Family Income and Early Child Care

The effects of poverty on children may be mediated not only by parents and the home environment, but by the early care and educational environments in which many children spend time. Most American children from all income levels spend significant amounts of time in nonmaternal care during their early years. A 1999 Survey by the National Survey of America's Families shows that 73% of children under age 5 with employed parents were in nonparental child-care arrangements, primarily center-based care and care by relatives (Sonenstein, Gates, Schmidt, Bolshun, 2002).

The work requirements set forth by the 1996 welfare law dramatically increased the number of children who require care while parents are working. Whereas the AFDC-based welfare system was originally intended to allow single mothers to remain at home to care for young children, the new TANF-based system is motivated by the goal of making single mothers economically self sufficient. As a result, low-income parents, especially those of very young children, are working more than ever before (Haskins & Primus, 2002). Under the 1996 law, parents with children under 1 year of age are exempt from the federal work requirements, and states may require parents of infants as young as 3 months to participate in employment activities. As a result of PWRORA, an estimated additional one million preschool-age children entered child-care settings (Fuller & Kagan, 2000). Children's experiences in child care may be one means by which the effects of poverty are exacerbated or ameliorated. These effects depend on the quality, type, and amount of care the child receives. We summarize here what we know about child care for low-income families, using data from studies in which Huston has participated as well as other research.

Income and Child-Care Quality

Child-care quality usually encompasses: (1) the child-provider relationship and child experiences (e.g., amount of verbal and cognitive stimulation, responsiveness, stability); (2) structural and caregiver characteristics (e.g., ratios, group size, caregiver education, physical environment/materials); and, (3) health and safety provisions (Lamb, 1997; Phillips, 1995). High quality care may serve as a protective factor for low-income children, and low quality care may compound other risk factors (Shonkoff & Phillips, 2000).

In the first 3 years of life, most child care occurs in home settings, either with relatives or family child-care providers. Data are now accumulating showing that the quality of these home settings is considerable lower for children from poverty-level families than for children of more affluent families. In a widely publicized Study of Family and Relative Care, infants and toddlers cared for by relatives (primarily grandmothers) and nonrelatives were observed. Children from low-income families received poorer quality care than those from more affluent families (Galinsky, Howes, Kontos, & Shinn, 1994).

The NICHD Study of Early Child Care is a longitudinal study in which Huston is an investigator following 1,364 children from birth through adolescence, and includes observations of whatever child-care settings parents selected for their children. Participants were recruited from hospitals located at 10 sites across the United States in 1991. Interviews with parents, evaluations of children, and observations of both parent-child interaction and children's child-care or school settings were conducted when children were 6, 15, 24, 36, and 54 months old and in first grade. Measures included maternal education (in years), child's race and ethnicity, maternal depressive symptoms, parenting quality scores derived from videotaped observations of mother-child interactions, and observed home environmental quality as measured by the Home Observation for Measurement of the Environment (HOME; Caldwell & Bradley, 1984).

Information about the number of hours, stability, and type of child care was collected from parents by telephone every 3 or 4 months. Children who were in nonmaternal care for more than 10 hours a week were observed in whatever child-care setting their parents used at each of the major assessment periods (6, 15, 24, 36, 54 months). Both structural features (e.g., adult-child ratio, group size, caregiver training) and processes were assessed during an observation covering several hours at each point of assessment. Quality was measured with the Observational Record of the Caregiving Environment (ORCE; NICHD Early Child Care Research Network, 1996, 2000), an instrument that included time-sampled records of the frequency of specific caregiver behaviors and observer ratings of the quality of caregiving on several scales. The ORCE was systematically adapted to be age-appropriate. At 6, 15, and 24 months, composite quality scores were calculated as the mean of five, 4-point rating scales (sensitivity to child's nondistress signals, stimulation of child's development, positive regard toward child, detachment [reflected], and flatness of affect [reflected]) (a ranged from .87 to .89). At 36 months, these five scales plus two additional subscales, "fosters child's exploration" and "intrusive" [reflected], were included in the composite (a = .83). At 54 months, the overall quality of caregiving composite was the mean of 4-point ratings of caregivers' sensitivity/responsivity, stimulation of cognitive development, intrusiveness [reflected], and detachment [reflected] (a = .72).

