quality regulation rules. For the research aim 2-B, two analyses were conducted to
examine the impacts of policy rules on the choices of specific types of care by selecting specific
populations: 1) policy influences among formal care, on the choice of center based care versus
family home care; 2) policy influences on the choice of family home care (formal care) versus
relative care (informal care).

44

Table 4. Bivariate correlation coefficients among sample demographics

1. Child’s age
2. Child’s gender
3. Receiving subsidy
4. Parent’s age
5. Parent’s race
6. Marital status
7. Parent’s education
8. Regular work
9. Full time
10. N of family
11. N of children
12. In poverty

1
-.01
.03
.19*
.04*
-.02
-.05*
.01
-.01
.08*
-.02
.04*

2

3

4

5

6

-.00
.04*
-.02
.01
.01
-.01
.01
.02
-.01
-.00

-.08*
.01
.16*
-.10**
-.02
.01
.03
.05*
.11*

-.04
-.37*
.28*
.10**
.04
.07*
-.07*
-.18*

.15*
-.19*
.02
.05*
.15*
-.04
.16*

-.35*
-.15*
-.04
-.06*
-.08*
.36*

7

8

9

10

11

.15*
.07* .26*
-.17* -.05* -.08*
.01 .01 -.08* .28*
-.33* -.15* -.20* .07*

.05*

* p <.05

Table 4 shows bivariate relationships among variables used in the analysis. A majority of
correlation coefficients were significant. The highest correlation coefficient as an absolute value
was found in the relationship between parent’s marital status (currently married, previously
married, never married) and parent’s age, but it was less than .4 (r=-.37). Parent’s marital status
also had significant correlations with parent’s education (r=-.35) and the income level (below the
poverty or not, r=.36). Additionally, multicollinearity diagnostics was conducted to examine the
presence of higher intercorrelated predicted variables in the model. There was no variance
inflation factor (VIF) which was larger than 3.

45

Table 5. Result for multilevel analysis: choice of center vs. other types (n=2742)

Intercept
N of children per teacher
Training hours per year
N of inspection per year
Eligibility threshold
Copayment rates
Reimbursement rates
Incentive to better quality service
Center capacity
Receiving Subsidy
% of subsidy eligible CHD
Expenditure level
Age of children (3~5 vs. 0~2)
Gender (female vs. male)
Parent's age
Number of family
Income under poverty
Race (vs. White)
Black
Latino
Others
Education (vs. high school grad)
Less than high school
College graduation
More than college grad
Marital Status (vs. currently married)
Previously married
Never married
Regular work schedule (vs. irregular)
Fulltime employed (vs. part time)
N of children under 5 (vs. 1)
Variance component Level 2
Level 3
ICC
Level 2
Level 3

Step1
Estimate (SE)
-1.35 (0.18)**

Step2

Final

1.14 (0.34)**

Estimate (SE)
-1.42 (0.18)**
0.18 (0.06)**
-0.002 (0.01)
-0.05 (0.07)
-0.01 (0.01)
0.07 (0.03)*
0.001 (0.01)
0.01 (0.01)*
-0.01 (0.01)
1.18 (0.34)**

1.43 (0.12)**
0.01 (0.09)
0.02 (0.01)*
-0.08 (0.04)
-0.21 (0.17)

1.44 (0.12)**
0.01 (0.09)
0.02 (0.01)*
-0.08 (0.04)
-0.22 (0.17)

Estimate (SE)
-1.42 (0.18)**
0.18 (0.06)**
-0.002 (0.01)
-0.05 (0.06)
-0.005 (0.01)
0.07 (0.03)*
0.001 (0.01)
0.01 (0.01)*
-0.01 (0.01)
1.19 (0.34)**
-0.01 (0.05)
0.0002 (0.002)
1.44 (0.12)**
0.01 (0.09)
0.02 (0.01)*
-0.09 (0.05)
-0.22 (0.17)

0.43 (0.16)**
-0.17 (0.19)
0.27 (0.18)

0.42 (0.15)**
-0.17 (0.18)
0.29 (0.18)

0.42 (0.16)**
-0.17 (0.18)
0.29 (0.18)

-0.32 (0.22)
0.28 (0.14)*
0.62 (0.17)**

-0.34 (0.22)
0.29 (0.14)*
0.62 (0.17)**

-0.34 (0.22)
0.29 (0.14)*
0.62 (0.17)**

-0.27 (0.19)
-0.12 (0.18)
0.37 (0.12)**
0.01 (0.16)
-0.17 (0.13)
1.37**

-0.28 (0.19)
-0.12 (0.18)
0.36 (0.12)**
0.003 (0.16)
-0.15 (0.13)
1.38**
0.057**
0.30
0.012

-0.28 (0.19)
-0.12 (0.18)
0.37 (0.12)**
0.01 (0.16)
-0.15 (0.13)
1.38**
0.057**
0.30
0.012

0.29

* p <.05. ** p <.01.

