OBJECTIVE. We examined the association between type of child care, participation in different types of child care in the year before kindergarten and the likelihood of obesity at the start of kindergarten.
METHODS. Using a nationally representative sample of 15 691 first-time kindergartners from the Early Childhood Longitudinal Study-Kindergarten Cohort, we used logistic regression to estimate the relationship between type of primary child care arrangement and children's likelihood of being obese at the start of kindergarten. Our models controlled for family and child characteristics associated with obesity and choice of child care. To examine differential effects of child care participation for groups at high risk for obesity, we tested interactions between children's ethnicity and income with primary type of child care.
RESULTS. At the start of kindergarten, 12% of the children were obese. Without controlling for other characteristics of children and families, children not in child care were significantly less likely and children in family, friend, and neighbor care were significantly more likely to be obese than children in other primary child care arrangements. White children were significantly less likely and Latino children more likely to be obese than children of other ethnic groups. After controlling for relevant child and family characteristics, children in family, friend, and neighbor care and non-Latino children in Head Start were more likely to be obese than children not in child care. For Latino children, however, participation in some types of nonparental child care had protective effects on their likelihood of being obese.
CONCLUSIONS. Primary type of child care is associated with children's obesity. For Latino children, who are at a greater risk of being obese, participation in nonparental child care seems to have a protective effect. These results suggest that child care settings may be an important site for policy intervention during a crucial developmental period. Efforts to help family, friend, and neighbor caregivers support children's physical health may be warranted.
Since 1980, the prevalence of obesity among 2- to 5-year-olds has increased steadily and is especially high among Latino children.1 Obesity during the preschool years is persistent and predicts overweight and obesity later in development.2,3 In the United States, children begin public education at age 5 or 6 years in full- or half-day kindergarten programs. Whereas some publicly funded child care is available for low-income families, we do not have a uniform system for early education before age 5. Working parents generally choose among several types of child care options from the private market or through informal networks. Although just more than 80% of children spend time in nonparental child care settings during the year before kindergarten,4,5 we know little about how prekindergarten child care environments influence children's likelihood of being obese. In this study, we use the term “obese” to refer to children who are at or above the 95th percentile on the Centers for Disease Control and Prevention (CDC) gender-specific BMI-for-age charts.6 This language parallels the language and definitions used for adults and is recommended by many professionals in the field.7–9 The CDC growth charts, however, refer to these children as overweight.
The preschool period may be crucial in the development of obesity. Typically between the ages of 3 and 7 years, children experience the “adiposity rebound,” a developmental milestone that marks the point at which BMI (a weight-for-height index age referenced for children) starts to increase after reaching its nadir.10,11 Because BMI does not distinguish between fat and lean tissue, the adiposity rebound could be an artifact reflecting developmental increases in height and lean tissue. Longitudinal studies, however, have found that the increases during this period are in body fat, not lean tissue or bone.12,13 Also, the fact that adiposity rebound at an early age (during the prekindergarten years) is a strong and consistent predictor of adult obesity suggests that it reflects something more than a transient deviation from age-normalized BMI values.10,14
As children spend more time in nonparental child care, maternal influence on children's diets diminishes15 and other caregivers play a more critical role in children's nutrition.16 This transition is most striking when children start eating a significant number of meals away from home, as many do in child care settings.17 Differences in preschool settings have been linked to differences in children's physical activity levels,18 a correlate of overweight and obesity in children.19
To date, few studies have examined the influence of child care on children's weight. One study relied on parental reports of children's height and weight and made comparisons between 2 broad categories of care types. The authors found that children in <15 hours/week of child care (a category that included family child care and center-based care) were less likely to be obese than children not in these types of care.20 A study of ethnic differences in early childhood obesity in a low-income urban sample revealed no relationship between children's primary child care arrangement and obesity but did not consider amount of care.21 Another study found a link between children's obesity status and hours of maternal employment; however, although maternal employment correlates highly with use of nonparental child care, child care was not included as a potential mediating variable.22 The relationship between type of child care and children's obesity status has yet to be investigated in a representative sample with direct assessments of children's weight and with more detailed care categories aligned with those used in the early care and education field.
In this study, we examined the following research questions rather than a predefined hypothesis about the influence of child care: Is there a significant association between type of child care participation in the year before kindergarten and children's obesity at the start of kindergarten? Do the associations vary for different subgroups of children? Although we discuss “effects,” we do not ascribe causation, knowing that we may not have accounted for all characteristics that influence choice of child care and children's weight status.
