View in EDS HTML Full Text PDF Full Text

The Role of State Subsidy Policies in Early Education Programs' Decisions to Accept Subsidies: Evidence from Nationally Representative Data

Saved in:
Bibliographic Details
Title: The Role of State Subsidy Policies in Early Education Programs' Decisions to Accept Subsidies: Evidence from Nationally Representative Data
Language: English
Authors: Gerilyn Slicker (ORCID 0000-0002-0826-6394), Christina Areizaga Barbieri, Jason T. Hustedt
Source: Early Education and Development. 2024 35(4):859-877.
Availability: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
Peer Reviewed: Y
Page Count: 19
Publication Date: 2024
Sponsoring Agency: Administration on Children, Youth, and Families (ACYF) (DHHS/ACF)
Contract Number: 90YE0248
Document Type: Journal Articles
Reports - Research
Education Level: Early Childhood Education
Descriptors: State Aid, Grants, Financial Policy, Educational Policy, Early Childhood Education, Family Needs, Low Income, Poverty, Access to Education, Child Care Centers, Child Development Centers, School Funds, Educational Finance, Financial Support, Programs, Participation, Income
DOI: 10.1080/10409289.2023.2244859
ISSN: 1040-9289
1556-6935
Abstract: Research Findings: Families need access to early care and education programs, both to ensure parents' ability to work and support children's development. Subsidies through the Child Care and Development Fund (CCDF) are a set of policies aimed at assisting families living in poverty with accessing early education. However, the number of early education centers that accept subsidies is declining. Using observational data from the National Survey of Early Care and Education and state subsidy policies from the CCDF Policies Database, we use an innovative application of propensity score methods to estimate causality and provide actionable findings regarding the effect of state-specific policies on centers' subsidy participation. We create comparable groups of centers that accept subsidies and centers that do not, based on research-supported program- and community-level predictors of subsidy participation. Logistic regression models using the matched sample demonstrated state-specific subsidy policies impacted subsidy participation. Specifically, as the state subsidy reimbursement rate increased, centers were more likely to accept subsidies. Practice or Policy: Findings point to meaningful state-level actions and policies that may incentivize centers' subsidy participation. In general, state policies that increase revenue for centers accepting subsidies (e.g. reimbursement for child absences) may result in higher rates of subsidy participation.
Abstractor: As Provided
Entry Date: 2024
Accession Number: EJ1420637
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwHvBy3Kzw_6-LzZ_C0IIfDnAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDPKJc8RaiE1pWt1YAgIBEICBm3iq_g1It4DFp3nUa7Lv6sUxuiXyh3uMAfYGkrH_bafJPdXqUmplDaW_x9vqK91HsSgDQBlzQfpbpjQ-0tlSa0VJiEFxbAP-EmlKWiKtXJIQWOXD0-aPQB9nZe0tftqhJoxhDvy3tA0kCKKZmXXeJ8N_9nX41fzuEL5HKzzq-JhwxSnOx2vh2JeTAEtgDu38_BL9osPbX8kVTbqE
Text:
  Availability: 1
  Value: <anid>AN0176533093;h4j01may.24;2024Apr15.04:43;v2.2.500</anid> <title id="AN0176533093-1">The Role of State Subsidy Policies in Early Education Programs' Decisions to Accept Subsidies: Evidence from Nationally Representative Data </title> <p>Research Findings: Families need access to early care and education programs, both to ensure parents' ability to work and support children's development. Subsidies through the Child Care and Development Fund (CCDF) are a set of policies aimed at assisting families living in poverty with accessing early education. However, the number of early education centers that accept subsidies is declining. Using observational data from the National Survey of Early Care and Education and state subsidy policies from the CCDF Policies Database, we use an innovative application of propensity score methods to estimate causality and provide actionable findings regarding the effect of state-specific policies on centers' subsidy participation. We create comparable groups of centers that accept subsidies and centers that do not, based on research-supported program- and community-level predictors of subsidy participation. Logistic regression models using the matched sample demonstrated state-specific subsidy policies impacted subsidy participation. Specifically, as the state subsidy reimbursement rate increased, centers were more likely to accept subsidies. Practice or Policy: Findings point to meaningful state-level actions and policies that may incentivize centers' subsidy participation. In general, state policies that increase revenue for centers accepting subsidies (e.g. reimbursement for child absences) may result in higher rates of subsidy participation.</p> <p>Families need access to affordable early care and education (ECE) programs for their children, both to ensure parents' ability to work and enhance children's readiness for school. However, affordable ECE is very difficult to locate, particularly for families from low-income backgrounds (Baldiga et al., [<reflink idref="bib8" id="ref1">8</reflink>]) resulting in income-based gaps in ECE enrollment (Magnuson & Waldfogel, [<reflink idref="bib37" id="ref2">37</reflink>]). The Child Care and Development Fund (CCDF) is a federal government program that provides child care subsidies to families living in poverty to offset the high cost of ECE. Subsidies are government-issued funds that families can use to access ECE at a participating program of their choice. Subsidies that allow children from disadvantaged backgrounds to attend high-quality early education are beneficial for children's development (Blau, [<reflink idref="bib11" id="ref3">11</reflink>]). While subsidies have the potential to increase ECE access for eligible families, fewer than one in six eligible children have access to subsidies (Child Care Aware of America, [<reflink idref="bib15" id="ref4">15</reflink>]). One factor influencing children's access to ECE is the persistent decline of ECE programs willing to accept subsidies (U.S. Department of Health and Human Services [DHHS], [<reflink idref="bib57" id="ref5">57</reflink>]).</p> <p>The sustainability of CCDF and children's access to affordable ECE is threatened by insufficient numbers of ECE programs accepting subsidies, yet little is known about the specific motivating factors and state policy contexts that may influence ECE programs' subsidy participation. Due to limited government funds, ECE programs often lose revenue by accepting subsidies because state subsidy reimbursement rates typically fall far below the prices that private-paying families are charged (Schulman, [<reflink idref="bib51" id="ref6">51</reflink>]). Researchers and policymakers speculate that state- specific subsidy policies – such as the subsidy reimbursement rate – may influence ECE program subsidy participation; however, there is an absence of causal evidence that provides policymakers with guidance about how to incentivize subsidy participation. In this study, we relied on a database of state-specific CCDF policies (CCDF Policies Database, 2011) and a nationally representative sample of ECE centers (National Survey of Early Care and Education Project Team [NSECE], [<reflink idref="bib42" id="ref7">42</reflink>]) to examine the predictive utility of state subsidy policies on ECE programs' subsidy participation. Using a novel application of propensity score methods (PSM), we used a set of research-supported center and community covariates to match ECE centers, creating a "treatment group" (i.e., accepts subsidies) and a "control group" (i.e., does not accept subsidies). Following the matching process, we determined if variation in subsidy policies across states shapes ECE centers' subsidy participation.</p> <hd id="AN0176533093-2">Background</hd> <p></p> <hd id="AN0176533093-3">The Child Care and Development Fund (CCDF)</hd> <p>The Child Care and Development Fund (CCDF), originally authorized in 1990, is a federal program with a goal of improving access to ECE for families from low-income backgrounds with children under the age of 13. At its inception, CCDF was established to primarily allow families to work, but over time and through the Child Care and Development Block Grant (CCDBG) Act of 2014—which reauthorized the CCDF program – CCDF has evolved to simultaneously prioritize the healthy development of children. As a result, CCDF now addresses both ECE access-related concerns of policymakers. CCDF allocates funds to states for disbursement. States use the funds to provide child care subsidies through vouchers to low-income families or contracts with ECE center-based or home-based programs. CCDF aims to provide families access to ECE with minimal costs. Though the majority of subsidies are used in ECE centers (U.S. Department of Health and Human Services, Office of Child Care [US DHHS OCC], [<reflink idref="bib57" id="ref8">57</reflink>]), there is evidence to suggest that families living in rural areas that use subsidies are more likely to attend home-based programs than in urban areas (De Marco & Vernon-Feagans, [<reflink idref="bib20" id="ref9">20</reflink>]). In fiscal year 2019, CCDF served approximately 1.4 million children (US DHHS OCC, [<reflink idref="bib57" id="ref10">57</reflink>]). Though CCDF is one of the largest federal ECE investments, the program only served 11% of eligible children each month in 2011 and 2012 (U.S. Government Accountability Office [GAO], [<reflink idref="bib58" id="ref11">58</reflink>]), when the data for this study were collected.</p> <hd id="AN0176533093-4">State Subsidy Policies</hd> <p>States set their own CCDF policies and practices – though they must fall within specified federal parameters – permitting variation in implementation of CCDF across states. State subsidy agencies establish a series of policies ranging from those that determine what families qualify for subsidies, the amount that parents must contribute toward ECE costs (i.e., copayments), and the amount the programs are paid for serving children using subsidies (i.e., the subsidy reimbursement rate). These state policies simultaneously affect families' ECE access and programs' ability to provide quality programming. Research suggests that variation in state subsidy policies can impact families' access to ECE (e.g., Davis et al., [<reflink idref="bib19" id="ref12">19</reflink>]; Weber et al., [<reflink idref="bib60" id="ref13">60</reflink>]), but has yet to establish the impact of state-specific subsidy policies on program participation in the subsidy system (i.e., if the center accepts subsidies).</p> <hd id="AN0176533093-5">Causal Research on Subsidies</hd> <p>A well-documented causal relationship has been established between a set of CCDF policies and continuity of subsidy use for families (Ha et al., [<reflink idref="bib29" id="ref14">29</reflink>]; Jenkins & Nguyen, [<reflink idref="bib32" id="ref15">32</reflink>]; Michalopoulos et al., [<reflink idref="bib39" id="ref16">39</reflink>]; Weber et al., [<reflink idref="bib60" id="ref17">60</reflink>]). Studies employing experimental designs in single states show that more generous policies (e.g., more generous eligibility policies, fewer out-of-pocket costs for families, improvement in bureaucratic processes) result in parents having access to more stable care arrangements and higher rates of satisfaction (Michalopoulos et al., [<reflink idref="bib39" id="ref18">39</reflink>]). Recently, Jenkins and Nguyen ([<reflink idref="bib32" id="ref19">32</reflink>]) provided causal evidence that lengthening the eligibility redetermination period for families increased children's continuous enrollment in ECE supported with subsidy funds, in a study using data from 38 states over a 10-year period. Another experimental study found that families who had access to a set of policies that were family-friendly – including access to caseworkers who provided information on ECE settings and fewer out-of-pocket expenses for families – more commonly selected center-based ECE as compared with other ECE options (Crosby et al., [<reflink idref="bib17" id="ref20">17</reflink>]). These studies suggest that variations in state- specific subsidy policies have the potential to impact subsidy participation for eligible families.