Table 5 presents results of logistic multilevel regression models predicting the use of
center based care compared to all other types of care. In the step1, demographic variables

46

including characteristics of children and family were included in level 1 and 2. Step 2 model
included individual child care policy variables to estimate an intercept (random intercept model)
and the final model added child care policy rules and state characteristics to estimate slope for
child care subsidy receipt. The variance component for the step 2 model was very low as 0.012
and it did not change in the final model after adding predictors of slope for subsidy receipt.
Throughout the steps, impact of demographic variables on the choice of center based care rarely
changed. In the final model, older children (age of 3~5) were more likely to use center based care
than younger children (age of 0~2, β=1.44, SE=0.12). Demographic characteristics, such as
parent’s age (β=0.02, SE=0.01), African American race (vs. White, β=0.42, SE=0.16), having
education equal to or more than college graduation (vs. high school graduation, β=0.29 and
SE=0.14 for college education, β=0.62 and SE=0.17 for more than college education), and
having a regular work schedule (vs. irregular schedule, β=0.37, SE=0.12) were significantly
associated with the choice of center based care. Among child care policy rules, child-staff ratio
for center based care, copayment rates, and incentive to better quality service had significant
positive impacts on the choice of center based care than other types of services. Higher childstaff ratios, which allow one staff member to take care of a large number of children, were
positively associated with using center based care among working parents (β=0.18, SE=0.06).
Also, there was significant association between higher copayment rates and the use of center
based care (β=-0.05, SE=0.02), showing that more incentives to center based care were
associated with using center based care among children under the age of 6 (β=0.01, SE=0.01).
Receiving child care subsidies had significant impacts on the choice of center based care (β=1.19,
SE=0.34), but there was no association between receiving child care subsidies and state
characteristics.

47

Research aim 2-B

Table 6 shows the results of logistic multilevel regression with random intercept and
slope to examine different impacts of child care policy on the choice of various types of care.
The first column describes the result of how child care policy influenced the choice of center
based care versus family home care (n=1744). Both policy rules regulating center based care and
family home care as well as child care subsidy rules were included to the analysis. In the final
model, demographic factors including older age of children (3~5 vs. 0~2 years old, β=1.42,
SE=0.19), parent’s age (β=-0.04, SE=0.01), African American race (vs. White, β=1.23, SE=0.38),
and regular work schedule (vs. irregular, β=0.52, SE=0.16) were associated with the choice of
center based care versus family home care. There was an association between choosing center
based care and having education more than college graduation, but it was marginally significant
(β=0.44, SE=0.24, p=0.071). There was no significant relationship between receiving subsidies
and selecting center based care versus family home care. Stringency of child-staff ratio in center
based care was significantly associated with the choice of center based care. Higher child-staff
ratios for center based care were positively associated with the choice of center based care versus
family home care (β=0.24, SE=0.10). Training requirements for family home care had a negative
association with choosing center based care versus choosing family home care (β=-0.07,
SE=0.03). The variance component for the final model was 0.05.

48

Table 6. Results of multilevel analysis: policy impacts on choice of specific care

Intercept
Center care (CC) capacity
N of children per teacher in CC
Training hours per year in CC
N of inspection per year in CC
Family care (FC) capacity
N of children per teacher in FC
Training hours per year in FC
N of inspection per year in FC
Eligibility thresholds
Copayment rates
Reimbursement rates
Incentive to better quality service
Receiving Subsidy
% of subsidy eligible CHD
Expenditure level
Age of children (3~5 vs. 0~2)
Gender (female vs. male)
Parent's age
Number of family members
Income under poverty
Race (vs. White)
Black
Latino
Others
Education (vs. high school grad)
Less than high school
College graduation
More than college grad
Marital Status (vs. currently married)
Previously married
Never married
Regular work schedule (vs. irregular)
Fulltime employed (vs. part time)
N of children under 6 (vs. 1)
Variance component: Level 2
Level 3
ICC
Level 2
Level 3

Center vs. family home care
(n=1744)
Estimate (SE)
0.07 (0.30)
-0.01 (0.01)
0.24 (0.10)*
-0.01 (0.02)
-0.001 (0.17)
-0.01 (0.02)
-0.19 (0.10)
-0.07 (0.03)*
-0.10 (0.15)
0.0004 (0.01)
0.05 (0.05)
-0.002 (0.01)
0.03 (0.02)
0.84 (0.52)
0.03 (0.07)
-0.004 (0.002)
1.42 (0.19)**
-0.20 (0.12)
-0.04 (0.01)**
0.03 (0.07)
-0.10 (0.24)

Family home vs. informal
(n=1447)
Estimate (SE)
-1.64 (0.33)**

1.23 (0.38)**
-0.08 (0.28)
0.30 (0.39)

-1.06 (0.35)**
-0.16 (0.30)
0.02 (0.39)

-0.09 (0.33)
0.29 (0.25)
0.44 (0.24)+

-0.19 (0.35)
0.20 (0.20)
0.43 (0.20)*

0.25 (0.24)
-0.31 (0.29)
0.52 (0.16)**
-0.28 (0.20)
-0.33 (0.22)
1.34
0.23**
0.29
0.05

-0.45 (0.29)
0.12 (0.27)
0.05 (0.15)
0.50 (0.26)+
0.36 (0.25)
1.67+
0.12**
0.34
0.024

+ p <0.075. * p <.05. ** p <.01.

