Supporting parents through Head Start-child care center partnerships.

Lim, Youngok ; Schilder, Diane ; Chauncey, Ben 等

Abstract

Partnerships between child care centers and Head Start can meet the increased child care needs of low-income parents that resulted from the welfare reform in 1996 and improve children's school readiness by providing full-day, full-year, and high quality child care services. They can also provide comprehensive services for low-income parents such as job training classes and employment referral services that will enhance parents' productivity and ease job searches. Using data collected from parents in Ohio (N=1,605), we estimate the probability of a parent selecting a child care center partnered with Head Start based on several parent characteristics. We find that parents in job training programs, in school, searching for a job, and working long hours are more likely to choose partnership centers. Next, we examine what types of family comprehensive services are offered through Head Start and child care partnerships. We find that parents of children in partnership centers are more likely to receive information about employment enhancement services than parents of children in unpartnered centers. Moreover, the spillover effects of employment enhancement services suggest that the benefits of such services extend to a larger population. These Head Start-child care center partnership services help low-income families become self-sufficient, a goal that cannot be achieved through child care subsidies alone. Not only do low-income working parents benefit, but communities and the wider economy as well.

Introduction

In 2002, 63 percent of the 19 million children under 5 years of age were enrolled in some form of regular child care during a typical week (Johnson, 2005). When we look at primary child care arrangements for children under age five, 24.3 percent were cared for at organized facilities such as day care centers, nursery schools and Head Start programs, 24.8 percent were cared for by relatives and another 17.2. percent were cared by non-relatives (2) (Johnson, 2005). However, this distribution of types of care shifts when we look at families in poverty. Families in poverty with employed mothers rely more heavily on care by a relative or father (60 percent) than on child care centers (14 percent) or family child care homes (7 percent), because they cannot afford the latter options. Moreover, about 22 percent of children under age five with employed mothers have multiple arrangements so that the mother can work. These statistics describe challenges that working families, particularly low-income working families, are facing today.

In 2002, in order to ease these burdens on low-income working families, the federal government devoted 11.2 billion dollars to child care and 6.5 billion dollars to Head Start (Center for Law and Social Policy, 2007). Simply increasing spending, however, may not solve all the challenges faced by low-income working families today. For example, low-income working parents who are eligible for Head Start programs and want to have Head Start services for their children may not be able to do so because Head Start programs are typically running on a part-day and part-year basis and working parents need child care arrangements full-day and full-year. As a result, children who could benefit the most from the enriched learning environments and comprehensive services of Head Start are not receiving them.

One way to address this issue is to encourage partnerships between Head Start and child care providers to jointly deliver services. Partnerships between Head Start and child care providers allow children from low-income working families to receive learning enrichment services and comprehensive services from Head Start while enabling their parents to work full-time. Such partnerships also offer potential economic benefits for parents. They can increase their labor force participation and their productivity if stable, high quality child care is available on a full-day and full-year basis (Carillo, 2004; Shellenback, 2004; Abt Associates, 2002). Partnership benefits to children are access to high quality learning environments and comprehensive services such as developmental screenings and health care referrals. Moreover, such partnerships offer parents the chance to enhance their self-sufficiency. These types of comprehensive services can yield positive parents outcomes.

Although these significant benefits to children and parents can be expected from Head Start and child care provider partnerships, there is no research to date that has examined such partnerships closely. This paper intends to remedy this by exploring the family comprehensive services provided through partnerships with child care centers to enhance low-income parents' self-sufficiency. Using data collected from parents in Ohio, we aim to fill existing holes in the current literature by examining the following two questions: (1) what are the sociodemographic characteristics of parents associated with selecting a Head Start and child care center partnership? Are there differences in selection patterns for low-income parents? (2) What types of family comprehensive services, especially services related to employability, are provided to parents by partnership centers? Are there spillover effects? The next section provides an overview of Head

Start-child care partnerships, followed by a description of Ohio's early childhood education policy. We review existing literature in section 5 and discuss a theoretical approach to examining the potential economic benefits of partnerships. In sections 6 and 7, we look at data and results. In the final section, we discuss needs for future research in this area.

Overview of Partnerships between Head Start and Child Care Providers

Head Start programs and child care subsidy programs were designed to support children from low-income families. Although these two programs share a common purpose, there exist several noteworthy differences between them. Originally, child care subsidies were allocated to support low-income parents' participation in the labor force, while Head Start programs were developed to promote low-income children's development and school readiness. In terms of program design, the child care subsidy system is a demand-side system, while Head Start more closely resembles a supply-side system. The federal government gives the Child Care Development Fund (CCDF), a block grant, to states to be used for child care subsidies. States have considerable flexibility in setting regulations for these subsidy programs. Head Start funding, on the other hand, flows directly from the federal government to local programs.

These differences between Head Start and CCDF may pose challenges for low-income working families and their children. While Head Start programs provide learning enrichment programs and other comprehensive services for children and their low-income parents, they are rarely a viable option for these families because Head Start usually runs on a part-day and part-year basis. This became a more serious issue after the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) (3)--typically known as welfare reform.