</p> <hd id="AN0176533093-6">State Subsidy Policies Impacting ECE Programs</hd> <p>There are also CCDF policies that directly impact ECE programs. For instance, each state determines its provider reimbursement rates for serving children using subsidies. State subsidy reimbursement rates are highly variable across and, in some cases within, states. The rates vary based on a series of factors including the geographical location, the type of program (i.e., center, home-based), the age group of the child (i.e., infant, toddler, preschool), the amount of care (i.e., full-time, part-time) or the quality of the program. In accordance with guidelines from the 2014 reauthorization of CCDF, states must establish these rates by conducting market rate surveys of the price of ECE in the state (or another approved method). The reauthorized CCDF law recommends that states set their reimbursement rates at 75% of the market rate, as this would be giving families using subsidies the opportunity to access three out of four programs without additional expenses beyond any family copayments (Office of the Inspector General [OIG], [<reflink idref="bib44" id="ref21">44</reflink>]). Yet, in 2019, only four states had reimbursement rates that met that guideline (Schulman, [<reflink idref="bib51" id="ref22">51</reflink>]). In a survey of state CCDF administrators, respondents shared that states prioritized serving more children over higher provider reimbursement rates. If the state set the rate at 75% of the market rate, the reimbursement costs would exhaust subsidy funds sooner, leaving fewer funds available for enhancing families' ECE access (OIG, [<reflink idref="bib44" id="ref23">44</reflink>]). At the same time, states must consider how the established reimbursement rate impacts ECE programs. Specifically, insufficient rates are likely to discourage provider subsidy participation (Schulman, [<reflink idref="bib51" id="ref24">51</reflink>]), but more research is needed to understand the specific role of subsidy policies.</p> <p>States also have the option to set maximum amounts that could be paid to ECE programs that accept subsidies. States may provide higher reimbursement rates for providers that qualify for increased payment for meeting additional criteria. A common practice, for instance, is to use a tiered reimbursement rating system that provides higher reimbursement rates for providers that achieve higher quality ratings in the state Quality Rating and Improvement System (QRIS). QRISs are accountability systems that give parents information about ECE program quality and have been shown to encourage program improvement through reputational and financial incentives (Bassok et al., [<reflink idref="bib9" id="ref25">9</reflink>]; Gomez et al., [<reflink idref="bib25" id="ref26">25</reflink>]). In some cases, states have a single higher payment rate while others have progressively higher payment rates for progressively higher levels of quality. Thirty-two states use a tiered reimbursement system (Dwyer et al., [<reflink idref="bib23" id="ref27">23</reflink>]). In states with tiered rates, the difference in tiers for a 4-year-old range from 5–117%. Importantly, even in the more than two-thirds of states with tiered reimbursement rates, the highest rates were still below the 75th percentile of market rates (Schulman, [<reflink idref="bib51" id="ref28">51</reflink>]).</p> <p>Other variations in state CCDF policies include the number of days providers are reimbursed for child absences (if any) and the number of days providers are reimbursed for closures for holidays, professional development, or other reasons (if any). In 2011 (when the data for the current study were collected), 25 states had a policy that provided payments for child absences and 24 states had a policy to provide payments for program closures (Minton et al., [<reflink idref="bib40" id="ref29">40</reflink>]). More recently, CCDF administrator surveys indicated that many states implemented provider-friendly payment practices – such as paying providers for child absences (41 states) – however, providers continued to cite unreliable and insufficient payments as challenges or barriers to subsidy participation (Halle et al., [<reflink idref="bib30" id="ref30">30</reflink>]; OIG, [<reflink idref="bib44" id="ref31">44</reflink>]).</p> <p>Further variation in state administration of subsidies includes how states disburse funds. Most states use vouchers, which families can use at any participating ECE program (Minton et al., [<reflink idref="bib40" id="ref32">40</reflink>]). Contracts – which are direct payments to ECE programs that are contractually obligated to the subsidy system – are used less frequently. However, research suggests contracts may enhance administrative support and make revenue more reliable for participating ECE programs (Adams & Rohacek, [<reflink idref="bib3" id="ref33">3</reflink>]). Given the potential for more revenue stability, the use of state subsidy contracts (vs. vouchers) may encourage provider subsidy participation.</p> <p>Since research has established a causal relationship between subsidy policies and families' ECE access, it is also possible that variation in subsidy policies impacting ECE programs – including the subsidy reimbursement rate, tiered reimbursement, payment for absences/closures, or the use of subsidy contracts – may also influence provider subsidy participation. Focusing on the implementation of subsidy policies is critical because they are amenable to policy intervention. While many factors may influence subsidy participation, features of state CCDF policies are within the direct control of state-level policymakers.</p> <hd id="AN0176533093-7">Provider Participation in the Subsidy System</hd> <p>Given the steady decline in the number of ECE programs that accept subsidies (US DHHS OCC, [<reflink idref="bib57" id="ref34">57</reflink>]), it is a federal priority to incentivize subsidy participation (Child Care and Development Fund Program [CCDF], [<reflink idref="bib14" id="ref35">14</reflink>]). As part of the reauthorization of CCDF, states are required to track provider subsidy participation as well as barriers to participation (CCDF, [<reflink idref="bib14" id="ref36">14</reflink>]). Research highlights substantial challenges for ECE programs that participate, including burdensome administrative requirements, communication problems between subsidy agencies and providers, and late or incorrect payment rates (Adams et al., [<reflink idref="bib4" id="ref37">4</reflink>]; Rohacek & Adams, [<reflink idref="bib47" id="ref38">47</reflink>]; Sandstrom et al., [<reflink idref="bib50" id="ref39">50</reflink>]). Of particular concern is that providers may lose revenue by accepting subsidies since most states offer provider reimbursement rates that are significantly lower than the market rate (Schulman, [<reflink idref="bib51" id="ref40">51</reflink>]). Not much is known about what motivates provider subsidy participation (Rohacek, [<reflink idref="bib46" id="ref41">46</reflink>]), though the evidence is growing that certain programmatic and community characteristics may play a role (Giapponi Schneider et al., [<reflink idref="bib24" id="ref42">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref43">53</reflink>]; Slicker et al., [<reflink idref="bib54" id="ref44">54</reflink>]). An adapted version of the Conceptual Framework of Childcare Provider Subsidy Participation (Giapponi Schneider et al., [<reflink idref="bib24" id="ref45">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref46">53</reflink>]) guided this study. This conceptual framework was originally developed and tested using Massachusetts administrative data for licensed ECE providers (Giapponi Schneider et al., [<reflink idref="bib24" id="ref47">24</reflink>]). The conceptual model considers several categories of possible predictors of subsidy participation for ECE providers, including provider factors, local market factors, and subsidy policies/practices. In a recent study, some refinements were made to the model using publicly-available nationally representative data (Slicker & Hustedt, [<reflink idref="bib53" id="ref48">53</reflink>]). Specifically, operational features of an ECE program – such as the administrative/operational capacity, center finances, and child enrollment – were associated with subsidy participation (Giapponi Schneider et al., [<reflink idref="bib24" id="ref49">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref50">53</reflink>]). Characteristics of the community – such as the poverty level of the surrounding area – were associated with subsidy participation. However, neither study was able to empirically test the impact of state-specific subsidy policies, such as the provider subsidy reimbursement rate. In the present study, the adapted conceptual framework (Figure 1) guided the selection of independent variables used in the initial logistic regression that guided the propensity score matching process.</p> <p>Graph: Figure 1. Conceptual framework: Program, community, and state influences on subsidy system participation.</p> <hd id="AN0176533093-8">The Present Study</hd> <p>We investigated the unique influence of state-specific subsidy policies (e.g., subsidy reimbursement rate, the presence of a tiered reimbursement policy) on provider subsidy participation using a nationally representative sample of ECE centers from the restricted-use 2012 NSECE and matching their location with state-specific subsidy policies in the 2011 CCDF Policies Database. Our study improved upon previous methods by using propensity score matching, allowing us to closely approximate a randomized experiment to estimate the predictive utility of state subsidy policies on provider subsidy participation. This national study focused specifically on state subsidy policies, making it unique from previous research that highlighted a series of center and community level predictors of subsidy participation. Though important, these center and community factors are not always (immediately) amenable to policy intervention. On the other hand, this study used a set of state-level subsidy policies directly impacting ECE centers and evaluated whether they predict provider subsidy participation. Our research question is: To what extent do state subsidy policies yield subsidy participation for ECE centers?</p> <hd id="AN0176533093-9">Method</hd> <p></p> <hd id="AN0176533093-10">Data Sources</hd> <p>Data were drawn from the CCDF Policies Database and the restricted-use National Survey of Early Care and Education (NSECE). Data on ECE centers, including information about subsidy system participation as well as a series of provider and community characteristics came from the NSECE. State subsidy policies came from the CCDF Policies Database.</p> <hd id="AN0176533093-11">NSECE</hd> <p>The NSECE is a survey that is sponsored by the Office of Planning, Research, and Evaluation (OPRE) in the Administration for Children and Families (ACF) in the U.S. DHHS. The data from this study were drawn from the 2012 center-based provider survey, which was conducted with ECE program directors licensed in the state in which they were located. Center-based providers were also drawn from all available state and national administrative lists, including license-exempt programs (e.g., Office of Head Start list of programs, YMCA programs, National Association for the Education of Young Children accredited programs, etc.) from various agencies in all 50 states and D.C. NSECE data collection took place from November 2011 through June 2012 (NSECE Project Team, [<reflink idref="bib41" id="ref51">41</reflink>]). In total, 8,265 center-based providers completed the survey and were asked a series of questions related to the program's enrollment and rates, workforce, and participation in government programs. The analytic sample for this study included only providers that enrolled at least one child five years of age or younger (and not yet enrolled in kindergarten, <emph>n</emph> = 7,771). In order to gain access to the state-level identifier, we relied on the restricted-level NSECE data. While we acknowledge that the 2019 NSECE is now available, not all of the variables we needed for the matching process (see below) are available in the more recent 2019 data. Since the validity of the propensity score estimate in our PSM design is dependent upon the inclusion of all relevant predictors, we used the 2012 data that allowed us to include all of the research-supported predictors of subsidy participation.</p> <hd id="AN0176533093-12">CCDF Policies Database</hd> <p>The CCDF Policies Database (Office of Planning, Research, & Evaluation [OPRE], [<reflink idref="bib43" id="ref52">43</reflink>]) is a separate database of child care subsidy policies for all 50 U.S. states and DC. The policies captured in the database are collected primarily through the states' caseworker policy manuals and then verified by the state administrators for accuracy. The database includes policies for family eligibility and payments, provider requirements, and reimbursement rates. We examined policies directly impacting center-based ECE providers that were in place as of October 1, 2011 to match the timeframe when providers were surveyed for the NSECE.