49

0.04 (0.01)*
0.11 (0.10)
0.07 (0.02)**
0.08 (0.08)
-0.01 (0.01)
0.05 (0.06)

1.25 (0.70)+
-0.08 (0.09)
0.006 (0.003)*
0.04 (0.13)
0.27 (0.11)*
0.07 (0.01)**
-0.20 (0.07)**
-0.30 (0.28)

The second column in Table 6 shows the policy impacts on the choice of family home
care compared to the choice of informal care (grandparent, siblings, or relatives). Policy rules
applied to family home care including regulations rules, eligibility thresholds, and copayment
rates were included in this model. Demographic characteristics including gender of child (female,
β=0.27, SE=0.11), parent’s age (β=0.07, SE=0.01), the number of family members (β=-0.20,

SE=0.07), African American race (vs. White, β=-1.06, SE=0.35), and having an education level
of college graduation (vs. high school graduation, β=0.43, SE=0.2) had significant associations
with using family home care versus informal care. Having full time jobs had a marginally
significant association with choosing family home care versus informal care (β=0.50, SE=0.26,
p=0.053). Among state variables, family home care capacity and training requirement for family
home care had significant associations with the use of family home care versus informal care.
Higher family home care capacity was significantly associated with choosing homes based care
versus informal care (β=0.04, SE=0.01). More stringent training requirement (hours for training)
also was positively associated with choosing family home care versus informal care (β=0.07,
SE=0.02). There was a marginally significant association between using family home care versus
informal care and receiving child care subsidies (β=1.25, SE=0.70, p=0.074) and it was affected
by state expenditure (β=0.006, SE=0.003). The variance component for the final model was
0.024.

50

V.

DISCUSSION

The purpose of this study was to explore variations in child care policy across states and
assess policy impacts on choices of child care services among working families with young
children. Specifically, this study focuses on revealing how states shape child care strategies by
adopting policy rules differently and examining how those policy levers shape state child care
environments and as a consequence, are associated with individual’s choices of care services. In
the first analysis of the current study, 50 states were grouped based on the child care subsidy and
quality regulation to uncover variations across state. Cross tabulations of two cluster
memberships showed diverse state child care policy combinations. In the second analysis, this
study examined the associations of policy rules with the choice of specific types of services
among working families with children age under 6. The result showed that high copayment rate,
high child to staff ratio (less stringent rules), and high incentive to quality providers had positive
associations with the choice of center based care. Subsequent analysis by choosing specific
populations using particular types of care revealed that quality regulations had significant
associations with the choice of services while subsidy rules did not. By revealing variations in
child care policy across states and influence of policy rules on the choice of specific types of care,
this research contributes to better understanding how states shape policy rules and provides
essential knowledge to help states designing child care strategies relevant for their circumstances.
States’ variation among clusters

Unlike previous studies focusing on one side of child care policy, the current study
provides comprehensive descriptions on policy strategies on child care by examining both
subsidy policy and quality regulation. The policies states have adopted have been a major factor
driving interstate differences in welfare provisions (Mettler, 2000; Rector & Youssef, 1999), and
51

this was also true in child care policy. Cluster analysis to navigate diverse policy choices among
states showed that each state had different policy choices and levels of emphasis in child care
subsidy policy and quality regulations. Cross tabulations of two different cluster memberships
(see Table 3) shed light on tendencies that states used in adopting their policy schemes; some
states had both less stringent regulation and relatively generous child care subsidy provision
while others had both less generous child care subsidy and relatively more stringent regulation.
Interestingly, there was no state which had both stringent quality regulations and generous child
care subsidy provisions. States which had both loose regulation and restrictive child care subsidy
provisions were also rare.
Critics have argued that generous subsidy provision and stringent regulations are
beneficial for families by imposing less financial burdens to families and providing better quality
of service (Grobe, Weber, Davis, 2008; Vandell & Wolfe, 2000). However, there was no such
combination for any state in this study. Examining with a state-level perspective provides some
insight into how and why variations exist in childcare service utilization. Among five states with
generous provision and less stringent child-staff ratios (cluster 1 and cluster B), three states had
very small capacity in family home care. In this situation, center based care is the only facility
which provides formal child care. Too stringent quality regulation leads to decreases in the
supply of center based care by discouraging market entry of service providers in those states
(Hotz & Xiao, 2011). Those states would like to intentionally increase or maintain the overall
number of center based care providers in the market by providing more subsidies and allowing
higher child-staff ratios. In addition, one state which provided generous child care subsidies but
had loose regulations (cluster 1 and cluster A) had the largest population of children under the
age of 6 and had relatively low center based care capacity among states. This state may

52

intentionally have loose regulations of center based care and provide generous subsidies to
encourage market entry of providers.
These findings suggest that a state’s particular environmental characteristics including
population composition and child care market situations may be substantial factors creating
different policy choices across states. States’ intentional utilization of child care policy rules
indicates that policy rules which were argued as ideal for working families and children may not
be beneficial in all cases.
Different policy impacts on individuals versus overall populations