Partnerships between Head Start and child care centers have increased in response to changes in the needs of low-income families. With the development of partnership centers, these families' children can receive Head Start services in addition to child care from a single provider, thus avoiding the need to be transported back and forth. At the same time, parents can work, participate in job training, or attend school without interruptions due to problematic child care arrangements. A child in a partnership center often finds himself or herself in a "mixed classroom", where both Head Start and non-Head Start children are placed together and where both Head Start performance standards and child care licensing regulations are met. As a result, a child in a partnership center is more likely to be cared for by a better qualified teacher in a classroom with better teacher-to-child ratios (4). Head Start and child care center partnerships offer full-day and full-year (5) child care services while parents work. Additionally, they have three other unique benefits: an education component for children, comprehensive services for children, and family comprehensive services for parents.

In order to provide continuous child care service for children and their working parents, partnerships typically blend funding from Head Start with other sources such as child care subsidies, parent fees, or both. Since children eligible for Head Start are usually eligible for child care subsidies, a partnership typically involves blending funding from Head Start with Child Care Development Fund (CCDF) (i.e. child care subsidies).

In this paper, we focus on partnerships between Head Start and child care centers. We narrowly define a partnership as a formal agreement, usually a contract, between a Head Start agency and a child care center, entered into with the aim of jointly delivering services to eligible families. However, Head Start and child care partnerships can take various forms. For example, it is possible to have an organization such as Head Start offer extended hours (6) of service by blending their own funding with child care subsidies or contracting with a family child care provider to serve Head Start-eligible children. These types of partnerships are less common than those between Head Start and child care centers. Since partnerships between Head Start agencies and child care centers are usually based on a formal written contracts, discontinuation of partnerships can occur when contracts end and are not renewed. Likewise, new partnerships can form when contracts are established between Head Start and previously non-partnered centers.

In this paper, we focus on the family comprehensive services for parents provided through Head Start and child care center partnerships. Head Start performance standards specify that Head Start help parents set family goals, and assist them with necessary services, referrals and other supports. Responding to each family's goals and needs; Head Start programs offer referrals, services and resources providing opportunities for continuing education and employment training and other employment services through formal and informal networks. Given Head Start performance standards, child care providers partnering with Head Start are likely to provide referrals or services to help parents became employed and stay employed, as indicated in their family goals. In contrast, a typical child care center is not likely to provide such services because licensing regulations do not specify delivering such services for parents. Hofferth and Kisker (1994) report that 2 percent of independent-for-profit centers offer physical exams for children, while 75 percent of Head Start programs provide such exams. They also report that 89 percent of Head Start programs offer psychological testing for children when 8 percent of independent-for-profit centers offer them for children. This report suggests that targeted population is more likely to access comprehensive services when the provision of comprehensive services is mandated by Head Start regulations. Although this report examines comprehensive services for children, it may still hold relevance to comprehensive services for parents.

Family comprehensive services vary widely and commonly include health care referrals, employment counseling, and adult literacy education. It is worth mentioning that Head Start programs typically do not offer these services directly but rather provide parents with referrals to other community agencies to promote low-income parents' self-sufficiency. To understand which types of family comprehensive services are provided through partnership, we examine data we collected from parents in Ohio in this paper.

1. Description of Ohio Policy

We selected Ohio as our study state for examining partnerships between Head Start and child care centers because in 1997 Ohio established state-wide initiatives and provided state funding to those Head Start programs that agreed to partner with local early care and education providers (Verzaro-O'Brien and Scott, 1999). Due to this incentive and encouragement from policy-makers, Ohio provides many examples of partnerships between Head Start and child care centers.

In addition, Ohio licensing regulations on structural indicators of quality, such as group size and ratios, placed it in the middle range of states nationwide (NCCIC, 2006). Ohio also placed in the middle range nationwide in terms of its number of child care centers--it had about 3,800 licensed centers in operation. Nationally some states have fewer than 1000 licensed centers and others have more than 10,000 (NARA, 2006). Nationally all Head Start programs follow one set of performance standards.

To be eligible for Head Start programs, a family's total annual income before taxes must be at or below the official federal poverty line (7) (U.S. Dept. of Health and Human Services, Office of Head Start 2007). Note that eligibility for participation in Head Start is not conditional to parents' working status or job training status. Policies for the child care subsidy program in Ohio, on the other hand, have made child care subsidies available to families at or below 185 percent of the official federal poverty level who meet employment and training participation requirements since 1997 (Karolak, 2002).

The Child and Adult Care Food Program (CACFP), funded by the U.S. Department of Agriculture (USDA), is administered to subsidize the cost of meals and snacks to eligible children at licensed centers, Head Start, family child care providers, and afterschool programs. In Ohio, children from families with incomes at or below 185 percent of the federal poverty level are eligible for free or reduced-price meals (Karolak, 2002).

The conditions described above ensured that we would have enough samples of partnerships and also that our findings would be generalizable and comparable to other states.

2. Literature Review

Although there is little research available that addresses the relationship between Head Start and child care center partnerships, some research relevant to the focus of this study does exist.

There are some studies on the effectiveness of two generation programs such as Even Start, Head Start Family Service Centers (FSCs), and New Chance. Most of these studies, however, focus on components of early childhood education for children and parenting education for parents. There are only a handful of studies examining the third component of two generation programs--adult education for parents to enhance parents' economic self-sufficiency. In terms of the short-term effects of adult education and job training, both the Even Start study and New Chance study found that such education and training had positive results as measured by GED certifications attained (Games et al, 1997; Quint et al, 1997; Pierre, Layzer & Barnes, 1995). Unfortunately, the higher education attained was not shown to translate into stable employment or higher earnings in the long run (Games et al, 1997; Pierre, Layzer & Barnes, 1995). Bernstein, Swartz, and Levin (2000) compared Head Start Family Service Centers (FSC) (8) to traditional Head Start services. They found higher levels of adult participation in educational programs and employment services, although participation in these activities was not shown to lead to higher literacy rates or greater levels of employment. It is possible that the adult education and job training services offered by FSCs were not intense enough to result in significant differences between these programs (Pierre, Layzer and Barnes, 1995).