</p> <hd id="AN0176533093-13">Measures</hd> <p>Most of the variables used in this study were from the NSECE except for the state subsidy policies, which were from the CCDF Policies Database. To merge the two datasets, we use a state-level identifier from the restricted-use NSECE dataset.</p> <hd id="AN0176533093-14">Subsidy System Participation</hd> <p>The dependent variable, subsidy system participation, was from a question in the NSECE that asked providers to report the percentage of children who were funded by "child care subsidy programs such as CCDF or TANF (including vouchers/certificates, state contracts)." Subsidy system participation is a dichotomous variable that captured whether the provider serves at least one child funded by subsidies. This item served as our outcome variable.</p> <hd id="AN0176533093-15">CCDF Policies</hd> <p>We captured state-specific variation in subsidy policies using a series of variables from the CCDF Policies Database. These variables served as the independent variables for the logistic regression following the propensity score matching process. The base subsidy reimbursement rate was the rate that centers received for accepting subsidies. Consistent with previous research (Greenberg et al., [<reflink idref="bib28" id="ref53">28</reflink>]; Madill et al., [<reflink idref="bib36" id="ref54">36</reflink>]), we adjusted the base reimbursement rate provided in the CCDF policies database using Regional Price Parities, developed by the Bureau of Economic Analysis, to capture state-to-state differences in cost of living. We used a z-score of this rate, where z = 0.0 is equal to the mean base rate, to improve interpretability. Since some states have a higher reimbursement rate for certain ECE centers, typically based on the program earning a higher quality rating within the state's QRIS, we included a variable capturing whether the state has a tiered reimbursement policy. We also included a variable capturing if the family was required to pay any difference between the ECE center's parent-pay rate and the state's reimbursement rate. To capture variation in how states administer subsidies, we used a variable indicating if the ECE program received subsidy funds through state contracts (vs. vouchers).</p> <p>Finally, we included a few variables directly impacting centers. The CCDF Policies Database included data about whether centers were paid for the days children are absent and whether centers were paid for the days the center is closed (for professional development, holidays, inclement weather, or other approved reasons). We used these variables as predictors.</p> <hd id="AN0176533093-16">Covariates in Propensity Score Matching</hd> <p>Our goal for propensity score matching (PSM) was to create two groups of centers that varied in whether they accepted child care subsidies but that were equal on a range of covariates – outside of state subsidy policies – that are likely to predict subsidy participation. Then, we used these equated groups as outcomes in a logistic regression that simultaneously assessed the predictive utility of a range of state subsidy policies. We use three categories of covariates from the NSECE in the propensity score matching to estimate ECE centers' propensity to accept child care subsidies. Each of the covariates included for the propensity score matching process have previously been identified as predictors of subsidy system participation (Giapponi Schneider et al., [<reflink idref="bib24" id="ref55">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref56">53</reflink>]). The first category of independent variables was provider factors, which included the center's legal status, total child enrollment, if the program enrolled children under age 3, and if the program had a quality rating. In the center-based provider survey, respondents reported if the program was for-profit or nonprofit. Previous research has identified for-profit status as a predictor of subsidy system participation (Giapponi Schneider et al., [<reflink idref="bib24" id="ref57">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref58">53</reflink>]). Respondents also reported the number of children enrolled in all age groups served and these numbers were combined to create a variable capturing the total number of children. In previous research, as the number of children enrolled increased, subsidy system participation became less likely (Giapponi Schneider et al., [<reflink idref="bib24" id="ref59">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref60">53</reflink>]). Using responses about the number of children by age group, the NSECE team also created a variable capturing whether the program enrolled any children under the age of 3 (infants and toddlers). In another study relying on NSECE data, programs that enrolled children under the age of 3 had higher odds of subsidy system participation (Slicker & Hustedt, [<reflink idref="bib53" id="ref61">53</reflink>]). Respondents also reported whether their program had an overall quality rating. Having a quality rating/accreditation is positively associated with subsidy system participation (Giapponi Schneider et al., [<reflink idref="bib24" id="ref62">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref63">53</reflink>]).</p> <p>We included several community characteristics as covariates. To capture the urban density of the program's surrounding community, the NSECE team matched the program's address to the 2010 U.S. Census data. Communities with 0.85 or above in the ratio of urban to total population were classified as having a high urban density. Programs that operated in areas with a ratio of 0.84 or below were classified as having a lower urban density. We included this variable as a covariate given that families in urban areas use subsidies at higher rates than rural areas (Davis et al., [<reflink idref="bib18" id="ref64">18</reflink>]). The NSECE team also created a community poverty density variable whereby programs operating in communities with greater than 20% of the population living below the Federal Poverty Level (FPL) were considered to have a high density of community poverty. Programs operating in a community with 20% or fewer of the population living below the FPL were classified as having a lower density of community poverty. Previous research suggests that as the median income of an area increased, subsidy system participation decreased (Giapponi Schneider et al., [<reflink idref="bib24" id="ref65">24</reflink>]) and, similarly, that when a program operated in an area of high poverty density, odds of subsidy system participation were higher (Slicker & Hustedt, [<reflink idref="bib53" id="ref66">53</reflink>]).</p> <p>Finally, we included public and private sources of funding as covariates for the propensity score matching. Centers were asked to report the sources of funding received as well as the number of children funded by a list of publicly funded sources. Centers reported whether they received any revenues from tuition or fees paid by parents. Receipt of parent payment has been shown to be positively associated with subsidy participation (Slicker & Hustedt, [<reflink idref="bib53" id="ref67">53</reflink>]). The NSECE Team identified a center as a Head Start center if directors reported receiving any Head Start funds. Research suggests that receipt of Head Start funds was negatively associated with subsidy density (Slicker, [<reflink idref="bib52" id="ref68">52</reflink>]). Similarly, if centers reported receiving any public pre-K funds, the centers were classified as a pre-K center. Previous research suggests that ECE centers that report receipt of public pre-K funds had lower odds of subsidy system participation (Slicker & Hustedt, [<reflink idref="bib53" id="ref69">53</reflink>]).</p> <hd id="AN0176533093-17">Analytic Approach</hd> <p>The goal of this study was to determine the impacts of state CCDF policies on provider subsidy system participation (i.e., whether the ECE center accepts child care subsidies). However, given the nature of the CCDF program and the variation in state policies, this cannot be tested experimentally. That said, to answer our research question, we applied propensity score methods (PSM) to a nationally representative sample to develop groups of providers that either did (we refer to them as "subsidy" centers throughout the paper) or did not ("non-subsidy" centers) accept subsidies. These groups were comparable on all key predictors of subsidy participation <emph>except</emph> in their state's related policies and practices. PSM involves generating probabilities that an individual will be in a "treatment" or "control" condition based on a collection of selected covariates and will be used to test predictors of state subsidy policies on provider subsidy system participation. We used PSM in a novel way because we estimated comparable "control" and "treatment" groups as our <emph>outcome</emph> and then estimated the predictive utility of specific variables on being in one of the two groups. Typically, studies estimate the "treatment" and "control" groups with PSM and then compare those groups on outcome(s). However, we were interested in simultaneously examining a range of policies that are not mutually exclusive (i.e., that cannot be split up into simple treatment vs. control groups as in traditional PSM methods) and that likely impact subsidy participation. Thus, we created comparable outcome groups that are equated in all crucial predictors of subsidy participation outside of state policies before estimating the impacts of these policies using logistic regression. We leveraged propensity score methods, which use observational data to estimate causality, to appropriately consider or account for the impact of center- and community-level predictors of provider subsidy system participation. However, we present results that estimated the influence of state subsidy policies, and thus should be especially relevant to policymakers making decisions at the state level.</p> <p>PSM can provide causal estimates because the covariates that predict receiving the treatment were carefully selected based on previous research and used to capture and reduce selection bias. A propensity score is the probability that an ECE center would be in a particular group (i.e., participate in the subsidy system) based on a set of covariates (Rosenbaum & Rubin, [<reflink idref="bib48" id="ref70">48</reflink>]). PSM reduces selection bias by balancing covariate distributions between the groups to be compared (Leite, [<reflink idref="bib35" id="ref71">35</reflink>]) and has become a common choice for equating groups with non-experimental data when random assignment is not possible (Bai & Clark, [<reflink idref="bib7" id="ref72">7</reflink>]; Thoemmes & Kim, [<reflink idref="bib56" id="ref73">56</reflink>]). In our study, what would traditionally be considered the "treatment" group consisted of the subsidy centers (<emph>n</emph> = 2,640), while the "control" group is comprised of the non-subsidy centers (<emph>n</emph> = 5,131). The goal of PSM is to achieve an appropriate counterfactual; in our case, the appropriate counterfactual was ECE centers that had a high likelihood of accepting subsidies (and thus being in the treatment group), but that did not accept subsidies. However, it is important to note that we are not estimating the effect of accepting subsidies on an outcome. Thus, we do not refer to the subsidy group as the treatment group beyond the description of the PSM procedure. Rather, by carrying out this process, any differences between groups on the dependent variables (i.e., subsidy system participation) could be more confidently attributed to variation in state CCDF policies in the logistic regression carried out after the matching process. Specifically, using a multistep process that involved multiple imputation (MI) and PSM (described below), we divided the analytic sample across subgroups designated on the basis of subsidy system participation and matched based on the probability that centers accept subsidies to then determine state CCDF policies' predictive utility on subsidy system participation.</p> <p>To carry out this study, we took the following steps. First, we used MI to address missing data. Second, we applied propensity score matching to find comparable samples of subsidy and non-subsidy centers using a series of research-supported center, community, and funding source covariates. Then, we used logistic regressions to calculate odds ratios to estimate the likelihood of subsidy participation for each of the state-specific subsidy policies included in the model.</p> <hd id="AN0176533093-18">Missing Data</hd> <p>We used MI to handle missing data since the data were assumed to be missing at random (MAR; Acock & Long, [<reflink idref="bib1" id="ref74">1</reflink>]). Since provider interviews were likely missing due to logistical constraints and not due to providers' potential responses, data were assumed to be MAR since the missingness may depend on the values of observed measures but not on unobserved measures (Graham, [<reflink idref="bib26" id="ref75">26</reflink>]). MI is the preferred method for handling missing data in propensity score methods (Leite, [<reflink idref="bib35" id="ref76">35</reflink>]).