The second analysis examined the impacts of policy rules on the choice of specific types
of care using multilevel analysis. Positive associations were found between higher (less stringent)
child-staff ratios, higher copayment rates, and the choice of center based care. This result
confirmed findings from previous studies that less stringent regulation increased the provision of
center based care (Hotz & Xiao, 2011) and the choice of center based care versus other types of
care (Rigby, Ryan, & Brooks-Gunn, 2007). As shown in the cluster analysis, this may be
especially true in states where center based care is the primary licensed care and allowing center
care providers to take care of more children is critical in increasing supply of licensed service.
Positive impacts of higher copayment rates were also argued in the previous research. Levy and
Michel claimed that copayment levels were used as a lever to limit the number of families served
by the (child care subsidy) system: the higher the level of copayments, the more money available
for child care and thus the wider the distribution of services over the eligible population (Levy &
Michel, 2002).
Interestingly, those findings are opposite to the arguments of previous studies on impacts
of policy rules on families and children. The positive association between higher copayment
53

rates and the choice of center based care is in contrast to arguments that higher copayments make
families unable to continue to receive a child care subsidy since they could not afford to pay their
share of child care costs, which could lead them to turn to less satisfactory child care
arrangements (Grobe, Weber, Davis, 2008; Long, Kirby, Kurka, & Waters, 1998). In addition,
higher child-staff ratios may increase the provision of services, but they may not be beneficial for
children’s development by decreasing overall quality of series (Vandell & Wolfe, 2000).
These opposing or contradictory results should be interpreted with caution. Since the
current study only considered the policy influence of copayment rates for the family of three
with income less than federal poverty line, it may be different when considering copayment rates
for families with different income levels. In terms of regulation impacts, this study considered
how child-staff ratios are associated with the choice of services, rather than how it was
associated with the quality of services. In this respect, interpreting results should consider the
ways in how these results can improve the policy implementation to better address its goals and
increase the service quality available for families so that it will provide more benefits to them, in
regard to various choices and preferences of child care services among families.
By shedding light on different influences of policy rules based on perspectives, these
findings suggest the ways how to design child care policy. Considering needs of target
populations and child care market situations, states should carefully design child care policy
which balances between quality and quantity of service, and between financial burdens of
individuals and child care subsidy provision to more families.
Significance of quality regulations on the choice of services

The incentive given to quality providers was found to have a significant association with
the choice of center based care. It seems warranted as an incentive can actually decrease
54

copayment rates and higher incentives may encourage potential providers to enter the market by
promising more public supports. However, the most noteworthy point is that the policy lever
which aims at improving quality of center based care is also positively associated with choosing
services with better quality. Significant impacts of policy levers targeting service providers on
the choice of better quality services were also revealed in the result of the research aim 2-B.
Results showed that the choices of formal care, both center based care and family home care,
were significantly affected by training requirement for family home care, while child care
subsidy rules did not have any associations with the choice of services. This is especially critical
when comparing the impacts of subsidy receipt on the choice of different types of care services.
While the choice of service among formal care, center based care versus family home care was
not affected by subsidy receipt, the choice of family home care versus informal care was
positively associated with child care subsidy receipt and state expenditures.
Those findings suggest that quality regulation which targets at maintaining and
improving service quality can have positive impact on choosing better services among families.
It indicates that quality improvement may have a substantial role on the choices of care services
with better quality and higher price, which child care subsidies aim at increasing. Especially,
increasing overall quality of services including family home care and informal care eventually
improve the impacts of subsidy receipt on children. However, current child care policy mainly
focuses on providing subsidies to demand-side (service users) and leave supply side in the
market function. Child care expenditures are mainly used to provide monetary benefits to service
users through the child care subsidy. Federal regulation requires states to spend more than 4% of
child care expenditure on quality activities, but it is still very low and the main target is licensed
providers. Under the current low investment on quality improvement system, increasing the

55

proportions of the funds used for quality improvement and expanding supports for provider to all
family home care providers and informal care will improve policy benefits of child care subsidy
policy.
Limitations

This study is a significant contribution to the literature as it reveals various impacts of
policy rules and how they operate in relation to the choice of specific types of care services, but
it is not without limitations. First, this analysis relies on child care policy rules and does not
include other state factors which may be associated with state’s policy choices, such as labor
market situation, related policy rules (for example, TANF or Head Start), and more detailed
population compositions. By affecting the child care needs, labor market participations of
mothers within states can be also related to the subsidy receipt and the choice of care services
among working families. Therefore, further research which comprehensively considers policy
related factors is needed to improve understanding of states’ choices of child care policy and the
associations with other public assistance programs.
Second, across all analyses, intra-class correlation coefficients which account for how
much the variance of higher level (state) account for total variance for individuals were shown to
be low. Studies suggest that the intra-class correlation coefficient is often between 0.05 to 0.20 in
social science research and in education research (Hedge & Hedberg, 2007; Snijders & Bosker,
1999). Based on these, the intra-class correlation coefficients for the research aim 2-A was
acceptable (0.05), while it was very low for the research aim 1 and for the research aim 2-B
(0.012 and 0.024 respectively). However, scholars argue that in multilevel analysis, higher level
variables which indicate environmental factors cannot be ignored because the intra-class
correlation coefficient cannot provide sufficient insight to the necessity of higher level variables
56

(Liska, 1990; Muthén & Satorra, 1995). In addition, contrary to normally distributed continuous
variables, having dichotomous response variables is tricky to investigate the proportions of
variance (Larsen & Merlo, 2005). In other words, using intra-class correlation coefficients in
logistic model may not be relevant to interpret the necessity of intervention and proportions of
variance explained by higher levels. Instead, by making it possible to test the theories that link
macro and micro levels in adequate ways, multilevel analysis is very useful in many studies
(Liska, 1990) and has provided meaningful insight in this study as well. It is more critical in
policy research, because in many cases, it is easier to intervene and change the policy level
factors than individual level factors. In this respect, even though the intra-class correlation
coefficients are very small, the analysis of this study can be meaningful.