Other groups of research relevant to this study offer more positive results. Oyemade, Washington, and Gullo (1989) found that low-income parents obtained higher education, were more often employed, and earned higher incomes after they participated in Head Start programs. Allen and Larson (1998) also evaluated comprehensive, individualized services for families through family child care programs. Their results showed evidence of an increase in women's participation in education programs and employment. Tallaferro (2005) suggested that Comprehensive Support Services (CSS) could be effective in providing services to families that would lead to employment. Grossman and Roberts (1989) support the effectiveness of subsidized employment and training programs, while job-search assistance programs were not effective.

Clearly, results from previous research are mixed regarding the relationship between employment outcomes and comprehensive services intended to enhance employability. Note that none of the previous studies examined Head Start and child care center partnership programs.

3. Theoretical Model for the Potential Benefits of Head Start and Child Care Center Partnerships

To analyze the benefits of Head Start and child care center partnerships holistically, we looked at partnerships from the perspective of Bronfenbrenner's (1979) ecological theory. The three ecological contexts examined were: the children's personal development (microsystem); the parents' degree of well-being, which constitutes the child's direct environment (exosystem); and the regional community, which constitutes the child's macro-environment (macrosystem).

Microsystem: Benefits of partnerships can be analyzed at the level of the child. Partnerships can benefit low-income children by promoting their cognitive and emotional development with enriched learning environments.

Exosystem: Partnership benefits can also be analyzed at the level of the parent. Head Start and child care center partnerships can benefit low-income working parents by providing adult education and other employment related services that may lead to the enhanced economic self-sufficiency of parents. In addition, partnerships can benefit parents by providing parenting education. These benefits to parents are likely to lead to benefits to children in the form of better home learning environments, improved parent-child interactions, and parents' increased involvement with their children's schools.

Macrosystem: Lastly, partnerships can benefit local communities by supporting working parents' needs for child care, by improving low-income parents' employability or earnings, and by reducing welfare dependency.

Ecological theory clearly demonstrates potential partnership benefits at multiple levels. In this paper, we focus on the benefits to parents, especially family comprehensive services, such as adult education, and other employment related services.

4. Data: Sampling of child care centers

We collected data from a sample child care centers in Ohio using mail surveys in 2002, 2003 and 2004. Data was collected data across multiple years to examine changes in centers, teachers and parents over time that can be attributed to the Head Start--child care partnership.

In 2001, there were a total of 3,492 licensed centers in Ohio, of which, 468 were partnership and 3,024 were non-partnership. We had randomly drawn our sample from licensed centers with stratification by partnership status and locale. Selected centers were contacted by telephone, asked initial screening questions, (9) and recruited. Based on the screening calls, we mailed out survey questionnaires to a total of 221 child care centers that provide full-day, full-year care and accept child care subsidies. (10) Of these selected centers, a total of 141 agreed to participate, a response rate of 64 percent.

The number of un-partnered centers in 2003 is greater than the number of un-partnered centers in 2002 because 17 partnership centers in 2002 discontinued their partnership and completed the survey as un-partnered centers in 2003. Reasons for discontinuing a partnership, that often mentioned by directors, are related to eligibility issues--not having a sufficient number of children who are eligible for partnership services. Note that it is also possible to lose partnership status if a child care center fails to meet Head Start performance standards, although this is not often mentioned by directors as reason for discontinuing a partnership.

To minimize attrition rates, we contacted centers multiple times and we also provided nominal financial incentives to encourage their participation. Despite our efforts, some attrition occurred the three-year period. For those who dropped out of the sample over time, we conducted a simple analysis to examine if there were significant differences between the centers that dropped out versus those that remained in our sample. We found there were no significant differences in total enrollment, demographics of the population or locale.

5. Data Collection from parents

In this section, we discuss details of data collected from parents in child care centers that participated in our study. As previously mentioned, we also surveyed directors and teachers from centers in our sample. Due to difficulties in obtaining personal contact information of parents, parent surveys were distributed to parents of children preschool aged (11) children enrolled in the center participating by the center directors. This method might have an adverse effect on the overall parent response rate. We sent out a total of 3,141 questionnaires to parents and received only 738 surveys in 2002, which gives us a 23.5 percent response rate. For 2003 and 2004, total sample sizes are 699 and 254 respectively. Response rates for 2003 and 2004 are about 13% and 10% respectively.

To increase response rates, we contacted directors multiple times and also provided a nominal financial incentive to those parents who completed surveys. Low response rates in the parent sample raise some caution in interpreting results from analyzing parent data as this smaller sample may not be representative and may limit the ability to generalize our findings. For instance, if parents who completed surveys are more likely to be actively involved in their children's school, data from this sample of parents are likely to overestimate access to services referred by child care centers. Although we were aware that nonresponse analysis can address concerns of low-response rates and potential bias in the sample of parents, it was not feasible for us to collect information about parents who did not participate in our survey. Due to this limitation, we are cautious in interpreting the results from parent data.

Data from parents is mostly repeated cross-sectional although there is a small sub-sample of parents who responded in more than one year. We included all the observations available to increase sample size using person-period format. Later in the analysis, we account for clusters within person by relaxing the assumption of independence within the same person.