</p> <p>We relied on the <emph>mice</emph> package in R (van Buuren & Groothuis-Oudshoorn, [<reflink idref="bib59" id="ref77">59</reflink>]), which uses chained equation to produce imputed datasets. Missing data on the predictors from the NSECE dataset ranged from 0 to 4% with the total number of children enrolled in the program containing the most missing values. Though most states had data available for each of the six CCDF policies of interest, there was also missing data in the CCDF Policies Database, ranging from 0% to 37% for policy variables. Aside from the variable about reimbursement for center closures which had having the most missing values (37%), only one other policy variable had missing values, which was a variable capturing whether the provider charged parents a copayment (15%). We imputed 20 datasets, following recommendations by Graham et al. ([<reflink idref="bib27" id="ref78">27</reflink>]) based on missingness rates.</p> <hd id="AN0176533093-19">Propensity Score Matching</hd> <p>We used propensity score matching to identify an equivalent group of subsidy and non-subsidy centers for each of the 20 imputed datasets. We carried out the matching process with the MatchThem package (Pishgar et al., [<reflink idref="bib45" id="ref79">45</reflink>]) in R. Logistic regression determined the probability for subsidy system participation based on the predictors identified in the conceptual model guiding this study (Giapponi Schneider et al., [<reflink idref="bib24" id="ref80">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref81">53</reflink>]), creating the propensity score based on multiple covariates (Bai & Clark, [<reflink idref="bib7" id="ref82">7</reflink>]). Next, we used propensity score matching – including nearest neighbor and caliper matching – to identify two groups: subsidy system participation/accepts subsidies and subsidy system non-participation/does not accept subsidies based on the proximity of their propensity scores (Caliendo & Kopeinig, [<reflink idref="bib13" id="ref83">13</reflink>]; Stuart, [<reflink idref="bib55" id="ref84">55</reflink>]). The nearest neighbor matching method involves choosing an individual in the treatment group and matching its propensity score with an individual in the comparison group that has its closest propensity score (West et al., [<reflink idref="bib61" id="ref85">61</reflink>]). To avoid a common problem with nearest neighbor matching whereby poor matches can occur because there is little overlap in the distribution of propensity scores in the two groups, we also used caliper matching. A caliper, or maximum distance between propensity scores in the two groups that is permitted for a match to be made, was specified following the recommendations of Cochran and Rubin ([<reflink idref="bib16" id="ref86">16</reflink>]). We used a caliper of 0.2 standard deviations, per recommendations for ideal caliper distance (Stuart, [<reflink idref="bib55" id="ref87">55</reflink>]), to determine the closest match. We also tested a final method of matching, optimal matching. Optimal matching matches treated individuals with untreated individuals by minimizing the total distance between the treated and untreated matched pairs for the entire sample, minimizing the reduction in sample size (Leite, [<reflink idref="bib35" id="ref88">35</reflink>]). Typically, optimal matching provides better matches than the other two methods (Bai & Clark, [<reflink idref="bib7" id="ref89">7</reflink>]), but given the large sample size available using NSECE data, results from optimal matching and nearest neighbor matching were similar. We tested each matching method described above to determine the method with the greatest bias reduction. We used the following formula:</p> <p>Graph</p> <p> <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mi mathvariant="italic">PB</mi><mrow><msub><mi>R</mi><mrow><mi mathvariant="italic">k</mi><mspace width="thickmathspace" /></mrow></msub></mrow><mspace width="thickmathspace" /><mo>=</mo><mspace width="thickmathspace" /><mrow><mfrac><mrow><mrow><mfenced open="|" close="|"><mrow><mspace width="thickmathspace" /><mrow><mrow><msub><mi mathvariant="italic">B</mi><mrow><mi mathvariant="italic">k</mi><mo>,</mo><mi mathvariant="italic">before</mi><mspace width="thickmathspace" /><mi mathvariant="italic">matching</mi></mrow></msub></mrow></mrow><mspace width="thickmathspace" /></mrow></mfenced><mspace width="thickmathspace" /><mo>−</mo><mspace width="thickmathspace" /><mfenced open="|" close="|"><mrow><mspace width="thickmathspace" /><mrow><mrow><msub><mi>B</mi><mrow><mi mathvariant="italic">k</mi><mo>,</mo><mi mathvariant="italic">after</mi><mspace width="thickmathspace" /><mi mathvariant="italic">matching</mi></mrow></msub></mrow></mrow><mspace width="thickmathspace" /></mrow></mfenced></mrow><mspace width="thickmathspace" /></mrow><mrow><mspace width="thickmathspace" /><mrow><mfenced open="|" close="|"><mrow><mspace width="thickmathspace" /><mrow><mrow><msub><mi>B</mi><mrow><mi mathvariant="italic">k</mi><mo>,</mo><mi mathvariant="italic">before</mi><mspace width="thickmathspace" /><mi mathvariant="italic">matching</mi></mrow></msub></mrow></mrow><mspace width="thickmathspace" /></mrow></mfenced></mrow></mrow></mfrac></mrow><mspace width="thickmathspace" /><mi mathvariant="italic">x</mi><mspace width="thickmathspace" /><mn>100</mn><mi mathvariant="normal">%</mi><mspace width="thickmathspace" /></math> </ephtml> </p> <p>where <emph>PBR</emph> is Percent Bias Reduction and <emph>B</emph> is bias. PBR larger than 80% is deemed effective. In the present study, we used the nearest neighbor matching with a 0.2 caliper, as it provided the greatest bias reduction (91.9% PBR). This is in contrast to nearest neighbor matching without the 0.2 caliper (86.9% PBR) and optimal matching (87.0% PBR).</p> <hd id="AN0176533093-20">Logistic Regression</hd> <p>Following the matching process, we conducted logistic regression to determine the predictive utility of state subsidy policies on centers' subsidy system participation. By carrying out the matching process described above, any differences between groups on the dependent variable (subsidy system participation) could be more strongly attributed to variation in state CCDF policies. In other words, because we divided the analytic sample across subsidy vs. non-subsidy groups, we used the logistic regression analysis to determine state CCDF policies' relative impact on subsidy system participation.</p> <hd id="AN0176533093-21">Results</hd> <p>The NSECE includes 7,771 ECE centers, including 2,640 subsidy centers and 5,131 non-subsidy centers. Across 20 imputed datasets, on average, subsidy centers had a mean propensity score of 0.460 (SD =.162) and non-subsidy centers had a mean propensity score of 0.278 (SD =.188) before matching. Using 1:1 nearest neighbor matching with caliper (0.2) matching, we matched a total of 4,788 centers across 20 imputations. There were 2,983 unmatched centers and we excluded these centers in further analyses. After matching, subsidy centers had a mean propensity score of 0.457 (SD =.159) and non-subsidy centers had a mean propensity score of 0.442 (SD =.151). The percent bias reduction (PBR) was 91.9%, well above the 80% that is deemed effective.</p> <p>Following the matching process, we assessed the balance on matched datasets using functions in the <emph>cobalt</emph> package in R and by running significance tests (t-tests for continuous variables and chi-square tests for dichotomous variables). We present results of the significance tests in Table 1.</p> <p>Table 1. Significance test results for variables used in the matching process.</p> <p> <ephtml> <table><thead><tr><td /><td>Before Matching</td><td>After Matching</td></tr><tr><td /><td>Subsidy Centers <italic>n</italic> = 2640</td><td>Non-Subsidy Centers <italic>n</italic> = 5131</td><td /><td /><td>Subsidy Centers <italic>n</italic> = 2394</td><td>Non-Subsidy Centers <italic>n</italic> = 2394</td><td /><td /></tr><tr><td>Variable</td><td>n(%)/mean(SD)</td><td>n(%)/mean(SD)</td><td>χ<sup>2</sup> / <italic>t</italic></td><td><italic>p</italic></td><td>n(%)/mean(SD)</td><td>n(%)/mean(SD)</td><td>χ<sup>2</sup> / <italic>t</italic></td><td><italic>p</italic></td></tr></thead><tbody><tr><td>For-Profit Program</td><td>1298(49.2)</td><td>1198(23.3)</td><td>532.94</td><td><.001</td><td>1209(50.5)</td><td>1029(43.0)</td><td>27.18</td><td><.001</td></tr><tr><td>Program has an overall quality rating</td><td>1363(51.6)</td><td>2431(47.4)</td><td>12.60</td><td><.001</td><td>1176(49.1)</td><td>1200(50.1)</td><td>0.48</td><td>.488</td></tr><tr><td>Program serves 0–3-year-olds</td><td>2283(86.5)</td><td>2545(49.6)</td><td>1007.47</td><td><.001</td><td>2038(85.1)</td><td>2052(85.7)</td><td>0.33</td><td>.566</td></tr><tr><td>Number of children enrolled</td><td>69.6(46.1)</td><td>63.9(48.5)</td><td>−5.01</td><td><.001</td><td>68.20(46.762)</td><td>68.96(46.540)</td><td>0.57</td><td>.570</td></tr><tr><td>School sponsorship</td><td>109(4.1)</td><td>491(9.6)</td><td>72.42</td><td><.001</td><td>103(4.3)</td><td>120(5.0)</td><td>1.36</td><td>.244</td></tr><tr><td>Church sponsorship</td><td>130(4.9)</td><td>419(8.2)</td><td>27.90</td><td><.001</td><td>128(5.3)</td><td>135(5.6)</td><td>0.20</td><td>.657</td></tr><tr><td>Program in an area with high poverty density</td><td>817(30.9)</td><td>1481(28.9)</td><td>3.63</td><td>.057</td><td>792(33.1)</td><td>687(28.7)</td><td>10.79</td><td>.001</td></tr><tr><td>Program in an area with high urban density</td><td>2296(87.0)</td><td>4318(84.2)</td><td>10.90</td><td><.001</td><td>2084(87.1)</td><td>2077(86.8)</td><td>0.09</td><td>.764</td></tr><tr><td>Program receives funds from parent pay</td><td>2437(92.3)</td><td>3226(62.9)</td><td>764.15</td><td><.001</td><td>2191(91.5)</td><td>2177(90.9)</td><td>0.51</td><td>.474</td></tr><tr><td>Program receives funds from Head Start</td><td>296(11.2)</td><td>1095(21.3)</td><td>121.69</td><td><.001</td><td>287(12.0)</td><td>210(8.8)</td><td>13.31</td><td><.001</td></tr><tr><td>Program receives funds from pre-K</td><td>545(20.6)</td><td>1189(23.2)</td><td>6.43</td><td>.011</td><td>529(22.1)</td><td>514(21.5)</td><td>0.28</td><td>.599</td></tr></tbody></table> </ephtml> </p> <p>As Table 1 shows, there were significant differences in all measures before matching, but the matching process considerably reduced the number of significant differences. Results show that there were a few significant differences for each of the measures between the matched samples of subsidy and non-subsidy ECE centers. However, because the standard mean difference was well below the standard cutoff of 0.01 for these comparisons these significant differences are not cause for concern.</p> <p>Significance tests can be misleading due to changes in the sample size (Imai et al., [<reflink idref="bib31" id="ref90">31</reflink>]), so we rely on a more common metric for evaluating covariate balance – the standardized difference in means. We used the absolute standardized mean differences (ASMDs) and Kolmogorov-Smirnov (KS) statistics as metrics to evaluate the extent to which propensity score matching created groups with similar covariate distributions. KS statistics are useful for evaluating balance because they measure the greatest distance between the empirical cumulative distribution functions for each variable between the two groups, and values close to 0 denote balance (Austin & Stuart, [<reflink idref="bib6" id="ref91">6</reflink>]). ASMDs and KS values in our matching evaluation ranged from 0.0062 (center accepts parent payment) − 0.0909 (center is a for-profit program). Evaluating matches revealed covariates were well-balanced because the ASMD and KS statistics for all covariates across the imputed datasets were closer to zero. In addition, these values were all less than 0.1, consistent with recommendations for ideal standardized differences in means (Stuart, [<reflink idref="bib55" id="ref92">55</reflink>]). The standardized difference in means metrics are presented as a "love plot" (Rudolph et al., [<reflink idref="bib49" id="ref93">49</reflink>]) in Figure 2, which depicts the balance on each covariate before and after matching. These methods for evaluating covariate balance suggest that matching was successful.</p> <p>Graph: Figure 2. Love Plot: Balance on each covariate before and after matching across all 20 imputations.</p> <hd id="AN0176533093-22">Descriptive Statistics</hd> <p>Table 2 depicts the variables that were used in the matching process across 20 imputations and their standard mean differences (SMD) before and after the matching process. 