57

VI.

CONCLUSION AND IMPLICATION

After welfare reform in 1996, states’ influences on local child care environments and
families’ care arrangements has dramatically increased, and states’ policy decisions in child care
funding, policy rules, quality control, and target populations have varied greatly. This study
contribute to increasing the knowledge of how state-level policies and regulations affect child
care decisions by families, offering policy areas that can be targeted for increasing provision of
quality service and in supporting the choice of better quality service by families.. Using
variations in policy implementation across states, this study subsequently reveals how policy
rules operate in the child care environment and influence individual’s choices. In addition,
examinations of the policy influences on the specific types of services by dividing care service
using various categories uncovers the importance of quality regulation on the choice of child care
services. Findings in this study provide key policy-level areas that could be emphasized in order
to make child care policy more beneficial to the families they serves. In addition, by revealing
that quality regulation was closely related with the choice of services, this study suggests specific
policy schemes which can benefit families.
Future research is needed to provide better understanding on state’s child care policy
decisions, especially those which more broadly examine factors affecting state’s child care
decisions are needed. The relationship with other policies, such as TANF and early childhood
education provision, can increase knowledge on state’s policy choices. Future research should
examine the ways that the state’s political environment, policy practices including political and
cultural attitudes to target populations, and detailed funding allocation under limited budget
constraint affect policy choices. In terms of improving overall child care system, contemplating
58

how to connect quality improvement with child care subsidy policy will be required for policy
makers, social workers, and researchers in the child care area in the future. In addition, studies
which more directly examine the policy impacts on the child care contexts in low income
communities, such as availability, market price, and compositions of various child care services,
will help to improve the understanding of policy operations and its subsequent impacts on target
populations.
Findings of the current study can increase knowledge of current child care system among
social workers. This will help social workers in child care policy area advocate policy choices
which increase available services and improve overall quality of child care services for low
income working families. In addition, advocacy for quality improvements of child care which
enhance children’s developments can be supported based on the findings of this study. Finally,
by increasing knowledge on current child care system among social workers, this study will help
social work practitioners working directly with low-income working families with children to be
well aware of and understand available supports and to recommend most relevant services for
them.

59

REFERENCES

60

REFERENCES

Adams, G., Rohacek, M., & Snyder, K. (2008). Child care voucher programs: Provider
experiences in five counties. Washington, DC: Urban Institute.
Administration for Children & Families. (2012). Office of Child Care Fact sheet. Administration
for Children & Families. Retrieved from http://www.acf.hhs.gov/programs/occ/factsheet-occ.
Albers, E. M., Riksen-Walraven, J. M., & de Weerth, C. (2010). Developmental stimulation in
child care centers contributes to young infants’ cognitive development. Infant Behavior
and Development, 33(4), 401-408.
Aldenderfer, M. S. & Blashfield, R. K. (1984). Cluster analysis. Sage University Paper Series on
Quantitative Applications in the Social Sciences, 07-044. Beverly Hills: Sage.
Alkon, A., To, K., Mackie, J. F., Wolff, M., & Bernzweig, J. (2010). Health and safety needs in
early care and education programs: What do directors, child health records, and national
standards tell us? Public Health Nursing, 27(1), 3-16.
Blau, D, M. (2003). Child care subsidy programs. In R. Moffitt (Ed.), Means-tested transfer
programs in the United States (pp.443-516). University of Chicago Press.
. (2007). Unintended consequences of child care regulations, Labour Economics, 14
(3), 513-538.
Blau, D. M. & Currie, J. (2006). Pre-School, Day Care, and After-School Care: Who's Minding
the Kids? Handbook of the Economics of Education, Elsevier.
Blau, D. M & Mocan, H. N. (2002). The Supply of Quality in Child Care Centers. Review of
Economics & Statistics, 84(3), 483-496.
Blau, D.M. & Tekin, E. (2007). The determinants and consequences of child care subsidies for
single mothers in the U.S.A. Journal of Population Economics, 20(4), 719-741.
Blau, F. D., Ferber, M. A., & Winkler, A. E. (2002). The Economics of Women, Men, and Work.
Upper Saddle River, NJ: Prentice-Hall.
Bromer, J., & Henly, J. R. (2009). The work-family support roles of child care providers across
settings. Early Childhood Research Quarterly, 24, 271-288.
Brooks, F. (2002). Impacts of child care subsidies on child and family wellbeing. Early
Childhood Research Quarterly, 17, 498 – 511.
Brown-Lyons, M., Robertson, A., & Layzer, J. (2001). Kith and Kin Informal Care: Highlights
from Recent Research. New York: National Center for Children in Poverty.
61