For this study, we used data from the Early Childhood Longitudinal Study-Kindergarten Cohort (ECLS-K).23 The ECLS-K is a nationally representative sample of children who entered kindergarten in the fall of 1998. From kindergarten through eighth grade, information was collected periodically on the children, their households, teachers, and schools. Because this data set is publicly available, we received exempt review status from our institutional review board for human subject's protection.
We used the first round of data, collected at the start of the kindergarten year in the fall, to examine the association between child care experiences in the year before kindergarten on children's weight status when they entered kindergarten. Because our research questions addressed the influence of preschool child care experiences, we restricted our sample to children who entered kindergarten for the first time in 1998 (17 219 of 21 260 total cases). Our analytic sample (n = 15 691) consisted of all individuals, with complete information on the variables described.
Our dependent variable was a dichotomous variable that measured whether a child was obese. To parallel the language and criteria used for adult obesity and conform with terminology in recent recommendations on childhood obesity,7–9 we classified children as obese if they were at or above the 95th percentile on the gender-specific CDC BMI-for-age charts. Following standard protocols, each child's height and weight were measured twice, and values were averaged across the 2 measurements. Height was measured by using a Shorr (Shorr Productions Growth Unlimited, Olney, MD) measuring board. Weight was measured by using a digital scale.24
Using Epi Info 3.3.2,25 we transformed the children's height and weight data to BMI-for-age percentiles and z scores. This program takes into account child's gender as well as height and weight in relation to child's age (in months) at the time of measurement. Following Nader et al,2 we excluded, as biologically implausible, children whose BMI z scores were less than −4 or more than 5. These cases represented ∼0.6% of the sample of first-time kindergartners. We tested our models with and without biologically plausible underweight children (defined as children below the fifth percentile on the gender-specific CDC BMI-for-age charts). The inclusion of these children did not affect our results and were retained in our final sample.
Independent Variables: Type of Child-Care Arrangement
Our primary independent variables were a set of mutually exclusive, dichotomous variables for the primary type of child care used in the year before kindergarten. Children not in a child care arrangement at least 10 hours/week (those referred to as being in parent care) served as the reference category. Primary child care is the type of regular, nonparental child care used most often and at least 10 hours/week during the year before kindergarten. We used the 10 hour/week minimum cutoff to define a primary care arrangement, because infrequent child care arrangements are not likely to exert a strong influence on children's obesity. We classified nonparental primary child care arrangements into 4 types: (1) family, friend, and neighbor care (paid or unpaid care by a relative, unpaid care by a nonrelative, or paid care by a nonrelative that occurs at least occasionally in a child's home); (2) family child care (paid, nonrelative care outside the child's home); (3) Head Start; and (4) center care (day care, nursery school, preschool, or prekindergarten). “Parental care” was the reference category and included children not in a regular, nonparental care arrangement at least 10 hours/week during the year before kindergarten.
We constructed the primary care variables from parents' responses to questions about different types of child care used in the year before kindergarten. Parents were asked whether they used the different arrangements (center, relative, nonrelative, and Head Start); whether the care was paid or unpaid and in-home or out-of-home; hours per week; and the duration (how many months during the year before kindergarten they used this arrangement). We used the duration and average hours per week to determine the primary care arrangement. Because the ECLS-K did not ask about the duration of Head Start participation, we used only average hours per week to determine the primary care arrangement for children who participated in both Head Start and other child care arrangements.
Other Independent Variables
We controlled for variables that, on the basis of previous research, might be expected to influence parental choice of child care. These variables included child's age, race, and gender; family income; number of children younger than 18 years old; age of the youngest child; and mother's education, employment, and marital status.26–28 Many of these variables (child's age, race, and gender; family income; and mother's education, employment, and marital status) have also been associated with children's weight status.1,7,21,29–31 We included additional controls for factors demonstrated or hypothesized to be associated with childhood obesity, including child's birth weight; number of weeks premature; coordination in moving arms and legs compared with other children of same age (mobility impairment); activity level compared with other children of same age; frequency of participation in sports or exercise with the family in a typical week; and mother's receipt of Special Supplemental Nutrition Program for Women Infants and Children (WIC) benefits when the child was an infant.32–35 We also included regional and rural/urban variables constructed by the ECLS-K from census definitions because of their associations with child care use and obesity levels.26,36,37
We discovered that the child's age variable provided in the ECLS-K database had not been updated as more accurate and complete information became available. Thus, we constructed child's age from updated information on child's birth date in relation to assessment date. We measured child's race by a series of dichotomous variables representing the following ethnic categories: Latino, black, Asian, Native American (including Native Hawaiian and other Pacific Islanders), and multiracial. The reference group was white children. Age of the youngest child is measured in years.