2,394 subsidy centers were matched with 2,394 non-subsidy centers. The SMD after matching was closer to zero than before matching on all variables, indicating matching was successful.</p> <p>Table 2. Covariates used in the matching process.</p> <p> <ephtml> <table><thead><tr><td /><td>Before Matching</td><td>After Matching</td></tr><tr><td>Variable</td><td>SMD</td><td>SMD</td></tr></thead><tbody><tr><td>Program has a quality rating</td><td>0.085</td><td>−0.020</td></tr><tr><td>Program serves 0–3-year-olds</td><td>1.078</td><td>−0.017</td></tr><tr><td>For-Profit program</td><td>0.516</td><td>0.150</td></tr><tr><td>Total enrollment</td><td>0.124</td><td>−0.017</td></tr><tr><td>Church-sponsored</td><td>−0.150</td><td>−0.014</td></tr><tr><td>School-sponsored</td><td>−0.274</td><td>−0.036</td></tr><tr><td>Program in an area with high poverty density</td><td>0.045</td><td>0.095</td></tr><tr><td>Program in an area with high urban density</td><td>0.084</td><td>0.009</td></tr><tr><td>Program receives funds from parent pay</td><td>1.105</td><td>0.022</td></tr><tr><td>Program receives funds from Head Start</td><td>−0.321</td><td>0.102</td></tr><tr><td>Program receives funds from pre-K</td><td>−0.063</td><td>0.016</td></tr></tbody></table> </ephtml> </p> <p>1 SMD = standard mean differences.</p> <p>Table 3 displays descriptive statistics for the state-specific subsidy policies from the CCDF Policies Database and is broken down by the ECE center's subsidy system participation status. We used these policy variables as our key predictors in the logistic regression analysis that occurred after matching to assess the influence of these policies on subsidy system participation (i.e., whether the ECE center accepts subsidies). Chi-square and t-tests indicate there were significantly fewer subsidy centers in states with a policy that families were responsible for paying the difference between the private pay rate and the state subsidy reimbursement rate before and after matching. There were also smaller percentages of subsidy centers located in states that have a policy to reimburse for center closures across the unmatched and matched samples.</p> <p>Table 3. Subsidy policies impacting sample, before and after matching.</p> <p> <ephtml> <table><thead><tr><td /><td>Before Matching</td><td>After Matching</td></tr><tr><td /><td>Subsidy Centers <italic>n</italic> = 2640</td><td>Non-Subsidy Centers <italic>n</italic> = 5131</td><td /><td /><td>Subsidy Centers <italic>n</italic> = 2394</td><td>Non-Subsidy Centers <italic>n</italic> = 2394</td><td /><td /></tr><tr><td>Policy Variable</td><td>n(%)/mean(SD)</td><td>n(%)/mean(SD)</td><td>χ<sup>2</sup> / <italic>t</italic></td><td><italic>P</italic></td><td>n(%)/mean(SD)</td><td>n(%)/mean(SD)</td><td>χ<sup>2</sup> / <italic>t</italic></td><td><italic>P</italic></td></tr></thead><tbody><tr><td>Base subsidy reimbursement rate, adjusted for RPP</td><td>599.5(132.4)</td><td>609.4(137.0)</td><td>3.08</td><td>.002</td><td>598.4(132.3)</td><td>602.2(138.7)</td><td>0.96</td><td>.338</td></tr><tr><td>Reimbursement for child absences</td><td>2483(94.1%)</td><td>4818(93.9%)</td><td>0.07</td><td>.788</td><td>2263(94.5%)</td><td>2257(94.3%)</td><td>0.14</td><td>.706</td></tr><tr><td>State subsidy contracts</td><td>1180(44.7%)</td><td>2457(47.9%)</td><td>7.12</td><td>.008</td><td>1065(44.5%)</td><td>1106(46.2%)</td><td>1.42</td><td>.234</td></tr><tr><td>Families responsible for paying difference between subsidy reimbursement rate and private pay rate</td><td>2184(82.7%)</td><td>4361(85.0%)</td><td>6.74</td><td>.009</td><td>1986(83.0%)</td><td>2044(85.4%)</td><td>5.27</td><td>.022</td></tr><tr><td>Tiered reimbursement</td><td>1218(46.1%)</td><td>2181(42.5%)</td><td>9.33</td><td>.002</td><td>1098(45.9%)</td><td>1052(43.9%)</td><td>1.79</td><td>.181</td></tr><tr><td>Reimbursement for center closures</td><td>1795(68.0%)</td><td>3972(77.4%)</td><td>80.81</td><td><.001</td><td>1630(68.1%)</td><td>1889(78.9%)</td><td>71.92</td><td><.001</td></tr></tbody></table> </ephtml> </p> <p>2 RPP = regional price parity.</p> <hd id="AN0176533093-23">Effect of Subsidy Policies on Subsidy System Participation</hd> <p>We examined the effect of the CCDF policies on subsidy participation using a logistic regression analysis. Results are displayed in Table 4. Predictors of centers' subsidy participation (1 = participate; 0 = did not participate) included: a continuous variable capturing the base subsidy reimbursement rate, adjusted for Regional Price Parity, and several dummy variables capturing the presence of specific policies including a tiered reimbursement policy (0=no, 1=yes), a policy that reimbursed ECE centers for child absences (0=no, 1=yes), a policy that provide subsidy reimbursement for center closures (0=no, 1=yes), a variable capturing if the family was required to pay any difference between the ECE center's parent-pay rate (0=no, 1=yes), and if the state used contracts (vs. vouchers) as the mechanism for subsidy reimbursement (0=no, 1=yes, contracts). We screened for multicollinearity and found that the predictors were not collinear.</p> <p>Table 4. Results of logistic regression.</p> <p> <ephtml> <table><thead><tr><td>Policy Variable</td><td>Logit</td><td>S.E.</td><td><italic>p</italic>-value</td><td>Odds Ratio</td><td>95% Confidence Intervals</td></tr></thead><tbody><tr><td>Base subsidy reimbursement rate, adjusted for RPP</td><td>0.076</td><td>0.036</td><td>.036*</td><td>0.927</td><td>0.855</td><td>0.998</td></tr><tr><td>Tiered reimbursement</td><td>0.142</td><td>0.070</td><td>.043*</td><td>1.152</td><td>1.015</td><td>1.289</td></tr><tr><td>Reimbursement for child absences</td><td>0.269</td><td>0.130</td><td>.039*</td><td>1.309</td><td>1.054</td><td>1.565</td></tr><tr><td>Reimbursement for center closures</td><td>−0.659</td><td>0.073</td><td><.001***</td><td>0.517</td><td>0.373</td><td>0.661</td></tr><tr><td>State subsidy contracts</td><td>0.140</td><td>0.070</td><td>.046*</td><td>1.151</td><td>1.013</td><td>1.289</td></tr><tr><td>Families responsible for paying difference between subsidy reimbursement rate and private pay rate</td><td>−0.055</td><td>0.083</td><td>.508</td><td>0.946</td><td>0.783</td><td>1.110</td></tr></tbody></table> </ephtml> </p> <p>3 RPP = regional price parity. ***<emph>p</emph> <.001, *<emph>p</emph> <.05.</p> <p>Results suggest that nearly all the state-specific subsidy policies we included in our study influenced the likelihood that an ECE center accepted subsidies (see Table 4). Specifically, the subsidy reimbursement rate and the policies that impacted the amount of money centers receive for accepting subsidies appeared to influence subsidy participation. Results suggest that as the base reimbursement rate increased, likelihood of subsidy system participation increased (Odds Ratio (OR) = 0.927, Confidence Intervals [CI]: 0.855–0.998, <emph>p</emph> =.036). Having a statewide tiered reimbursement policy that provides increased reimbursement rates for higher quality programs increased the odds of subsidy system participation (OR = 1.152, CI = 1.015–1.289, <emph>p</emph> =.043).</p> <p>A few other subsidy payment policies directly impacting centers appeared to influence subsidy system participation. Results suggest that if states reimbursed ECE programs for days the child was absent, ECE centers were more likely to accept subsidies (OR = 1.309, CI = 1.054–1.565, <emph>p</emph> =.039). On the other hand, when states reimbursed ECE centers for closures, centers had lower odds of accepting subsidies (OR = 0.517, CI = 0.373–0.661, <emph>p</emph> <.001).</p> <p>Our results also suggest that the way a state administered subsidies may also influence centers' subsidy participation. When ECE centers had contracts with states (vs. vouchers), subsidy system participation was more likely (OR = 1.151, CI = 1.013–1.289, <emph>p</emph> =.046).</p> <hd id="AN0176533093-24">Discussion</hd> <p>To provide clear policy recommendations for incentivizing ECE centers to accept subsidies, the present study used propensity score methods to estimate the predictive utility of state-specific subsidy policies on program participation in the subsidy system. Our results are timely and important given the continued decline in the number of centers that accept child care subsidies. While previous work has provided some initial recommendations for motivating subsidy participation among ECE programs (Giapponi Schneider et al., [<reflink idref="bib24" id="ref94">24</reflink>]; Slicker & Hustedt, [<reflink idref="bib53" id="ref95">53</reflink>]; Slicker et al., [<reflink idref="bib54" id="ref96">54</reflink>]), the center- and community-level associations with subsidy participation are more difficult to translate into immediate state action. Our findings, on the other hand, have important policy implications that have the potential to increase subsidy participation through modifications to a series of state-specific subsidy policies. Consistent with suggestions that subsidy reimbursements may impact programs' subsidy participation (Schulman, [<reflink idref="bib51" id="ref97">51</reflink>]), we found that subsidy policies that impact the center's revenue for accepting subsidies are important to consider. Furthermore, given the documented administrative difficulties associated with subsidy system participation (Adams et al., [<reflink idref="bib4" id="ref98">4</reflink>]; Sandstrom et al., [<reflink idref="bib50" id="ref99">50</reflink>]), our results appear to suggest that the way the state administers subsidies was also predictive of centers' subsidy participation.</p> <p>Our results suggest that the state subsidy reimbursement rate impacts centers' subsidy system participation. As the base subsidy rate increased, ECE centers were more likely to accept subsidies. Given that subsidy reimbursement rates were lower than the market rate paid by private-paying families in 47 states in 2019 (Schulman, [<reflink idref="bib51" id="ref100">51</reflink>]), this finding strengthens the argument for increased subsidy reimbursement rates as a mechanism to encourage ECE centers to accept subsidies. Subsidy reimbursement rates are important to the financial structure of the ECE center. These payments determine the salaries that centers can offer staff, the teacher-child ratio in the classroom (so long as it meets licensing and other state requirements), and the materials and supplies that can be purchased for use in classrooms. This finding suggests that policymakers should prioritize increasing reimbursement rates to incentivize subsidy participation, as inadequate payments may not only discourage participation but could also diminish the quality of the center programming if slots are filled with children using subsidies as payment as opposed to (higher) parent payments (Antle et al., [<reflink idref="bib5" id="ref101">5</reflink>]).</p> <p>Our findings suggest the importance of tiered reimbursement policies, as the presence of a tiered reimbursement policy in the state increased the odds of centers' subsidy participation. This finding may suggest that having meaningfully higher reimbursement rates available for ECE centers meeting certain state-specified quality benchmarks is an incentive for centers to accept subsidies. According to the CCDF Policies Database used for this study, at the time the data for this study was collected in 2011, 22 states and D.C. had tiered reimbursement policies in place and the number of states with a tiered reimbursement policy has increased over time (Build Initiative & Child Trends, [<reflink idref="bib12" id="ref102">12</reflink>]). Our results are interesting to compare alongside the results of a recent statewide study on tiered reimbursement, which found that ECE centers were not motivated to achieve higher levels of quality given the presence of a tiered reimbursement policy (Lee, [<reflink idref="bib34" id="ref103">34</reflink>]). However, another recent study in a state where the highest reimbursement rate for centers achieving the highest quality rating was more than 90% of the market rate found that a tiered reimbursement policy did incentivize subsidy participation (Slicker et al., [<reflink idref="bib54" id="ref104">54</reflink>]). Our current results may suggest the presence of a tiered reimbursement policy is motivating to ECE centers on a national scale because the tiered payments are predictive of subsidy participation.</p> <p>Findings from our study suggest that other subsidy policies that make it easier for ECE centers to manage subsidies may also incentivize centers' participation in the subsidy system. For example, our findings suggest that if centers were reimbursed for days that children were absent, they were more likely to accept subsidies. This is important to consider because, in many ECE programs, private paying families are expected to pay even when their child is sick or otherwise unable to attend. Only half of U.S. states had a policy that reimbursed ECE centers when children were absent in 2011 (Minton et al., [<reflink idref="bib40" id="ref105">40</reflink>]). Having policies in place whereby ECE centers that accept subsidies do not receive payment for child absences could serve as a disincentive to accept subsidies when centers could instead enroll a private-paying family and be guaranteed payment irrespective of a child's attendance.