Burstein, N. & Layzer, J. (2007). National study of child care for low-income families: Patterns
of child care use among low-income families. Cambridge, MA: Abt Associates.
Campbell, F. A., Pungello, E. P., Miller-Johnson, S., Burchinal, M., & Ramey, C. T. (2001). The
development of cognitive and academic abilities: Growth curves from an early childhood
educational experiment. Developmental Psychology, 37(2), 231-242.
Capizzano, J. & Adams, G. (2003). Children in low-income families are less likely to be in
center-based child care. Snapshots of America's Families III, 16. Retrieved from
http://www.urban.org/publications/310923.html.
Capizzano, J., Adams, G., & Jason, O. (2006). The Child Care Patterns of White, Black and
Hispanic Children (Occasional papers No. 72). Washington D.C.: The Urban Institute.
Retrieved from http://www.urban.org/publications/311285.html.
Chaudry, A. (2004). Putting children first: How low-wage working mothers manage childcare.
New York: Russell Sage Foundation.
Coley, R. L., Li-Grining, C. P., Chase-Lansdale, P. L. (2006). Low-income families’ child-care
experiences: Meeting the needs of children and families. In N. Cabrera, R. Hutchens, &
H. E. Peters (Eds.), Welfare to childcare: What happens to young children when single
mothers exchange welfare for work? (pp. 149-170). Mahwah, NJ: Lawrence Erlbaum.
Connelly, R., & Kimmel, J. (2003). Marital status and full–time/part–time work status in child
care choices. Applied Economics, 35(7), 761-777.
Crosby, D. A., Gennetian, L., & Huston, A. (2005). Child care assistance policies can affect the
use of center-based care for children in low-income families. Applied Developmental
Science, 9(2), 86-106.
Douglas, T., & Flores, K. (1998). Federal and state funding of children's programs. Washington,
DC: Urban Institute.
Early, D. M., & Burchinal, M. (2001). Early childhood care: Relations with family
characteristics and preferred care characteristics. Early Childhood Research Quarterly,
16, 475-497.
Ertas, N., & Shields, S. (2012). Child care subsidies and care arrangements of low-income
parents. Children and Youth Services Review, 34(1), 179-185.
Fischer, Sheri & Martella, J. (2013). Child Care Licensing Study, 2011 [Data file and code book].
doi:10.3886/ICPSR34550.v1.
Fuller, B., Kagan, S. L., Caspary, G. L., & Gauthier, C. A. (2002). Welfare Reform and Child
Care Options for Low-Income Families. Future of Children (Packard Foundation), 12
(1), 96-119.

62

Fuller, B., Kagan, S. L., Loeb, S., & Chang, Y. W. (2004). Child care quality: Centers and home
settings that serve poor families. Early Childhood Research Quarterly, 19(4), 505-527.
Fuller, B., & Strath, A. (2001). The child-care and preschool workforce: Demographics, earnings,
and unequal distribution. Educational Evaluation and Policy Analysis, 23(1), 37-55.
Gennetian, L. A., Crosby, D. A., Huston, A. C., & Lowe, E. D. (2004). Can child care assistance
in welfare and employment programs support the employment of low-income families?
Journal of Policy Analysis and Management, 23(4), 723 – 743.
Giannarelli, L., Minton, S., Durham, C., & Department of Health and Human Services. (2012).
Child Care and Development Fund (CCDF) Policies Database, 2011. [Data file and code
book]. doi:10.3886/ICPSR34390.v1.
Gormley, W. T. (1999). Regulating child care quality. The Annals of the American Academy of
Political and Social Science, 563 (1), 116-129.
Gormley, W. T., Gayer, T., Phillips, D., & Dawson, B. (2005). The effects of universal pre-K on
cognitive development. Developmental Psychology, 41, 872–884.
Grobe, D., Weber, R. B., & Davis, E. E. (2008). Why do they leave? Child care subsidy use in
Oregon. Journal of Family and Economic Issues, 29(1), 110-127.
Ha, Y., & Ybarra, M. (2013). Are Strong Work-First Welfare Policies Aligned With Generous
Child Care Provisions? What States Are Doing and the Implications for Social Work.
Families in Society: The Journal of Contemporary Social Services, 94(1), 5-13.
Hedges, L. V., & Hedberg, E. C. (2007). Intraclass correlation values for planning grouprandomized trials in education. Educational Evaluation and Policy Analysis, 29(1), 60-87.
Helburn, S. (1995). Cost, quality and child outcomes in child care centers. Technical report.
Denver, CO: Department of Economics, Center for Research and Social Policy,
University of Colorado at Denver.
Henly, J. R., Ananat, E. O., & Danziger, S. K. (2006). Nonstandard Work Schedules, Child Care
Subsidies, and Child Care Arrangements. Chicago, IL: University of Chicago.
Hiedemann, B., Joesch, J. M. & Rose, E. (2004). More Daughters in Child Care? Child Gender
and the Use of Nonrelative Child Care Arrangements. Social Science Quarterly, 85(1),
154-168.
Hofferth, S. L., & Chaplin, D. D. (1998). State regulations and child care choice. Population
Research and Policy Review, 17(2), 111-140.
Hotz, V. J., & Xiao, M. (2011). The impact of regulations on the supply and quality of care in
child care markets. American Economic Review, 101(5), 1775-1805.