Due to typical "mixed" classroom settings in the partnership centers, we categorize parents into three groups for analysis: (1) Head Start parents in partnering centers; (2) non-Head Start parents in partnering centers; (3) Non-Head Start parents in un-partnered centers. Analyses, such as the access to family comprehensive services, were conducted using these classifications.

6. Measures

Parent surveys included questions about socio-demographics, services/information received from centers, and employment related issues.

Socio-demographic characteristics: Three measures--marital status, the level of highest education and monthly household income--were asked and analyzed. To learn about parents' marital status, we asked "What is your current marital status?" and gave parents six choices to check off only one choice--(1) Single; (2) Not married, living with a partner; (3) Married and living with my spouse; (4) Separated or living apart from my spouse; (5) Divorced; (6) Widowed. After coding all six options as independent responses, for analysis purposes, we collapsed six options into two categories: 'Live with a spouse or partner'--(2) or (3)--versus the rest of others--(1), (4), (5) or (6)--who do not live with spouses or partners.

For the levels of education, we asked "What is the highest level of education you have completed?" with the following response options: (1) some school, but no diploma; (2) High school diploma or GED; (3) Trade license or certificate; (4) Associate's degree; (5) Bachelor's degree; (6) Graduate degree. Again, for analysis, we collapsed these six options into three categories such as 'High school graduate or less'--(1) or (2); 'AA or some college'--(3) or (4); 'BA or above'--(5) or (6).

For monthly household income, we asked "What was the total income for your household last month?" with seven options--(1) Less than $250; (2) $250--$499; (3) $500-$999; (4) $1,000--$1,499; (5) $1,500--$1,999; (6) $2,000-$2,499; (7) Over $2,500. Then we collapsed these seven options into three categories--'Less than $1,500'; '$1,500-$2,499'; '$2,500 or above.'

Services reported by parents: We asked parents "Does the center give you information about these services?" and the following: Health care services; Head Start; Mental health services; English proficiency classes; Adult education or training; GED preparation; Employment services; Immigration services; Food stamps; Financial aid for school; Housing assistance; Parenting, Healthy marriage; Social services; Legal services; Energy/fuel assistance; Dental services; Transportation. (12) For each item, parents could mark "Yes," "No" or "Not Applicable".

Employment-related questions: We asked "How many jobs do you currently have?" and parents either marked the choice "Not currently working" or fill out the number of jobs they were working directly in the blank. Responses of "Not currently working" were coded as zero. As a separate question, we asked "How many hours per week do you work?" and allowed parents to respond "Not currently working" or to fill out the total number of hours. As a follow-up question, we asked "Are you in school or job training?" and parents chose either "Yes" or "No".

For job search activity, we asked a simple question "Are you looking for a job?" and parents could choose from one of three choices--"Yes", "No" or "Already have a job".

7. Results

Data collected from parents over a three year period has total 1,691 person-period observations. However, missing values in some key variables such as Head Start enrollment status, monthly household income and the variables related to employment status reduce the number of observations used in the analysis to a total of 1,305 person-period observations. These 1,305 person-period observations are from a total of 1,193 persons (parents) over three years. Only 105 parents, 9 percent of 1,193 parents, are observed more than one year. We use person-period data format for the analysis to increase the sample size. Employing clustered statistical analysis technique we estimate robust standard errors to account for observations clustered within person.

7.1. Socio-Economic Characteristics of Parents

Table 3 presents the overall characteristics of parents in the sample and in a subgroup of parents--whose monthly household income is less than $1,500 dollars. Note that the federal poverty threshold for a family of four with two children in 2002 is $18,244 dollars, which is approximately $1,520 dollars per month (Census Bureau, 2004). Therefore, this subgroup--whose monthly household income is less than $1,500 dollars--implies a low-income group close to the federal poverty level. In this paper, we refer to this group as low-income parents. This low-income group is likely to meet the income eligibility for Head Start programs. Also note that the median household income for Ohio in 2002 is $42,684 dollars, which is about $3,500 dollars per month (US Census Bureau, 2004; US Census Bureau, 2007).

Table 3 shows that 53% of the parents in the sample have a child enrolled in a partnership center, while 47% have children attending an un-partnered center. Due to "mixed" classroom structure in partnership centers, parents are divided into three subgroups: (1) Head Start parents at partnership centers (18%); (2) non-Head Start parents at partnership centers (35%); (3) non-Head Start parents at un-partnered centers (47%)

For the low income subsample--parents earn less than $1,500 dollars per month, about 35% of low-income parents are in un-partnered centers; 20% are in partnership centers without Head Start services; 45% are receiving Head Start services at partnership centers. This shows that not all of low-income subsample is served by Head Start programs.

Table 3 also reports the average number of jobs parents are currently working and the average number of hours worked per week. Parents who do not currently work are coded as zero for the number of jobs which explains why the average number of jobs that parents are currently working is less than one. Similarly, parents who do not currently work are coded, again, as zero, for the number of hours worked per week.

7.2. Socio-Economic characteristics of parents associated with selecting a partnership center

To examine which characteristics of parents are likely to select a partnership center, we employ a probit model and estimate the probability of selecting a partnership center for all and also for a subgroup of low-income parents separately. To account for some observations measured repeatedly, we relax the assumption of independence to allow correlations within person and report the robust standard error in Table 4. Table 4 reports marginal effects--the changes in the probability for an infinitesimal change in each independent covariate.