To ensure that observers at the ten sites were making comparable ratings, all observers were certified before beginning data collection and tested for observer drift every 3-4 months. Agreement with master-coded videotapes and with other examiners in live observations was evaluated using intra-class correlations (i.e., Pearson correlations and the repeated measures ANOVA formulation) (Winer, 1971). Reliability exceeded .80 at all ages.

In the first year of life, home settings for children from low-income families were much lower in observed quality than home settings for children from nonpoor families (NICHD, 1997). Moreover, the cumulative quality of all types of child care over the first 4.5 years of life was lower for children from chronically-poor families than from children in nonpoor families or those who experienced short-term poverty (NICHD ECRN, in press).

On average, child-care centers attended by low-income children are better quality than the home settings their families use (Loeb, Fuller, Kagan, & Carrol, 2004; Votruba-Drzal, Coley, & Chase-Lans-dale, 2004), and are of more-or-less comparable quality to centers attended by children from more affluent families. Most studies have not found strong relations between family income and quality (Phillips, 1995). One reason may be that children from poor families receive slightly better care than those from lower- to middle-income families. In the National Child Care Staffing Study, which investigated a large sample of centers serving infants and preschoolers, centers serving children from moderate-income families provided poorer quality care than those serving children from families with very low incomes or high incomes. High-income families can purchase quality center care and can use child-care tax credits; families with very low incomes have more access to government subsidies for care than do those just above eligibility levels (Phillips, Voran, Kisker, Howes, & Whitebook, 1994; Hofferth, 1995).

Type of Care

Poor families disproportionately use informal, home-based arrangements (Brown-Lyons, Robertson, & Layzer, 2001; Hofferth, Brayfield, Ceich, Holcomb, 1991), and this appears to be the case for most families leaving welfare as well (Schumaker & Greenberg, 1999). Parents at all income levels choose different types of child-care arrangements based on a complex combination of their preferences and the constraints they face. For low-income parents, these constraints are often substantial.

Poor families are less likely than others to use paid care, but when they do pay for care they spend about 20% of their income on care, compared to about 6% for families who are not poor (Smith, 2002). As a result, child care often represents the second or third greatest expense for low-income working families (Isaacs, 2002). Recent studies indicate that a large proportion of families turn to informal, unlicensed child care when trying to fulfill new work requirements because it offers more flexible hours and is less expensive.

Although some low-income parents prefer home-based care, others would choose center-based care if given the economic resources to pay for it (Fuller et al., 2002). The Next Generation Study is an analysis of approximately 20 experimental tests of policies designed to promote employment for single parents with low incomes. Although all parents in the studies, including those in the control groups, had access to some types of subsidies for child care, some of the programs included additional child-care assistance (e.g., reducing bureaucratic barriers, convenient resource and referral, and higher eligibility thresholds). Almost all of the programs increased parental employment and, as a result, almost all increased use of nonparental child care. However, the type of care chosen depended on child-care assistance policies. Parents in programs that provided expanded child-care assistance were more likely to place their children in center-based care; those in programs with standard types of assistance were more likely to put their children in home-based arrangements (Crosby et al., 2004). The reasons appear to be relatively straightforward. In the expanded assistance programs, parents were more likely to use subsidies, and they paid less out of their own pockets for child care. They also reported fewer child-care problems that interfered with their ability to get and keep a job (Gennetian, Crosby, Huston, & Lowe, 2004).

Availability places additional constraints on the types of care families can use. Care for infants and school-age children is often more difficult to find than for preschool-age children. Market supply conditions vary greatly across states and communities. In general, the supply of child care tends be lower in low-income neighborhoods than in higher-income neighborhoods, and is particularly scarce during nonstandard hours and for children with special needs (U.S. GAO, 1997). Only 10% of centers and 6% of family child-care homes offer care on weekends (Phillips, 1995), and fewer offer care during nighttime hours. Low-income parents, especially those that have transitioned from welfare, are much more likely than other parents to work evenings, early mornings, weekends, and rotating or inconsistent shifts, making it difficult to put together reliable, stable child-care arrangements (Hofferth, 1995; Lowe & Weisner, 2003; Mishel & Bernstein, 1994).