</p> <p>Previous research suggests that contracts, which are direct payments to programs that accept subsidies (as compared to vouchers), can be very beneficial to ECE programs. For instance, contracts can help to stabilize revenue and reduce overall administrative burden for ECE programs (Adams & Rohacek, [<reflink idref="bib3" id="ref106">3</reflink>]; Matthews & Schumacher, [<reflink idref="bib38" id="ref107">38</reflink>]). Our findings suggest that when centers were in states where programs received subsidy payments through state contracts, the odds of ECE centers participation in the subsidy system increased. In 2011, only 17 states and DC administered subsidies through state contracts (Minton et al., [<reflink idref="bib40" id="ref108">40</reflink>]), suggesting an opportunity for expansion of this policy to other states.</p> <p>Our findings do not provide support for continuing payments when centers were closed. In fact, when states provided payments for center closures, centers were less likely to accept subsidies. This finding should be interpreted with caution because the variable used for this analysis was a dichotomous variable capturing whether the state reimbursed the center for closures at all and did not capture variation in the type of closures (i.e., closures for state holidays, inclement weather, professional development) or the number of days a center could be reimbursed for center closures. For example, in 2011 both Mississippi and South Carolina were technically paid for center closures, but while Mississippi could receive payment for up to 11 holidays per year, South Carolina could only receive state subsidy payments if they also billed private paying families for center closures and there was seemingly no limit to the number of days centers could be reimbursed. Further caution when interpreting this finding should be exercised because 19 states had missing data on this variable.</p> <p>In addition, a policy that required families to pay any difference between the center's parent payment rate and the state's reimbursement rate did not predict subsidy system participation. This finding is important because though the policy could theoretically benefit ECE programs financially, it has the potential to limit ECE access for families who cannot afford to pay that difference. According to state subsidy policies in place in 2019, 32 states required the family to pay the difference between the parent payment rate and the subsidy reimbursement rate and an additional 7 states have a policy that a family "may" be asked to pay the difference (Dwyer et al., [<reflink idref="bib21" id="ref109">21</reflink>]). This finding suggests that there may be more effective state policies for assisting ECE programs that accept subsidies to cover the difference between private pay rates and subsidy reimbursement rates than asking families to cover that difference.</p> <hd id="AN0176533093-25">Limitations & Future Directions</hd> <p>This study provides important evidence for the predictive utility of state-specific subsidy policies on centers' participation in the subsidy system; however, there were limitations to our approach. First, though we used nationally representative data from the NSECE and the corresponding subsidy policies from the CCDF Policies Database, the data used predate the reauthorization of CCDF in 2014. The reauthorization added requirements for programs that accept subsidies. Since the reauthorization, states have adjusted their subsidy policies and these modifications could have important implications for ECE centers as they weigh subsidy participation. More recent data would provide opportunities for examining the relationship between state subsidy policies and ECE centers' subsidy participation under the new CCDF program. Yet, our study provides valuable information about the subsidy system participation prior to these changes that may help make future comparisons that could increase understanding about the potential role of the CCDF reauthorization. Future research should also examine the impact of CCDF policies for home-based ECE programs. For example, states' base monthly reimbursement rates for an infant in home-based program ranged from $200 to $1,254 in 2019 with 31 states offering tiered reimbursement rates for home-based providers (Dwyer et al., [<reflink idref="bib22" id="ref110">22</reflink>]), and variation in these policies may impact decisions these providers make about accepting subsidies. Giapponi Schneider et al. ([<reflink idref="bib24" id="ref111">24</reflink>]) found that higher subsidy rates predicted subsidy participation for home-based providers, suggesting that our findings about the importance of state-specific CCDF policies such as reimbursement rates and tiered reimbursement for subsidy participation may also be true for home-based providers. However, there could be other factors unique to home-based care that shape subsidy participation decisions. This work is especially important because the largest decline of ECE programs that accept subsidies in recent years is amongst home-based providers (Adams & Dwyer, [<reflink idref="bib2" id="ref112">2</reflink>]).</p> <p>Though the present study has important implications for policymakers regarding specific subsidy policies that could be prioritized to incentivize centers' subsidy participation, states have a variety of policies to consider as it relates to subsidy administration. In fact, states often must balance the aims of CCDF with the unfortunate reality that federal funding is not sufficient to adequately serve all eligible children and families. An opportunity for future research could be to adopt an approach similar to previous research (Madill et al., [<reflink idref="bib36" id="ref113">36</reflink>]) by examining "packages" of multiple subsidy policies – including subsidy reimbursement rates alongside income and employment eligibility requirements for families – and their relationships with subsidy system participation. In addition, there could be other important policies or sources of motivation, such as a program's mission to serve specific children or families (Slicker et al., [<reflink idref="bib54" id="ref114">54</reflink>]), that may play a role in decisions around subsidy participation that were not available in our merged dataset. Relatedly, interviews with center administrators about their subsidy participation decision-making process could provideimportant context about the role of state subsidy policies.</p> <p>Similarly, this study did not specifically consider the impact of state variation in policies that impact subsidy eligibility for families. For example, the maximum family income that allows a new family to become eligible for subsidies for a family of three ranges from $1,423 to $5,802 per month across states (Dwyer et al., [<reflink idref="bib22" id="ref115">22</reflink>]). Variation in family eligibility also means that subsidy-eligible families may or may not be eligible for other publicly funded ECE in their state (e.g., Head Start, pre-K). While variation in these state policies may not directly impact ECE centers, future research may want to consider their impact on centers' subsidy participation.</p> <p>While beyond the scope of this study, previous research suggests that the quality of ECE centers is also higher in states with subsidy payments for absent children, higher subsidy reimbursement rates, and tiered reimbursement, particularly when there is a larger gap between the highest and lowest reimbursement tiers (Greenberg et al., [<reflink idref="bib28" id="ref116">28</reflink>]). Future research should gather causal evidence about the relationship between subsidy policies that positively impact the financial structure of ECE centers (i.e., subsidy reimbursement rates) and the quality of care. Given that these policies impact subsidy participation and have the potential to positively shape the quality of ECE services provided (e.g., Bassok et al., [<reflink idref="bib10" id="ref117">10</reflink>]; Johnson et al., [<reflink idref="bib33" id="ref118">33</reflink>]) and children's outcomes (Zanoni & Johnson, [<reflink idref="bib62" id="ref119">62</reflink>]), policymakers may want to consider prioritizing these types of subsidy policies, particularly given that federal subsidy contributions, however generous, are typically not sufficient to reach all eligible families.</p> <hd id="AN0176533093-26">Conclusion</hd> <p>The present study relied on a nationally representative sample of ECE centers and the state-specific subsidy policies in place to examine the predictive utility of state subsidy policies on centers' subsidy participation. The findings point to several specific policies that predicted subsidy system participation: tiered reimbursement, payment for child absences, and state subsidy contracts as the method of payment. In addition, our findings suggest that as the base subsidy reimbursement rate increased, subsidy system participation was more likely. State policymakers evaluating all their subsidy policies may want to consider prioritizing these items when making decisions about how to create their CCDF plans and policy agenda.</p> <p>On the other hand, a state policy that required families to cover the difference between the state reimbursement rate and the ECE center's private pay rate was not a significant predictor of subsidy participation. In addition, we find that a policy that reimburses for center closures was negatively associated with subsidy participation. Taken together, these findings may suggest that while centers must consider the financial well-being of the center, ECE centers may also be dissuaded by policies that have the potential to negatively impact working families by charging families more to cover the cost of care or limiting families' daily access to ECE by reimbursing centers for closing, even if it is for important purposes like providing professional development for teachers. These findings may mean that state-level policymakers should prioritize policies that are beneficial to ECE programs, but not at the expense of families, as they try to implement subsidy policies in their state that incentivize centers' subsidy participation.</p> <p>In the context of limited funds available for use by states as they administer the subsidy program, these findings may be important as states consider how to incentivize more ECE centers to accept subsidies. Our results suggest the importance of policies that increased the amount of state-issued funds an ECE center receives for children using subsidies (but not necessarily funds that come from/at the expense of families). These incentivizing policies include the actual reimbursement rate, but also policies that may provide supplemental funds to ECE centers such as tiered reimbursement and payment for child absences. In addition, policies that supported ECE centers as they managed subsidies, such as state subsidy contracts, also appeared to motivate subsidy participation. Programs that are willing to accept subsidies are critical to ensuring equitable ECE access for families living in poverty and, as a result, this study has implications for increasing access to ECE.</p> <hd id="AN0176533093-27">Disclosure Statement</hd> <p>No potential conflict of interest was reported by the author(s).</p> <ref id="AN0176533093-28"> <title> References </title> <blist> <bibl id="bib1" idref="ref74" type="bt">1</bibl> <bibtext> Acock, A. C., & Long, D. L. (2012). What to do about missing values. In H. Cooper, P. M. Camic, A. T. Panter, D. Rindskopf, & K. J. Sher (Eds.), APA handbooks in psychology. APA handbook of research methods in psychology, Vol. 3. Data analysis and research publication (pp. 27 – 50). American Psychological Association. https://doi.org/10.1037/13621-002</bibtext> </blist> <blist> <bibl id="bib2" idref="ref112" type="bt">2</bibl> <bibtext> Adams, G., & Dwyer, K. (2021, April). Child care subsidies and home-based child care providers: Expanding participation. The Urban Institute.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref33" type="bt">3</bibl> <bibtext> Adams, G., & Rohacek, M. (2002). More than a work support? Issues around integrating child development goals into the child care subsidy system. Early Childhood Research Quarterly, 17 (4), 418 – 440. https://doi.org/10.1016/S0885-2006(02)00184-9</bibtext> </blist> <blist> <bibl id="bib4" idref="ref37" type="bt">4</bibl> <bibtext> Adams, G., Rohacek, M., & Snyder, K. (2008). Child care voucher programs: provider experiences in five counties. The Urban Institute.