The National Institute of Child Health and Human Development (NICHD) Study of Early Child Care, begun in 1990, investigates whether and how much early child care affects infants' and young children's social, emotional, physical and cognitive development. The longitudinal investigation was designed as an ecological study of children from birth through first grade. A cohort of over 1,300 socially and racially diverse children from 10 sites around the United States was identified and has been assessed at multiple age points in various settings (at home, in child care, and in standard tests). The results of the early assessments, at age 6, 15, 24, and 36 months, are now available.
Previous research on the effects of early child care had led to controversy and confusion. The most provocative finding was that when infants were in nonmaternal child care 20 or more hours a week, starting in the first year of life, they were less likely than infants without such experience to form a secure emotional attachment to their mothers. But this finding was open to interpretation. Was the difference robust and substantial? Was the measure of attachment appropriate? Were the children in child care at risk for psychological maladjustment because they lacked a secure relationship with their mothers? The studies that had produced the finding were relatively small, and focused on white middle-class children. They did not take into consideration differences in the type or quality of nonmaternal care that the children received in the mother's absence. They did not follow the children over time to see the continuing effects of child care, nor did they assess other aspects of psychological well-being.
Recognizing the limitations in the available research and the importance of the issue for American children, families and society, the NICHD director, Duane Alexander, M.D., initiated a large scale, comprehensive study of early childhood. It was a collaborative study, involving investigators at more than 10 research institutions as well as NICHD.
The collaborative study was an improvement over previous research in a number of ways. Nearly 9,000 mothers in more than 23 hospitals were screened for participation. Excluded from the sample were mothers under 18, infants from multiple births, mothers or infants who stayed in the hospital more than one week after the birth, and mothers or infants with disabilities, medical problems, or substance abuse. From the eligible families, 1,364 were randomly selected and agreed to participate in the study (approximately 130 at each site).
The families in the sample were diverse. They lived in urban, suburban and rural areas, from Boston to Los Angeles, Seattle to Little Rock. They varied in income, education, ethnicity (24% of the children were ethnic minorities), family structure (14% were single mothers), and mother's employment (six months after the baby's birth, 46% of the mothers worked or were in school full-time 30 or more hours per week; 21% worked or were in school part-time; 33% were not employed or in school). The characteristics of the families in the study were representative of the communities from which they were drawn. However, the mothers in the study were somewhat more likely to be employed than other mothers in the community or in the United States (in the study, 65% of the mothers were employed at 12 months; nationally the figure in 1990 was 53%).
Mirroring national statistics, most of the families in the study used nonmaternal care during the first year of the infant's life. Approximately 80% of the infants in the study experienced some regular nonmaternal child care during their first year. On average, they entered care at 3 months of age, for 29 hours per week. By the time they were 6 months old, half of the infants in the study were in care for at least 30 hours a week (Figure 1).
When infants first entered care, the most common nonmaternal care provider was the father (25%), another relative (23%) or a nonrelated caregiver in a child-care home (24%). Care in a child-care center (12%) or by a sitter (nanny, housekeeper) in the child's own home (12%) was less common. Mothers who were employed only part-time or not at all were relatively more likely to use family members or in-home sitters; full-time employed mothers were relatively more likely to have their children in the more "institutional" settings of child-care homes or centers. The use of these care arrangements was quite stable over the course of the infant's first year for employed mothers, but for mothers who were not working, the likelihood of fathers and other relatives providing care declined by half, and care by unrelated in-home providers and in centers doubled over the year.
The care arrangements experienced by individual infants themselves were varied: only 38% of the infants were enrolled in just a single nonparental arrangement, and of these, one-third returned to parental care at some time during the first year. Over one-third of the infants in nonparental care (36%) were in at least three different nonparental arrangements in their first year. Experiencing such multiple care arrangements was more common for infants whose mothers worked variable hours. Of the infants in more than one nonparental care arrangement, 93% experienced the termination of at least one arrangement in the first year.
Infants who were in a regular nonmaternal care arrangement for at least 10 hours per week were visited there by trained observers, at 6, 15, 24, and 36 months of age. The observers recorded the quality of the physical environment, the number and kinds of people present, and the nature of care provided by the caregivers. They interviewed the caregivers to assess their background characteristics and their attitudes toward child rearing. The quality of care they observed was quite high.