63

Huston, A., Y. E. Chang, & Gennetian, L. (2002). Family and Individual Predictors of Child
Care Use by Low-Income Income Families in Different Policy Contexts. Early
Childhood Research Quarterly, 17(4), 441-469.
Kimmel, J., & Powell, L. M. (2006). Nonstandard work and child care choices of married
mothers. Eastern Economic Journal, 32(3), 397-419.
Larsen, K., & Merlo, J. (2005). Appropriate assessment of neighborhood effects on individual
health: integrating random and fixed effects in multilevel logistic regression. American
journal of epidemiology, 161(1), 81-88.
Layzer, J. & Goodson, B. (2006). National Study of Child Care for Low-Income Families – Care
in the Home: A Description of Family Child Care and the Experience of Families and
Children That Use It. Wave 1 Report. Cambridge, MA: Abt Associates.
Levy, D. U. (2000). Child Care Policies in America: Inter State Differences and the Progress of
Devolution. Chicago, IL: University of Illinois at Chicago.
Levy, D.U. & Michel, S. (2002). More can be less: Childcare and welfare reform in the United
States. In S. Michel & R. Mahon (Eds.), Child care policy at the crossroad: Gender and
welfare state restructuring (pp.239-263). New York, NY: Routledge.
Liang, X., Fuller, B. & Singer, J.D. (2000). Ethnic differences in child care selection: The
influence of family structure, parental practices, and home language. Early Childhood
Research Quarterly, 15(3), 357-384.
Li-Grining, C. P., & Coley, R. L. (2006). Child care experiences in low-income communities:
Developmental quality and maternal views. Early Childhood Research Quarterly, 21(2),
125-141.
Liska, A. E. (1990). The significance of aggregate dependent variables and contextual
independent variables for linking macro and micro theories. Social Psychology Quarterly,
53(4), 292-301.
Loeb, S., Bridges, M., Bassok, D., Fuller, B., & Rumberger, R. W. (2007). How much is too
much? The influence of preschool centers on children's social and cognitive development.
Economics of Education Review, 26(1), 52-66.
Long, Sharon K., Kirby, G., Kurka, R., & Waters, S. (1998), Child Care Assistance under
Welfare Reform: Early Responses by the States", Assessing the New Federalism
Occasional Paper No.15. Washington D.C.: The Urban Institute.
Lowe, E. D., & Weisner, T. S. (2004). ‘You have to push it—who’s gonna raise your kids?’:
situating child care and child care subsidy use in the daily routines of lower income
families. Children and Youth Services Review. 26 (3), 143-171.
Magnuson, K. A., Ruhm, C., & Waldfogel, J. (2007). Does prekindergarten improve school
preparation and performance? Economics of Education Review, 26(1), 33-51.
64

Mettler, Suzanne. (2000). States' Rights, Women's Obligations: Contemporary Welfare Reform
in Historical Perspective. Women & Politics, 21, 1-34.
Meyers, M. K., Gornick, J. C., & Peck, L. R. (2001). Packaging Support for Low-Income
Families: Policy Variation across the United States. Journal of Policy Analysis and
Management, 20(3), 457-483.
Meyers, M. K., Heintze, T., & Wolf, D. A. (2002). Child Care Subsidies and the Employment of
Welfare Recipients. Demography, 39 (1), 165-179.
Meyers, M. K. & Jordan, L. P. (2006) Choice and Accommodation in Parental Child Care
Decisions. Journal of the Community Development Society, 37 (2), 53-70.
Michalopoulos, C., & Robins, P. K. (2002). Employment and Child-Care Choices of SingleParent Families in Canada and the United States. Journal of Population Economics, 15,
465-493.
Morris, J. R. (1999). Market constraints on child care quality. Annals of the American Academy
of Political and Social Science, 563 (1), 130-145.
Muthén, B., & Satorra, A. (1995). Complex sample data in structural equation modeling. In P.V.
Marsden (Ed.), Sociological Methodology (pp. 267-316). Washington, DC: American
Sociological Association.
National Association of Child Care Resource & Referral Agencies. (2011). We Can Do Better:
2011 Update: NACCRRA’s Ranking of State Child Care Center Regulation and
Oversight. National Association of Child Care Resource & Referral Agencies.
Office of the Assistant Secretary for Planning and Evaluation. (2012). Estimates of child care
eligibility and receipt for fiscal year 2009: ASPE Issue brief, December. Washington, DC:
US Department of Health and Human Services. Retrieved from
http://aspe.hhs.gov/hsp/12/childcareeligibility/ib.pdf
Pittard, M., Zaslow, M., Lavelle, B., and Porter, T. (2006). Investing in quality: A survey of state
Child Care and Development Fund initiatives. Washington, DC: Child Trends.
Powell, L. M. (2002). Joint Labor Supply and Childcare Choice Decisions of Married Mothers.
Journal of Human Resources, 37(1), 106-128.
Raikes, H.A., Raikes, H. H., & Wilcox, B. (2005). Regulation, subsidy receipt and provider
characteristics: What predicts quality in child care homes? Early Childhood Research
Quarterly, 20(2), 164-184.
Raudenbush, S. W. & Bryk, A. S. (2002). Hierarchical linear models: Applications and data
analysis methods. (2nd ed.). Thousand Oaks, CA: Sage Publications.
Rector, R. E. & Yousef, S.E. (1999). The Determinants of Welfare Caseload Decline: Report No.
99-04. Washington D.C.: The Heritage Center for Data Analysis, Heritage Foundation.
65