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref101" type="bt">5</bibl> <bibtext> Antle, B. F., Frey, A., Barbee, A., Frey, S., Grisham-Brown, J., & Cox, M. (2008). Child care subsidy and program quality revisited. Early Education and Development, 19 (4), 560 – 573. https://doi.org/10.1080/10409280802230999</bibtext> </blist> <blist> <bibl id="bib6" idref="ref91" type="bt">6</bibl> <bibtext> Austin, P. C., & Stuart, E. A. (2015). Moving towards best practice when using Inverse Probability of Treatment Weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Statistics in Medicine, 34 (28), 3661 – 3679. https://doi.org/10.1002/sim.6607</bibtext> </blist> <blist> <bibl id="bib7" idref="ref72" type="bt">7</bibl> <bibtext> Bai, H. & Clark, M. (2019). Propensity score methods and applications. SAGE Publications, Inc. https://doi.org/10.4135/9781071814253</bibtext> </blist> <blist> <bibl id="bib8" idref="ref1" type="bt">8</bibl> <bibtext> Baldiga, M., Joshi, P., Hardy, E., & Acevedo-Garcia, D. (2018). Data-for-equity research brief: Child care affordability for working parents. Brandeis University. https://<ulink href="http://www.nichq.org/sites/default/files/resource-file/ChildCare%20Affordability%20brief.pdf">www.nichq.org/sites/default/files/resource-file/ChildCare%20Affordability%20brief.pdf</ulink></bibtext> </blist> <blist> <bibl id="bib9" idref="ref25" type="bt">9</bibl> <bibtext> Bassok, D., Dee, T. S., & Latham, S. (2019). The effects of accountability incentives in early childhood education. Journal of Policy Analysis and Management, 38 (4), 838 – 866. https://doi.org/10.1002/pam.22149</bibtext> </blist> <blist> <bibtext> Bassok, D., Magouirk, P., & Markowitz, A. J. (2021). Systemwide quality improvement in early childhood education: Evidence from Louisiana. AERA Open, 7, 233285842110116. https://doi.org/10.1177/23328584211011610</bibtext> </blist> <blist> <bibtext> Blau, D. M. (2022). The case for targeted preschool and childcare subsidies. Journal of Policy Analysis and Management, 41 (3), 929 – 937. https://doi.org/10.1002/pam.22400</bibtext> </blist> <blist> <bibtext> Build Initiative & Child Trends. (2019). A catalog and comparison of quality initiatives [Data system]. <ulink href="http://qualitycompendium.org/">http://qualitycompendium.org/</ulink></bibtext> </blist> <blist> <bibtext> Caliendo, M., & Kopeinig, S. (2008). Some practical guidance for the implementation of propensity score matching. Journal of Economic Surveys, 22 (1), 31 – 72. https://doi.org/10.1111/j.1467-6419.2007.00527.x</bibtext> </blist> <blist> <bibtext> Child Care and Development Fund Program. (2016). Rules and regulations, 81 Federal Register 190 (codified at 45 C.F.R., pt. 98).</bibtext> </blist> <blist> <bibtext> Child Care Aware of America. (2019). The U.S. and the high price of child care: An examination of a broken system. https://cdn2.hubspot.net/hubfs/3957809/2019%20Price%20of%20Care%20State%20Sheets/Final-TheUSandtheHighPriceofChildCare-AnExaminationofaBrokenSystem.pdf?utm_referrer=https%3A%2F%2Fwww.childcareaware.org%2Four-issues%2Fresearch%2Fthe-us-and-the-high-price-of-child-care-2019%2F</bibtext> </blist> <blist> <bibtext> Cochran, W. G., & Rubin, D. B. (1973). Controlling bias in observational studies: A review. The Indian Journal of Statistics, Series A, 35 (4), 417 – 446.</bibtext> </blist> <blist> <bibtext> Crosby, D. A., Gennetian, L., & Huston, A. C. (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. https://doi.org/10.1207/s1532480xads0902_4</bibtext> </blist> <blist> <bibtext> Davis, E. E., Grobe, D., & Weber, R. B. (2010). Rural-urban differences in childcare subsidy use and employment stability. Applied Economic Perspectives and Policy, 32 (1), 135 – 153. https://doi.org/10.1093/aepp/ppp004</bibtext> </blist> <blist> <bibtext> Davis, E. E., Krafft, C., & Forry, N. D. (2016). The role of policy and practice in short spells of child care subsidy participation. Journal of Public Administration Research and Theory, 1 (1), 1 – 19. https://doi.org/10.1093/jopart/muw039</bibtext> </blist> <blist> <bibtext> De Marco, A., & Vernon-Feagans, L. (2015). Child care subsidy use and child care quality in low-wealth, rural communities. Journal of Family and Economic Issues, 36 (3), 383 – 395. https://doi.org/10.1007/s10834-014-9401-8</bibtext> </blist> <blist> <bibtext> Dwyer, K., Minton, S., Kwon, D., & Weisner, K. (2020). Key Cross-State Variations in CCDF Policies as of October 1, 2019: The CCDF Policies Database Book of Tables OPRE Report 2021-07. Office of Planning, Research, and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services. https://ccdf.urban.org/sites/default/files/state_2019_ccdf_policies_dec_2020.pdf</bibtext> </blist> <blist> <bibtext> Dwyer, K., Minton, S., Kwon, D., & Weisner, K. (2021, February). Key cross-state variations in CCDF policies as of October 1, 2019. The Urban Institute.</bibtext> </blist> <blist> <bibtext> Dwyer, K., Tran, V., & Minton, S. (2019). Child care subsidies under the CCDF program: An overview of policy differences across states and territories as of October 1, 2018 (OPRE Report 2019-116). Urban Institute.</bibtext> </blist> <blist> <bibtext> Giapponi Schneider, K., Erickson Warfield, M., Joshi, P., Ha, Y., & Hodgkin, D. (2017). Insights into the black box of child care supply: Predictors of provider participation in the Massachusetts child care subsidy system. Children and Youth Services Review, 79, 148 – 159. https://doi.org/10.1016/j.childyouth.2017.06.014</bibtext> </blist> <blist> <bibtext> Gomez, C. J., Whitaker, A. A., & Cannon, J. S. (2022). Do early care and education programs improve when enrolled in quality rating and improvement systems? Longitudinal evidence from one system. Early Education and Development, 34 (5), 1236 – 1253.</bibtext> </blist> <blist> <bibtext> Graham, J. W. (2012). Missing data: Analysis and design. Springer Science and Business Media.</bibtext> </blist> <blist> <bibtext> Graham, J. W., Olchowski, A. E., & Gilreath, T. D. (2007). How many imputations are really needed? Some practical clarifications of multiple imputation theory. Prevention Science, 8 (3), 206 – 213. https://doi.org/10.1007/s11121-007-0070-9</bibtext> </blist> <blist> <bibtext> Greenberg, E., Isaacs, J. B., Derrick-Mills, T., Michie, M., & Stevens, K. (2018). Are higher subsidy payment rates and provider-friendly payment policies associated with child care quality? The Urban Institute.</bibtext> </blist> <blist> <bibtext> Ha, Y., Joshi, P., Giapponi Schneider, K., & Hardy, E. (2020). Can administrative changes improve child-care stability? Social Service Review, 94 (2), 285 – 338. https://doi.org/10.1086/70944</bibtext> </blist> <blist> <bibtext> Halle, T., Banghart, P., Zaslow, M., Cook, M., Kane, M., Bartlett, J. D., Llyod, C. M., Cox, A., Lloyd, C. M., & Redd, Z. (2019). Implementation lessons from six early head start- Child care partnerships. Early Education & Development, 30 (8), 990 – 1008. https://doi.org/10.1080/10409289.2019.1656320</bibtext> </blist> <blist> <bibtext> Imai, K., King, G., & Stuart, E. A. (2008). Misunderstandings between experimentalists and observationalists about causal inference. Journal of the Royal Statistical Society: Series A (Statistics in Society), 171 (2), 481 – 502. https://doi.org/10.1111/j.1467-985X.2007.00527.x</bibtext> </blist> <blist> <bibtext> Jenkins, J. M., & Nguyen, T. (2022). Keeping kids in care: Reducing administrative burden in state Child Care Development Fund policy. Journal of Public Administration Research and Theory, 1 (1), 23 – 40. https://doi.org/10.1093/jopart/muab020</bibtext> </blist> <blist> <bibtext> Johnson, A. D., Ryan, R. M., & Brooks-Gunn, J. (2012). Child-care subsidies: Do they impact the quality of care children experience? Child Development, 83 (4), 1444 – 1461. https://doi.org/10.1111/j.1467-8624.2012.1780.x</bibtext> </blist> <blist> <bibtext> Lee, E. S. (2021). A mixed-methods study of Maryland's monetary incentives to improve the quality of child care centers. Early Childhood Research Quarterly, 55, 349 – 362. https://doi.org/10.1016/j.ecresq.2021.01.002</bibtext> </blist> <blist> <bibtext> Leite, W. (2016). Practical propensity score methods using R. Sage Publications. https://doi.org/10.4135/9781071802854</bibtext> </blist> <blist> <bibtext> Madill, R., Lin, V.-K., Friese, S., & Paschall, K. (2018). Access to early care and education for disadvantaged families: Do levels of access reflect states' child care subsidy policies? Child Trends.</bibtext> </blist> <blist> <bibtext> Magnuson, K. A., & Waldfogel, J. (2016). Trends in income-related gaps in enrollment in early childhood education: 1968 to 2013. AERA Open, 2 (2), 1 – 13. https://doi.org/10.1177/2332858416648933</bibtext> </blist> <blist> <bibtext> Matthews, H., & Schumacher, R. (2008). Ensuring quality care for low-income babies: Contracting directly with providers to expand and improve infant and toddler care. Center for Law and Social Policy.</bibtext> </blist> <blist> <bibtext> Michalopoulos, C., Lundquist, E., & Castells, N. (2010). The effects of child care subsidies for moderate-income families in Cook County, Illinois (OPRE 2011-3). Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services.</bibtext> </blist> <blist> <bibtext> Minton, S., Durham, C., Huber, E., & Giannarelli, L. (2012). The CCDF policies database book of tables: Key cross-state variations in CCDF policies as of October 1, 2011 (OPRE Report 2012-51). Office of Planning, Research, and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services.</bibtext> </blist> <blist> <bibtext> National Survey of Early Care and Education Project Team. (2013). National Survey of Early Care and Education: Summary Data Collection and Sampling Methodology OPRE Report #2013-46. Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services. https://<ulink href="http://www.acf.hhs.gov/sites/default/files/documents/opre/nsece%5fsummarymethodology%5ftoopre%5f102913.pdf">www.acf.hhs.gov/sites/default/files/documents/opre/nsece%5fsummarymethodology%5ftoopre%5f102913.pdf</ulink></bibtext> </blist> <blist> <bibtext> NSECE Project Team. (2012). National Survey of Early Care and Education.</bibtext> </blist> <blist> <bibtext> Office of Planning, Research, & Evaluation. (2008). Child Care and Development Fund (CCDF) policies database, 2008-2019. Administration for children and families. https://<ulink href="http://www.acf.hhs.gov/opre/research/project/child-care-and-development-fund-ccdf-policies-database-2008-1013">www.acf.hhs.gov/opre/research/project/child-care-and-development-fund-ccdf-policies-database-2008-1013</ulink></bibtext> </blist> <blist> <bibtext> Office of the Inspector General, U.S. Department of Health and Human Services. (2019). States' payment rates under the Child Care and Development Fund. https://oig.hhs.gov/oei/reports/oei-03-15-00170.pdf</bibtext> </blist> <blist> <bibtext> Pishgar, F., Greifer, N., Leyrat, C., & Stuart, E. (2021). MatchThem: Matching and weighting after multiple imputation. The R Journal. https://journal.r-project.org/archive/2021/RJ-2021-073/</bibtext> </blist> <blist> <bibtext> Rohacek, M. (2012). A Summary of Research on How CCDF Policies Affect Providers. Urban Institute. https://<ulink href="http://www.urban.org/research/publication/summary-research-how-ccdf-policies-affect-providers">www.urban.org/research/publication/summary-research-how-ccdf-policies-affect-providers</ulink></bibtext> </blist> <blist> <bibtext> Rohacek, M., & Adams, G. (2017). Providers in the child care subsidy system: Insights into factors shaping participation, well-being, and quality. Urban Institute.</bibtext> </blist> <blist> <bibtext> Rosenbaum, P. R., & Rubin, D. B. (1985). Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. The American Statistician, 39 (1), 33 – 38. https://doi.org/10.1080/00031305.1985.10479383</bibtext> </blist> <blist> <bibtext> Rudolph, K. E., Stuart, E. A., Glass, T. A., & Merikangas, K. R. (2014). Neighborhood disadvantage in context: The influence of urbanicity on the association between neighborhood disadvantage and adolescent emotional disorders. Social psychiatry and psychiatric epidemiology, 49 (3), 467 – 475. https://doi.org/10.1007/s00127-013-0725-8</bibtext> </blist> <blist> <bibtext> Sandstrom, H., Coffey, A., Henly, J., Bromer, J., Spalding, A., Thomas, W., Greenberg, E., & Derrick-Mills, T. (2018). Learning from child care providers across settings: A critical step to improving the quality and stability of subsidized care. Urban Institute.</bibtext> </blist> <blist> <bibtext> Schulman, K. (2019). Early progress: State child care assistance policies 2019. National Women's Law Center.