Table 4 presents that the probability of being in a partnership center is decreased by 0.0685 (z=-1.86) for parents with an Associate degree or some college, and by 0.1204 (z=-2.93) for parents with a BA degree or above, compared to parents with a high school diploma or below. The probability that parents earning more than $2,500 dollars per month select into a partnership center is lower (dF/dx= -0.0827, z=-1.65) than parents earning less than $1,500 dollars per month, although this is marginally significant (p=0.099). White parents are less likely (dF/dx= 0.2099, z=-5.06) to choose a partnership center than non-white parents. Table 4 indicates that the levels of education, monthly household income and being a 'white' are negatively associated with the probability of selecting a partnership center.

For characteristics of parents that increase the probability of selecting into a partnership center, they include number of working hours per week, attending in job training or in school, and searching for a job. For example, the probability that parents in job training or in school select into a partnership center increases by 0.1516 (z=3.73), compared to parents who are not in training or in school. Parents who are currently searching for a job are more likely (dF/dx= 0.0894, z=1.74) to be in a partnership center than parents who are not searching. One additional hour of working slightly increases (dF/dx=0.0024, z=1.74) the probability of selecting a partnership center.

The analysis for the low-income parents reveals a similar pattern. Being a 'white' and having an Associate degree lower the probability of selecting a partnership center by 0.2580 (z=-4.32), and by 0.2095 (z=-3.61) respectively. Marital status seems to be associated with the probability of choosing a partnership center for a lowincome group, but not for all. The probability that parents who are either married or living with a partner select a partnership is lower by 0.2227 (z=-3.27) than parents who are not. Being in job training or in school increases the probability of selecting a partnership center by 0.1568 (z=2.55).

7.3. Types of Family Comprehensive Services provided by partnership centers

Head Start performance standards specify a collaborative process with families to identify needs and goals of the family and to assist the family by providing necessary support and comprehensive services. Therefore, Head Start programs provide referrals for comprehensive services not only for children but also for parents. In this section, we focus on family comprehensive services for parents.

To understand what types of family comprehensive services are referred to parents by partnership centers and also whether or not un-partnered centers provide referrals for any type of family comprehensive services to parents as well, we asked parents to indicate the types of family comprehensive services they received through their child care centers. Due to "mixed" classroom settings, some partnering centers may offer referral for comprehensive services to all children and families at the center including non-Head Start children and families, if some services do not incur any additional costs in providing the information or services to non-Head Start families. Therefore, non-Head Start children and their parents at partnering centers may receive benefits from Head Start. Thus, it is possible to have some spillover effects for those parents at partnering centers who are not enrolled in Head Start.

Table 5 describes various types of family comprehensive services reported by three subgroups of parents--(1) Head Start parents at partnering centers; (2) non-Head Start parents at partnering centers; (3) Non-Head Start parents at un-partnered centers. Table 5 presents the proportion of parents who reported receiving the information on a specific service. Then we conducted simple t-tests to examine whether or not there exist significant differences between subgroups of parents--especially between (1) Head Start parents at partnering centers and (3) non-Head Start parents at un-partnered centers to see if partnership centers are indeed engaged in providing information about family comprehensive services and to see what types of services are reported. We also conducted simple t-tests between (2) non-Head Start parents at partnering centers and (3) non-Head Start parents at un-partnered centers to examine if there are spillover effects in providing comprehensive services, which is often indicated by anecdotal evidence.

Generally, Table 5 verifies that Head Start parents at partnering centers are receiving more information about various types of family comprehensive services and this is significantly different from non-Head Start parents at unpartnered centers. Two the most common types of referrals aside from Head tart appear to be parenting education (69.7%) and health care services (64.0%) for Head Start parents in a partnership center. Services related to employability such as adult education or training, GED preparation, English classes or employment services are not as common as parenting or health care services. However, results from t-tests indicate that those services are more likely to be offered to Head Start parents in partnering centers than non-Head Start parents at un-partnered centers. For instance, 40.4% of Head Start parents at partnering centers reported that their centers offered the information about adult education or training, while only 7.3% of non-Head Start parents at un-partnered centers reported that they received such information (t=-9.66, p<0.0001). 'Providing transportation' is the only item yielding no statistically significant difference between Head Start parents at partnering centers and non-Head Start parents at un-partnered centers (t=-0.27, p=0.7896).

As previously mentioned, it is possible that non-Head Start parents at partnership centers may receive benefits of spillover effects. Table 5 suggests that spillover effects indeed occur in partnership centers. Compared to non-Head Start parents in un-partnered centers, non-Head Start parents in partnership centers are more likely to access the information about parenting (49.2%), mental health services (17.1%), and financial aid for school (16.5%). These differences in mental health services (t=-2.74), and financial aid for school (t=-3.34) are statistically significant at the 1% level, while the difference in parenting (t=-2.03) is statistically significant at the 5% level. For services related to employability, 15.6% of non-Head Start parents at partnership centers reported to receive the information about adult education or training while 7.3% of non-Head Start parents at un-partnered centers reported so (t=-4.17, p<0.0001). 5.4% and 7.6% of non-Head Start parents at partnership centers reported they received the information about GED preparation and employment services respectively. These percentages are significantly higher than non-Head Start parents in un-partnered centers for GED preparation (2.4%, t=-2.44, p=0.015) and employment services (3.2%, t=-3.05, p=0.0024). In Table 3, nearly 20% of families whose monthly income is less than $1,500 dollars are not in Head Start programs although they send their children to partnership centers. Therefore, spillover effects may benefit those non-Head Start families in partnership centers.