Amount of Care

Children in poor families spend less time in nonmaternal care than do those in affluent families because, on average, their mothers work fewer hours. For example, in the NICHD SECC, infants in poor families spent an estimated 14 hours per week less in child care than did children in affluent families; these estimates are adjusted for maternal education, ethnic group, family structure, and sex of the child. With income held constant, children who were in care for the longest hours had mothers who worked many hours and who had relatively low levels of education, but whose incomes were slightly above poverty (NICHD ECRN, 1997). Over the first 4.5 years of life, children in chronically poor families spent approximately 6 hours per week less in child care than did children in families that were never poor, but the amount of child care increased over time for those in chronically poor families as well as those in families that left poverty after the child's first 3 years of life, again suggesting that maternal work was an important means of escaping poverty (NICHD ECRN, in press).

Effects of Child-Care Quality, Type, and Quantity on Children

The primary challenge in examining the causal relationships between child-care characteristics and child outcomes is the fact that child care is selected by parents, not randomly assigned. Families who choose different types of care differ in many ways, making it difficult to disentangle the effects of care on children. Despite methodological challenges, child-care research over the last two decades has provided a fairly consistent and convincing argument that child care matters for developmental outcomes (Lamb, 1997; Vandell & Wolfe, 2000).

One major purpose of the NICHD Study of Early Child Care and Youth Development (SECC) was to understand the relations of child care to children's health and cognitive and social development. Information about overall health, illnesses, and injuries was collected from parents at regular intervals. Children's development of cognitive and language skills was measured at 24, 36, and 54 months and in first, third, and fifth grades. At 24 months, the Mental Development Index from the Bayley II (Bayley, 1993) was used to assess overall developmental status. At 36 months, the Bracken Scale of Basic Concepts (Bracken, 1984) and the Reynell Developmental Language Scales (Reynell, 1991) were administered. From 54 months onward, cognitive, language, and academic performance was assessed with reading, math, and cognitive subtests from the Woodcock-Johnson Psycho-Educational Battery-Revised (Woodcock & Johnson, 1989, 1990) and the Preschool Language Scale (PLS-3; Zimmerman, Steiner, & Pond, 1992).

Mothers' and caregivers' or teachers' reports of the children's behavior problems were obtained using the Child Behavior Checklist (CBCL), administered almost annually from 24 months on. The CBCL contains two broad-band subscales: externalizing problems (acting out, aggression) and internalizing problems (fearfulness, anxiety). A complete list of measures at each age can be found at


High quality care predicts children's academic functioning, including language, reading, and math skills (Phillips & Adams, 2001; Vandell & Wolfe, 2000; see review by Isaacs, 2002). Several longitudinal studies, including the NICHD Study, demonstrate that this relationship persists even after controlling for demographic and parenting characteristics (NICHD ECRN, 2000; 2002). The effects of child-care quality on cognitive functioning and school achievement have been linked specifically to levels of engagement between children and caregivers (Shonkoff & Phillips, 2000). Quality of care may be especially important for children from low-income families. In analyses of the NICHD study (NICHD ECRN, 2001; McCartney, 2003), quality of care predicted language and cognitive development for children from low-income families, but was less important for children from middle and upper income families.

Although quality of care is also related to positive social behavior and lowered behavior problems, the effects on social behavior are less consistent than the effects on cognitive development (e.g., NICHD ECRN, 2003). Other investigations following children over periods of time have shown relations between quality and social-emotional outcomes. For example, Howes (1988; 1990) demonstrated that quality of early care predicted children's levels of play and social competence in the child-care setting and, a few years later, when they reached school age. In the Child Care Costs and Quality Study, 826 four-year-olds were observed in child-care centers in four states. The quality of the adult-child interactions predicted vocabulary, math competence, children's perceptions of their own competence, and teacher ratings of positive behavior and sociability at age 8 (Peisner-Feinberg & Burchinal, 1997; Peisner-Feinberg, Burchinal, Clifford, Culkin, Howes, Kagan, & Yazejian, 2001). In another investigation, the observed quality of care at age 4 predicted children's social and academic adjustment in school at age 8 (Vandell & Corasaniti, 1990).