</bibtext> </blist> <blist> <bibtext> Slicker, G. (2022). Examining Provider Participation in the Child Care Subsidy System: A Mixed Methods Study. https://<ulink href="http://www.proquest.com/openview/2feaf2ff3ce71a19996dbed7ec28f4f2/1?pq-origsite=gscholar&cbl=18750&diss=y">www.proquest.com/openview/2feaf2ff3ce71a19996dbed7ec28f4f2/1?pq-origsite=gscholar&cbl=18750&diss=y</ulink></bibtext> </blist> <blist> <bibtext> Slicker, G., & Hustedt, J. T. (2022). Predicting participation in the child care subsidy system from provider features, community characteristics, and use of funding streams. Children & Youth Services Review, 136, 106392. https://doi.org/10.1016/j.childyouth.2022.106392</bibtext> </blist> <blist> <bibtext> Slicker, G., Hustedt, J. T., & Stoffers, M. (2023). Participation of early care and education centers in the child care subsidy system: A statewide mixed methods investigation. Early Childhood Research Quarterly, 65, 317 – 331. https://doi.org/10.1016/j.ecresq.2023.07.009</bibtext> </blist> <blist> <bibtext> Stuart, E. A. (2010). Matching methods for causal inference: A review and a look forward. Statistical Science, 25 (1), 1 – 25. https://doi.org/10.1214/09-STS313</bibtext> </blist> <blist> <bibtext> Thoemmes, F. J., & Kim, E. S. (2011). A Systematic review of propensity score methods in the social sciences. Multivariate Behavioral Research, 46 (1), 90 – 118. https://doi.org/10.1080/00273171.2011.540475</bibtext> </blist> <blist> <bibtext> U.S. Department of Health and Human Services, Office of Child Care. (2021). FY 2019 Child Care and Development Fund data tables (preliminary). https://<ulink href="http://www.acf.hhs.gov/occ/data/fy-2019-ccdf-data-tables-preliminary">www.acf.hhs.gov/occ/data/fy-2019-ccdf-data-tables-preliminary</ulink></bibtext> </blist> <blist> <bibtext> U.S. Government Accountability Office. (2016). Access to subsidies and strategies to manage demand vary across states. https://<ulink href="http://www.gao.gov/assets/690/681652.pdf">www.gao.gov/assets/690/681652.pdf</ulink></bibtext> </blist> <blist> <bibtext> van Buuren, S., & Groothuis-Oudshoorn, K. (2011). Mice: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software, 45 (3). https://doi.org/10.18637/jss.v045.i03</bibtext> </blist> <blist> <bibtext> Weber, R. B., Grobe, D., & Davis, E. E. (2014). Does policy matter? The effect of increasing child care subsidy policy generosity on program outcomes. Children and Youth Services Review, 44, 135 – 144. https://doi.org/10.1016/j.childyouth.2014.06.010</bibtext> </blist> <blist> <bibtext> West, S. G., Cham, H., Thoemmes, F., Renneberg, B., Schulze, J., & Weiler, M. (2014). Propensity scores as a basis for equating groups: Basic principles and application in clinical treatment outcome research. Journal of Counseling and Clinical Psychology, 82 (5), 906 – 919. https://doi.org/10.1037/a0036387</bibtext> </blist> <blist> <bibtext> Zanoni, W., & Johnson, A. D. (2019). Child care subsidy use and children's outcomes in middle school. AERA Open, 5 (4), 4. https://doi.org/10.1177/2332858419884540</bibtext> </blist> </ref> <aug> <p>By Gerilyn Slicker; Christina Areizaga Barbieri and Jason T. Hustedt</p> <p>Reported by Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib37" firstref="ref2"></nolink> <nolink nlid="nl2" bibid="bib11" firstref="ref3"></nolink> <nolink nlid="nl3" bibid="bib15" firstref="ref4"></nolink> <nolink nlid="nl4" bibid="bib57" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib51" firstref="ref6"></nolink> <nolink nlid="nl6" bibid="bib42" firstref="ref7"></nolink> <nolink nlid="nl7" bibid="bib20" firstref="ref9"></nolink> <nolink nlid="nl8" bibid="bib58" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib19" firstref="ref12"></nolink> <nolink nlid="nl10" bibid="bib60" firstref="ref13"></nolink> <nolink nlid="nl11" bibid="bib29" firstref="ref14"></nolink> <nolink nlid="nl12" bibid="bib32" firstref="ref15"></nolink> <nolink nlid="nl13" bibid="bib39" firstref="ref16"></nolink> <nolink nlid="nl14" bibid="bib17" firstref="ref20"></nolink> <nolink nlid="nl15" bibid="bib44" firstref="ref21"></nolink> <nolink nlid="nl16" bibid="bib25" firstref="ref26"></nolink> <nolink nlid="nl17" bibid="bib23" firstref="ref27"></nolink> <nolink nlid="nl18" bibid="bib40" firstref="ref29"></nolink> <nolink nlid="nl19" bibid="bib30" firstref="ref30"></nolink> <nolink nlid="nl20" bibid="bib14" firstref="ref35"></nolink> <nolink nlid="nl21" bibid="bib47" firstref="ref38"></nolink> <nolink nlid="nl22" bibid="bib50" firstref="ref39"></nolink> <nolink nlid="nl23" bibid="bib46" firstref="ref41"></nolink> <nolink nlid="nl24" bibid="bib24" firstref="ref42"></nolink> <nolink nlid="nl25" bibid="bib53" firstref="ref43"></nolink> <nolink nlid="nl26" bibid="bib54" firstref="ref44"></nolink> <nolink nlid="nl27" bibid="bib41" firstref="ref51"></nolink> <nolink nlid="nl28" bibid="bib43" firstref="ref52"></nolink> <nolink nlid="nl29" bibid="bib28" firstref="ref53"></nolink> <nolink nlid="nl30" bibid="bib36" firstref="ref54"></nolink> <nolink nlid="nl31" bibid="bib18" firstref="ref64"></nolink> <nolink nlid="nl32" bibid="bib52" firstref="ref68"></nolink> <nolink nlid="nl33" bibid="bib48" firstref="ref70"></nolink> <nolink nlid="nl34" bibid="bib35" firstref="ref71"></nolink> <nolink nlid="nl35" bibid="bib56" firstref="ref73"></nolink> <nolink nlid="nl36" bibid="bib26" firstref="ref75"></nolink> <nolink nlid="nl37" bibid="bib59" firstref="ref77"></nolink> <nolink nlid="nl38" bibid="bib27" firstref="ref78"></nolink> <nolink nlid="nl39" bibid="bib45" firstref="ref79"></nolink> <nolink nlid="nl40" bibid="bib13" firstref="ref83"></nolink> <nolink nlid="nl41" bibid="bib55" firstref="ref84"></nolink> <nolink nlid="nl42" bibid="bib61" firstref="ref85"></nolink> <nolink nlid="nl43" bibid="bib16" firstref="ref86"></nolink> <nolink nlid="nl44" bibid="bib31" firstref="ref90"></nolink> <nolink nlid="nl45" bibid="bib49" firstref="ref93"></nolink> <nolink nlid="nl46" bibid="bib12" firstref="ref102"></nolink> <nolink nlid="nl47" bibid="bib34" firstref="ref103"></nolink> <nolink nlid="nl48" bibid="bib38" firstref="ref107"></nolink> <nolink nlid="nl49" bibid="bib21" firstref="ref109"></nolink> <nolink nlid="nl50" bibid="bib22" firstref="ref110"></nolink> <nolink nlid="nl51" bibid="bib10" firstref="ref117"></nolink> <nolink nlid="nl52" bibid="bib33" firstref="ref118"></nolink> <nolink nlid="nl53" bibid="bib62" firstref="ref119"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1420637
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: The Role of State Subsidy Policies in Early Education Programs' Decisions to Accept Subsidies: Evidence from Nationally Representative Data
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Gerilyn+Slicker%22">Gerilyn Slicker</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-0826-6394">0000-0002-0826-6394</externalLink>)<br /><searchLink fieldCode="AR" term="%22Christina+Areizaga+Barbieri%22">Christina Areizaga Barbieri</searchLink><br /><searchLink fieldCode="AR" term="%22Jason+T%2E+Hustedt%22">Jason T. Hustedt</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Early+Education+and+Development%22"><i>Early Education and Development</i></searchLink>. 2024 35(4):859-877.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 19
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2024
– Name: SourceSuprt
  Label: Sponsoring Agency
  Group: SrcSuprt
  Data: Administration on Children, Youth, and Families (ACYF) (DHHS/ACF)
– Name: NumberContract
  Label: Contract Number
  Group: NumCntrct
  Data: 90YE0248
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: <searchLink fieldCode="EL" term="%22Early+Childhood+Education%22">Early Childhood Education</searchLink>
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22State+Aid%22">State Aid</searchLink><br /><searchLink fieldCode="DE" term="%22Grants%22">Grants</searchLink><br /><searchLink fieldCode="DE" term="%22Financial+Policy%22">Financial Policy</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Policy%22">Educational Policy</searchLink><br /><searchLink fieldCode="DE" term="%22Early+Childhood+Education%22">Early Childhood Education</searchLink><br /><searchLink fieldCode="DE" term="%22Family+Needs%22">Family Needs</searchLink><br /><searchLink fieldCode="DE" term="%22Low+Income%22">Low Income</searchLink><br /><searchLink fieldCode="DE" term="%22Poverty%22">Poverty</searchLink><br /><searchLink fieldCode="DE" term="%22Access+to+Education%22">Access to Education</searchLink><br /><searchLink fieldCode="DE" term="%22Child+Care+Centers%22">Child Care Centers</searchLink><br /><searchLink fieldCode="DE" term="%22Child+Development+Centers%22">Child Development Centers</searchLink><br /><searchLink fieldCode="DE" term="%22School+Funds%22">School Funds</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Finance%22">Educational Finance</searchLink><br /><searchLink fieldCode="DE" term="%22Financial+Support%22">Financial Support</searchLink><br /><searchLink fieldCode="DE" term="%22Programs%22">Programs</searchLink><br /><searchLink fieldCode="DE" term="%22Participation%22">Participation</searchLink><br /><searchLink fieldCode="DE" term="%22Income%22">Income</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1080/10409289.2023.2244859
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 1040-9289<br />1556-6935
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Research Findings: Families need access to early care and education programs, both to ensure parents' ability to work and support children's development. Subsidies through the Child Care and Development Fund (CCDF) are a set of policies aimed at assisting families living in poverty with accessing early education. However, the number of early education centers that accept subsidies is declining. Using observational data from the National Survey of Early Care and Education and state subsidy policies from the CCDF Policies Database, we use an innovative application of propensity score methods to estimate causality and provide actionable findings regarding the effect of state-specific policies on centers' subsidy participation. We create comparable groups of centers that accept subsidies and centers that do not, based on research-supported program- and community-level predictors of subsidy participation. Logistic regression models using the matched sample demonstrated state-specific subsidy policies impacted subsidy participation. Specifically, as the state subsidy reimbursement rate increased, centers were more likely to accept subsidies. Practice or Policy: Findings point to meaningful state-level actions and policies that may incentivize centers' subsidy participation. In general, state policies that increase revenue for centers accepting subsidies (e.g. reimbursement for child absences) may result in higher rates of subsidy participation.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2024
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1420637
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1420637
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1080/10409289.2023.2244859
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 19
        StartPage: 859
    Subjects:
      – SubjectFull: State Aid
        Type: general
      – SubjectFull: Grants
        Type: general
      – SubjectFull: Financial Policy
        Type: general
      – SubjectFull: Educational Policy
        Type: general
      – SubjectFull: Early Childhood Education
        Type: general
      – SubjectFull: Family Needs
        Type: general
      – SubjectFull: Low Income
        Type: general
      – SubjectFull: Poverty
        Type: general
      – SubjectFull: Access to Education
        Type: general
      – SubjectFull: Child Care Centers
        Type: general
      – SubjectFull: Child Development Centers
        Type: general
      – SubjectFull: School Funds
        Type: general
      – SubjectFull: Educational Finance
        Type: general
      – SubjectFull: Financial Support
        Type: general
      – SubjectFull: Programs
        Type: general
      – SubjectFull: Participation
        Type: general
      – SubjectFull: Income
        Type: general
    Titles:
      – TitleFull: The Role of State Subsidy Policies in Early Education Programs' Decisions to Accept Subsidies: Evidence from Nationally Representative Data
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Gerilyn Slicker
      – PersonEntity:
          Name:
            NameFull: Christina Areizaga Barbieri
      – PersonEntity:
          Name:
            NameFull: Jason T. Hustedt
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2024
          Identifiers:
            – Type: issn-print
              Value: 1040-9289
            – Type: issn-electronic
              Value: 1556-6935
          Numbering:
            – Type: volume
              Value: 35
            – Type: issue
              Value: 4
          Titles:
            – TitleFull: Early Education and Development
              Type: main
ResultId 1