Overall, three-quarters of the caregivers observed in the first year were rated as moderately or highly sensitive to the demands and needs of the infants, and expressed positive feelings toward them. Fathers, grandmothers and in-home providers were particularly likely to be sensitive and positive (more than 80% of them received these high ratings); care providers in centers, who were responsible for taking care of more infants, were least likely to be rated as moderately or highly sensitive (48%). Higher quality caregiving was observed in settings with fewer children, better child-adult ratios, and safer, cleaner and more stimulating physical surroundings. The caregivers providing higher quality care held more progressive (nonauthoritarian) beliefs about child rearing.
In general, the care received by economically poor children who were cared for in home settings by a family member or a child-care home provider was of lower quality than that of other children. Poor children in child-care centers received care equivalent in quality to that received by children from more affluent families. Children just above the poverty level received lower quality care in child-care centers than children in poverty, probably because of the subsidies available to families below the poverty line.
But what effect does being in nonmaternal child care have on children's behavior and well-being? The study included measures of the children's relationships with their mothers, fathers and peers. It included measures of children's adjustment and self concepts, their general intellectual functioning, language skills, and readiness for school. All these were to be analyzed in the contexts of the children's experiences at home and in child care (Figure 2). Although not all the analyses have been completed, three sets of analyses were available.
The first set of analyses demonstrated that the more hours the infants spent in nonmaternal care, the less sensitive their mothers were and the less engaged and affectionate the children were when they were observed playing with toys in front of our video camera. This difference was statistically significant (when other family variables were controlled) at ages from 6 to 36 months, but it was very small in absolute terms. Mothers whose children spent more hours in care were also observed to be more negative toward their children when they were visited at home at 15 months (more likely to shout, slap, spank, criticize, or interfere)-again, a small but statistically significant difference. Mothers whose children were in higher quality child care, on the other hand, were more involved with their children when they were observed with them at home (vocalizing, responding, voicing positive feelings, hugging, kissing, praising) and more sensitive to them in the videotaped interaction; their children were more engaged and affectionate. These differences were also statistically significant but small in absolute terms.
The second set of analyses followed up on these results to see whether these small differences in maternal sensitivity and involvement and children's engagement and affection were reflected in the nature of the child's emotional attachment to the mother. The standard assessment of infants' emotional attachments (Ainsworth's Strange Situation) was administered when the children were 15 months old. No differences in the likelihood of secure attachment were observed to be related to child care-whether the child was in child care, how many hours he or she spent there, how many different care arrangements he or she was in, how old the child was when care began, the type of care (father, relative, in-home, child-care home, center), or the quality of care received. However, where children were already "at risk" because their mothers were less involved and sensitive toward them being in child care, particularly when the child care was of low quality, was related to an elevated rate of insecure attachment (Table). These findings are discussed in more detail in the December 1997 issue of Child Development.
The third set of analyses investigated the effects of child care on children's cognitive and language development. After statistically controlling for the contributions of family income and mother's IQ and cognitive stimulation of the infant, we found that the quality of child care predicted children's cognitive performance, school readiness, and language scores. The most predictive variables were the amount of language stimulation children received from their care providers in the child-care setting and whether they were in a center program or not.
At 15 and 24 months, after adjustments for these family contributions had been made, children in high-quality care (with lots of language stimulation) were about 12 percentage points higher than children in low-quality care in terms of the number of words they used and understood; at 36 months the difference in language production and comprehension was about seven percentage points. Children in high-quality care were also about seven percentage points higher on the Bracken test of school readiness at 36 months. Differences on infant mental tests (the Bayley Scale of Mental Development) at 15 and 24 months, although statistically significant, were much smaller (two percentage points). At 36 months, children in centers were about seven percentage points higher on school readiness than children in child-care homes; they were about eight percentage points higher than children in any kind of home care (including home care by mother) in their language comprehension.
This is a study still in progress. There are more data to collect, and more analyses to complete. Results to date indicate that across the first three years of life, children's progress is predicted more strongly by family factors than by the characteristics of child care. Nevertheless, because of the small but significant contributions child care makes, above and beyond the family, it merits continued scrutiny and assessment.