Rigby, E., Kagan, S.L., & Brooks-Gunn, J. (2004, April). Child Care Policymaking in an Era of
Devolution: Determinants of State Policy Variation. Paper presented at the Annual
Conference of the State Politics and Policy Section of the American Political Science
Association. Akron, OH.
Rigby, E., Ryan, R. M., & Brooks-Gunn, J. (2007). Child care quality in different state policy
contexts. Policy Analysis, 26(4), 887-907.
Ryan, R. M., Johnson, A., Rigby, E., & Brooks-Gunn, J. (2011). The impact of child care
subsidy use on child care quality. Early childhood research quarterly, 26(3), 320-331.
Schott, L., Pavetti, L., & Finch, I. (2012). How States Have Spent Federal and State Funds under
the TANF Block Grant. Washington, DC: Center on Budget and Policy Priorities.
Retrieved from http://www.cbpp.org/cms/?fa=view&id=3808.
Schulman, Karen, and Blank, Helen. (2011). State child care assistance policies 2011: reduced
support for families in challenging times. Washington, DC: National Women's Law
Center. Retrieved from http://www.nwlc.org/resource/state-child-care-assistancepolicies-2011-reduced-support-families-challenging-times.
Shlay, A. B. (2010). African American, White and Hispanic child care preferences: A factorial
survey analysis of welfare leavers by race and ethnicity. Social Science Research, 39(1),
125-141.
Snijders, T.A.B. & Bosker, R.J. (1999). Chapter 14: Discrete Dependent Variables. In Snijders &
Bosker, Multilevel Analysis: An introduction to basic and advanced multilevel modeling
(pp.207-238). London: Sage Publications.
Tekin, E. (2005). Child care subsidy receipt, employment, and child care choices of single
mothers. Economics Letters, 89(1), 1-6.
Tang, S., Coley, R. L., & Votruba-Drzal, E. (2012). Low-income families' selection of child care
for their young children. Children and Youth Services Review, 34 (10), 2002–2011.
Tout, K., Zaslow, M., Papillo, A. R., & Vandivere, S. (2001). Early Child Care and Education:
Work Support for Families and Developmental Opportunity for Young Children.
Assessing New Federalism Occasional Paper No. 51. Washington, DC: Urban Institute.
United States Department of Commerce. Bureau of the Census. (2012). Survey of Income and
Program Participation (SIPP) 2008 Panel Wave 8. The National Bureau of Economic
Research. Retrieved from http://www.nber.org/data/sipp.html
U.S. Department of Health and Human Services, Administration for Children and Families,
Office of Child Care. (2012a). Fiscal Year 2011 CCDF State Expenditure Data.
Retrieved from http://www.acf.hhs.gov/programs/occ/ccdf/factsheet.htm.

66

U.S. Department of Health and Human Services, Administration for Children and Families,
Office of Child Care. (2012b). Fiscal Year 2011 TANF Financial Data. Retrieved from
http://www.acf.hhs.gov/programs/ofa/resource/tanf-financial-data-fy-2011.
U.S. Department of Health and Human Services, Administration for Children and Families,
Office of Child Care. (2013). Fiscal Year 2011 CCDF Data Tables (preliminary).
Retrieved from
http://www.acf.hhs.gov/sites/default/files/occ/final_tab_b_attachment_i_xls.xls
Vandell, D. L. (2004). Early child care: The known and unknown. Merill-Palmer Quarterly, 50,
387-414.
Vandell, D. L., Belsky, J., Burchinal, M., Steinberg, L., & Vandergrift, N. (2010). Do effects of
early child care extend to age 15 years? Results from the NICHD study of early child care
and youth development. Child development, 81(3), 737-756.
Vandell, D. L. & Wolfe, B. (2000). Child care quality: Does it matter and does it need to be
improved? Washington, D.C.: Office of the Assistant Secretary for Planning and
Evaluation: U.S. Department of Health and Human Services.
Weinraub, M., Shlay, A. B., Harmon, M., & Tran, H. (2005). Subsidizing child care: How child
care subsidies affect the child care used by low-income African American families. Early
Childhood Research Quarterly, 20, 373 – 392.
Westat and Mathematica Policy Research (2001). Survey of Income and Program Participation
Users’ Guide. Report to U.S. Department of Commerce Economics and Statistics
Administration, U.S. Bureau of the Census.
Wolfe, B. (2002). Incentives, challenges, and dilemmas of TANF: A case study. Journal of
Policy Analysis and Management, 21(4), 577-586.

67