8. Discussion

Past research has focused primarily on whether or not Head Start has a positive effect on preparing children from low-income families for school. These studies do not explore the benefits of family comprehensive services for parents, a component of Head Start, and the positive effect that such services may have on low-income parents' self-sufficiency. Therefore, in our research, we focus on family comprehensive services provided to parents through Head Start and child care center partnerships, especially those related to parents' employment.

The results from probit estimations reveal that parents in job training or in school are more likely to select a partnership center for their children, while parents with higher education or a higher income are less likely to choose partnership centers. When we examine the types of family comprehensive services referred by partnership centers, we find that parents with children in partnership centers received significantly more information about services related to employment than parents with children in un-partnered centers. We also find spillover effects in providing information about family comprehensive services on non-Head Start parents at partnership centers. Given the fact that a substantial proportion of low-income parents have children who are not in the Head Start program but are nevertheless in child care centers partnered with Head Start, spillover effects are encouraging and imply an even larger potential benefits from Head Start and child care partnership.

However, these findings should be cautiously interpreted due to several limitations. There were very low response rates from the parents whose data was used for the analysis. Also, because the surveys were administered by the directors of the child care centers, we cannot exclude the possibility that parents over-reported services they received from centers. It is also possible that parents who filled out surveys participate more actively in Head Start programs and activities, thus leading to an over-estimation of family comprehensive services provided to parents through partnership centers.

Despite the potential benefits from partnerships between child care centers and Head Start, there are several barriers preventing the Head Start--child care center partnership to be more widely adapted. For example, the director of a partnered child care center must navigate complicated rules and regulations governing two different programs as well as center-specific changes necessary for the partnership to work. In addition, partnerships usually complicate the management of a child care center in areas such as staffing and scheduling. From the perspective of Head Start agencies, partnerships make the monitoring and quality assurance of each site more difficult.

Researchers are now calling for comprehensive approaches to the early child care education system as opposed to single aspect approaches (Stoney et. al, 2006). Morrissey and Warner (2007) show that comprehensive early child care education services have economic returns as great or greater than approaches targeting only one aspect of the system. Along the same line, this paper shows the economic benefits of Head Start and child care center partnerships and family comprehensive services provided through partnerships. Furthermore, we present the potential economic benefits that partnerships may bring to communities, to society as a whole, and to low-income parents. In doing so, we look closely at the characteristics of parents who choose partnership centers and examine the types of comprehensive services provided through partnerships.

Unfortunately, this paper cannot provide conclusive evidence about the effects partnerships have on parents and whether or not family comprehensive services enhance their economic self-sufficiency because we lack longitudinal data on parents' labor earnings and employment. It remains for future research to examine the relationship between family comprehensive services and the employment changes of low-income parents

References

Abt Associates. (2002). National report on work and family. Cambridge, MA.

Allen, M. & Larson, J. (1998). Healing the whole family: A look at family care programs. Children's Defense Fund, Washington D.C.

Bernstein, L., Swartz, J. & Levin, M. (2000). Evaluation of the Head Start Family Service Center Demonstration Projects. Executive summary. Washington D.C., US Department of Health and Human Services.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press.

Carillo. C. (2004). A totally new way to think of back-up care. Work & Family Connection. Retrieved March 15, 2007 from http://www.workfamily.com/worklifeclearinghouse/guestcolumns/ cindycarrillo.htm

Center for Law and Social Policy (2007). Combined total child care spending from the Child Care and Development Block Grant (CDBG) and Temporary Assistance for Needy Families (TANF) Block Grant fact sheet. Retrieved from March 15, 2007, from http://www.clasp.org/ChildCareAndEarlyEducation/map102606us.pdf

Duch, H. (2005) Redefining parent involvement in Head Start: a two-generation approach. Early Child Development and Care. 175(1). Pp23-35.

Games, B.C., Conger, D., & McCarty, M. (1997) Follow-up study of families in the Even Start in-depth study: final report. Cambridge, MA, Abt Associates, Inc.

Gennetian, L.A., Huston, A.C., Crosby, D.A., Chang, Y.E., Lowe, E.D., & Weisner, T.S. (2002). Making child care choices: How welfare and work policies influence parents' decision. New York: Manpower Demonstration Research Corporation.

Grossman , J. B. & Roberts, J. (1989). Welfare savings from employment and training programs for welfare recipients. The Review of Economics and Statistics. 71(3) pp.532-537.

Hofferth, S. & Kisker, E. E. (1994) Comprehensive services in child care wettings: prevalence and correlates. Pediatrics. 94, pp.1088-1091.

Johnson, J.O. (2005). Who's minding the kids? child care arrangements: Winter 2002, Current Population Reports, pp.70-101, U.S. Census Bureau. Retrieved on November 1, 2007 from http://www.census.gov/prod/2005pubs/p70-101.pdf

Karolak, E. J. (2002). A study of the child care payment system in Ohio pursuant to Am. Sub. H.B. 283. Retrieved on November 10, 2007 from http://eclkc.ohs.acf.hhs.gov/hslc/Program%20Design%20and%20Management/ Head%20Start%20Requirements/Head%20Start%20Requirements/1304#community

Morrissey, T.W. & Warner, M.E. (2007) Why early care and education deserves as much attention, or more, than prekindergarten alone. Applied Developmental Science. 11(2): 57-70.