Much of the information on child-care effects is correlational in nature, so one cannot infer that child care causes outcomes. Some studies have examined changes in children's performance over time as a means of controlling for family and child differences; that is, a child's performance at one time is compared to his/her earlier performance. Two large-scale studies of children from very low-income families show that observed quality of care is associated with cognitive gains (controlling for initial levels) over the preschool years (Loeb et al., 2004; Coley et al., 2004). A similar analysis of the NICHD SECC demonstrated small, but reliable effects of child-care quality (NICHD ECRN & Duncan, 2003).

When naturally occurring child care is studied, the "effects" can be estimated only for the environments that exist, not for environments that might be optimal but rare. If most existing child care is of low quality, then examining effects of quality on child outcomes will not tell us how much good quality could contribute to children's development. There are, however, true experimental tests showing the efficacy of educational child care for children from low-income families during the first few years of life (Barnett, 1995; McLoyd, 1997). For instance, children who received educational child care from infancy to school age in North Carolina performed better than controls on measures of school achievement as late as age 21 (Campbell, Ramey, Pungello, Sparling, & Miller-Johnson, 2002). In the Infant Health and Development Project, low birth weight children were randomly assigned to an intervention that included educational child care from age 1 to 3 or to an untreated control group. The intervention had positive effects on IQ, vocabulary, and incidence of behavior problems at age 3 and 5 (Smith & Brooks-Gunn, 1994). However, the size and endurance of these effects is not yet clear. One meta-analysis demonstrated that well-designed interventions with experimental and control groups had lasting effects on children's school performance; children who had received the interventions were less likely to be retained in a grade or placed in special education classes (Lazar & Darlington, 1982). On the other hand, the positive effects of Head Start may not last if children go on to attend poor quality schools (Currie & Thomas, 2000).

Type of Care

We already noted that children from low-income families generally receive somewhat higher quality care in centers than they do in home settings. Even with quality controlled, however, low-income children showed more gains in cognitive development when they were in center care than when they were in home care (Loeb et al., 2004). Similarly, for children from across the income range, children who received center-based care during the preschool years showed more gains in cognitive performance than did those who did not, even when child-care quality was controlled (NICHD ECRN & Duncan, 2003). In the Next Generation studies of low-income families who participated in experimental welfare and employment programs, we used the experimental impacts of the programs to demonstrate that center-based care, in conjunction with improved family income, during the preschool years led to small improvements in children's school achievement in the first few years of school (Gennetian et al., 2004).

The effects of center-based care on children's social behavior are more mixed. In the NICHD study and in some other studies, there is evidence that children with center-care experience, particularly in infancy, show more behavior problems at ages 4.5 and kindergarten (NICHD ECRN, 2003; in press). Among low-income children, however, one observational study found no positive or negative effects of center care on problem behavior (Loeb et al., 2004). Our experimental analyses in the Next Generation study indicated that, once selection factors are controlled, children with center care experience in preschool are rated by elementary school teachers as having fewer behavior problems than those with other types of prior experience. It seems likely that children can learn aggressive and assertive behaviors when they spend a large amount of time with peers; on the other hand, centers can vary considerably in the ways in which social behavior and social conflicts among children are handled.

Quantity of Care

There is evidence from the NICHD study that children in high amounts of child care from infancy onward have less sensitive interactions with their mothers and, by age 4.5, are seen by teachers as having more externalizing behavior problems than children with less overall child-care experience (NICHD ECRN, 2003). As noted earlier, however, children from low-income families spend less time in child care than do children in more affluent families, so quantity of care is not likely to account for poverty effects on intellectual and social development.


In summary, children from poor families receive less nonmaternal care, and the care they do receive is more likely to be informal and home-based care of relatively low quality than is the case for other children. When they receive center-based care, it is closer in quality to that received by higher income children. Public policies, including subsidies for child care and high-quality preschool educational programs, can enable low-income parents to provide supportive environments for their children. Both center care and quality care contribute to children's cognitive development and academic preparation. Quality care also promotes positive social behavior, but the effects of center care on social behavior are more mixed.

Single Parenting

Single Parenting

Finally! You Can Put All Your Worries To Rest! You Can Now Instantly Learn Some Little-Known But Highly Effective Tips For Successful Single Parenting! Understand Your Role As A Single Motherfather, And Learn How To Give Your Child The Love Of Both Parents Single Handedly.

Get My Free Ebook

Post a comment