National Child Care Information Center (2006). Child care center licensing regulations. Retrieved on November 11, 2006 from http://nccic.acf.hhs.gov/pubs/cclicensingreq/ratios.html

National Association for Regulatory Administration (2006). The 2005 child care licensing study. Retrieved on November 15, 2007 from http://nara.affiniscape.com/displaycommon.cfm?an=1&subarticlenbr=148

Oyemade, U. J., Washington, V. & Gullo, D. (1989). The relationship between Head Start parental involvement and the economic and social self-sufficiency of Head Start families. The Journal of Negro Education, 58(1) pp.5-15.

Peirre, R. G., Layzer, J. I. & Barnes, H. (1995) Two-generation programs: design, cost, and short-term effectiveness. The Future of Children, 5(3) pp.76-93.

Quint, J.C., Bos, J.M. & Polit, D.F. (1997). New Chance: final report on a comprehensive program for young mothers in poverty and their children. New York, MDRC.

Shellenback, K. (2004). Child care and parent productivity: Making the business case. Ithaca, NY: Department of City and Regional Planning, Cornell University.

Schilder, D., Kiron, E., & Elliott, K. (2003, February). Early Care and Education Partnerships: State Actions and Local Lessons. (Res. Brief Vol. 1, No. 1). Newton, MA: Education Development Center, Inc. Retrieved June 20, 2004, from http://ccf.edc.org/documents/PDF/EDC_ExecBrief.pdf

Stoney, L., Michell, A., & Warner, M. (2006). Smarter reform: moving beyond single-program solutions to an early care and education system. Community Development: Journal of the Community Development Society. 37(2) p.101-115.

Tallaferro, Jocelyn. (2005). Local Welfare Reform: Challenges and Triumphs of the Comprehensive Support Services intervention. Journal of Human Behavior in the Social Environment. 12 (2/3) pp.261-280.

U.S. Census Bureau (2004). Poverty: Poverty Thresholds 2002. Retrieved March 5, 2007, from http://www.census.gov/hhes/www/poverty/threshld/thresh02.html

U.S. Census Bureau (2007). Income: Historical Income Tables--Households. Retrieved March 5, 2007, from http://www.census.gov/hhes/www/income/histinc/h08.html

U.S. Department of Health and Human Services / Administration for Children and Families/ Administration on Children Youth and Families/ Office of Head Start. (2007). Head Start performance standard. Retrieved March 15, 2007, from http://eclkc.ohs.acf.hhs.gov/hslc/Program%20Design%20and%20Management/ Head%20Start%20Requirements/Head%20Start%20Requirements/1304#community

Verazaro-O'Brien M. & Scott, J. A. (1999) The Ohio Head Start-Child Care Collaboration Initiative: Implementing Full-Day/Full-Year Services in 1997-98. Columbus, Ohio.

Youngok Lim, Diane Schilder, Ben Chauncey (1)

(1) Youngok Lim, Ph.D., Diane Schilder, Ed.D., Ben Chauncey are staff at Abt Associates. Funding for this work was supported by ?????. Thanks to the anonymous reviewers for helpful comments.

(2) About 10 percents reported as no regular child care arrangement and 3 percent reported kindergarten as their primary child care arrangement.

(3) The PRWORA replaced Aid to Families with Dependent Children (AFDC), a guaranteed cash assistance program to eligible low-income families, with Temporary Assistance for Needy Families (TANF), a temporary assistance program. TANF is designed to end needy parents' dependence on government benefits by requiring recipients to spend a minimum number of hours each week in one or more allowable work or job preparation activities. Therefore, the PRWORA resulted in a large increase in demand for child care subsidies (Schilder et al., 2003; Gennetian et al. 2002).

(4) Meeting both Head Start performance standards and child care licensing requirements implies that child care providers are supposed to meet the stricter one, which is usually Head Start performance standard in Ohio.

(5) Full-day child care is defined in this paper as providing child care services at least 8 hours per day. Full-year child care is defined in this paper as providing child care services for 12 months.

(6) Here extended hours of service means 8 or more hours of service per day to meet parents' needs for work. There exist some Head Start programs offering full-day services defined as services for 5 hours per day. Even full-day Head Start programs usually cannot fully cover parents' working hours.

(7) For the purpose of eligibility, a child from a family that is receiving public assistance or a child in foster care is eligible even if the family income exceeds the income guidelines. However, at least 90 percent of the children who are enrolled in each Head Start program must meet income eligibility.

(8) Initiated in 1990, Head Start Family Service Centers were developed to help Head Start programs assess needs for the family and address problems of inadequate literacy, low employability and substance abuse by coming up with service place and referral to specific services in the community (Bernstein et al., 2000).

(9) We asked following questions for screening purposes: "How many preschool teachers work at your center?"; "How many preschool children are enrolled at your center?"; "How many preschool children receive a subsidy?"; How many preschool children receive services from Head Start? (if a partnership center).

(10) We tried to match partnership centers with un-partnered centers in terms of their locales (urban, suburban or rural) and proportion of children who receive child care subsidy.

(11) 'Preschool-aged' children is defined as 3 to 5 year-olds before kindergarten.

(12) Among those items, 'social services' was included only in 2003 and in 2004, while 'legal services,' 'Energy/fuel assistance,' 'Dental services,' 'Transportation' were included in 2004 only. All other items were included in all three years, 2002, 2003 and 2004.

Table 1 Number of Centers participated

Year Partnership Centers Un-partnered Centers

2002 78 63
2003 47 66
2004 40 54

Table 2
Number of Parents who participated in surveys

Year Partnership Centers Un-partnered Centers

2002 415 323
2003 311 388
2004 133 121

Table 3
Socio-Economic Characteristics of Parents: Summary of
Descriptive Statistics.

 All Low-Income ([dagger])
 (N=1305) (N=339)

 Mean Mean
 (St. Dev.) (St. Dev.)
Marital Status
 Married or Live with a Partner 66.4% 21.2%
Education
 High school graduate or less 37.4% 60.5%
 AA or some college 31.9% 35.4%
 BA or above 30.7% 4.1%
Monthly Income
 < $1,500 26.0% 100.0%
 $1,500 - $2,499 23.1% --
 $2,500+ 50.9% --
Race
 White 84.4% 72.6%
Enrolled in Head Start 17.5% 45.4%
Enrolled in a partnership center 52.8% 65.2%
Non-Head Start at un-partnered 47.2% 34.8%
 center
Non-Head Start at a partnership
 center 35.3% 19.8%
Head Start at a partnership center 17.5% 45.4%
Number of jobs currently working 95.0% 90.0%
 (0.43) (0.50)
Working hours(/week) 33.02 28.89
 (14.89) (15.98)
In Job training or in school 19.2% 32.4%
Search for a job 9.7% 22.40

([dagger]) Low-income is defined as monthly household income
less than $1,500 dollars.

Table 4
Probit Estimation of selecting into a partnership
center--sandwich estimator of variance

 All Low-Income ([dagger])
 (N=1305) (N=339)
 Marginal Marginal
 Effect Effect
 (Robust (Robust
Selecting a partnership center Std. Err) Std. Err)

Marital Status
 Married or Live with a Partner -0.0074 -0.2227
 -0.0417 (0.0708)
Education
 (Omitted High school
 graduate or less)
 AA or some college -0.0685 -0.2095
 (0.0368) (0.0587)
 BA or above -0.1204 -0.1194
 (0.0409) (0.1536)
Monthly Income
 (Omitted < $1,500)
 $1,500 - $2,499 -0.0529
 (0.0452)
 $2,500+ -0.0827
 (0.0500)

Race
 White -0.2099 -0.2580
 (0.0385) (0.0514)
Number of jobs currently working 0.0222 -0.0321
 (0.0448) (0.0689)
Working hours(/week) 0.0024 0.0008
 (0.0014) (0.0024)
In Job training or in school 0.1516 0.1568
 (0.0406) (0.0595)
Search for a job 0.0894 -0.0321
 (0.0504) (0.0727)

Year (Omitted 2002)
 Year 2003 -0.0816 -0.0889
 (0.0297) (0.0585)
 Year 2004 -0.0008 -0.1165
 (0.0423) (0.0556)

Low-income is defined as monthly household income
less than $1,500 dollars.

Table 5
Types of comprehensive family services provided by
partnership and un partnered centers and spillover effects.

 Non-Head
 Head Start Start parents
 parents at a at a
 partnership partnership
 center center
 N=228 N461

Does your center give
your information
about...? Mean Mean

Health Care Services 64.0% 28.6%
Head Start 95.2% 35.1%
Mental Health Services 41.7% 17.1%
English (Proficiency) 18.0% 4.6%
 Classes
Adult Education or 40.4% 15.6%
 Training
GED Preparation 28.5% 5.4%
Employment Services 24.6% 7.6%
Immigration Services 13.2% 2.6%
Food Stamps 26.3% 8.9%
Financial Aid for 28.1% 16.5%
 School
Housing Assistance 29.4% 6.1%
Parenting 69.7% 49.2%
Healthy Marriage 20.2% 7.6%
Social Services ([section]) 51.9% 17.8%
Legal Services ([paragraph]) 29.0% 5.5%
Energy/Fuel Assistance ([paragraph]) 45.2% 6.8%
Dental Services ([double dagger]) 61.3% 19.4%
The center provides 29.0% 20.8%
 transportation ([double dagger])

 Non-Head
 Start parents
 at an
 un-partnered
 center
 N=616

Does your center give
your information
about...? Mean

Health Care Services 24.5%
Head Start 16.2%
Mental Health Services 11.2%
English (Proficiency) 4.1%
 Classes
Adult Education or 7.3%
 Training
GED Preparation 2.4%
Employment Services 3.2%
Immigration Services 1.1%
Food Stamps 5.5%
Financial Aid for 9.4%
 School
Housing Assistance 2.6%
Parenting 43.0%
Healthy Marriage 10.2%
Social Services ([section]) 15.6%
Legal Services ([paragraph]) 2.4%
Energy/Fuel Assistance ([paragraph]) 3.6%
Dental Services ([double dagger]) 10.8%
The center provides 26.5%
 transportation ([double dagger])

([section]) Social Services: N=108 for Head Start parents at a
partnership center, N=247 for Non-Head Start parents at a
partnership, N=352 for non-Head

Start parents at an un-partnered center.

([paragraph]) Legal services & Energy/Fuel assistance: N=31 for
Head Start parents at a partnership center, N=73 for non-Head
Start parents at a partnership center, N=83 for non-Head Start
parents at an un-partnered center.

([double dagger]) Dental services & Transportation: N=31 for Head
Start parents at a partnership center, N=72 for non-Head Start
parents at a partnership center, N=83 for non-Head Start parents
at an un-partnered center.