Because exploratory descriptive analyses revealed college education as the threshold at which mother's education exerted the largest influence on children's obesity status, we used a dichotomous variable (whether or not the mother had obtained a bachelor's degree or more) as our measure of mother's education. We measured mother's employment status by a set of dichotomous variables: whether a mother was employed full-time (≥35 hours), part-time (<35 hours), or not employed (the reference category). We defined mother's marital status according to dichotomous variables for married, cohabiting, or single and used single as the reference category. For the few cases in which a mother or female guardian was not present in the household, we used the father's information for education, employment, and marital status rather than excluding those cases.
Typically, the parent reported child's birth weight in pounds and ounces; if ounces were missing, we used pounds only. “Weeks premature” was defined as the number of weeks before 37 weeks' gestation of a child's birth. If a child was born at ≥37 weeks' gestation, this variable was equal to 0. We included 3 measures of children's activity levels. The first was a dichotomous variable representing whether the parent perceived the child to be mobility impaired with respect to arm and leg coordination. The second was a continuous variable representing the frequency per week with which the child participated in sports or exercise with the family. The third was a dichotomous variable measuring whether the parent perceived the child to be more active, on average, than other children.
We also included a variable indicating whether a child's parent participated in WIC with the target child. An infant/child is eligible to receive WIC if he or she is (1) younger than 1 year (infant) or younger than 5 years (child), (2) in a family with income between 100% and 185% of the federal poverty line as set by each state, and (3) considered to have a nutrition risk that can be helped by WIC foods and nutrition information. Nutrition risk can include such things as mother's obesity. This variable controls both for household income and risk factors since the child's birth and, thus, adds a historical dimension to the model.
For the descriptive analysis, we calculated means and SEs for the whole sample and separately according to children's obesity status. We used 2-sample t tests to compare the characteristics of obese and nonobese children and their households. We used multivariate logistic models to measure the association between child care in the year before kindergarten and children's obesity status at the start of kindergarten, controlling for other child, family, and regional differences. Using multinomial logistic regressions, we also tested for the influence of child care on children's likelihood of being overweight (≥85th percentile on CDC growth charts), normal weight (<85th percentile), or obese (≥95th percentile). These analyses revealed no significant influence of child care on children's likelihood of being overweight, so we focused our analysis on the logistic regression models with obesity as the dependent variable.
We examined the independent influence of child care on children's obesity and child care's interactive effects with ethnicity and income. We tested these interactions because low-income children and children of some ethnic groups have greater risks of obesity and child care may show different patterns of influence for these groups. Rather than testing interactions with all our variables and the subgroups of interest by running a fully stratified model, we focused solely on the interactions with child care. In our final models, we retained only the interactions that showed a significant association with children's obesity status.
We weighted all data with population weights provided in the ECLS-K. Because a multistage complex sampling design was used to collect the ECLS-K data, standard analyses intended for simple random samples underestimate SEs. Therefore, we used the survey commands in Stata38 to account for the design effects of complex survey sampling and correct the SEs. We explored concerns about selection bias in parents' choice of child-care setting. Because of the complexity of our primary variable of interest–type of child care standard methods for addressing possible self-selection bias do not apply. Thus, assumptions about causality within our results should be made with caution. It is possible other factors not available in the data are associated with both choice of child care and childhood obesity.
Table 1 presents descriptive statistics on the dependent and independent variables in the full sample and separately for obese and nonobese children. At the start of kindergarten, 11.6% of the sample was obese. Other studies using data from the same time period (1997–2001) have revealed similar obesity rates in children 2 to 6 years of age.2,39–42 During the year before kindergarten, 16.9% of children had a family, friend, and neighbor care primary care arrangement; 8.4% had family child care; 9.2% were in Head Start; and 32.7% had center care. Almost 33% of children were not in a regular nonparental child care at least 10 hours/week.
Two-sample t tests revealed some significant differences between obese and nonobese children's characteristics and households. Obese children were significantly more likely than nonobese children to be in family, friend, and neighbor care (20.2% vs 16.5%) and less likely to be in parent care (29.9% vs 33.1%). Obese children were significantly more likely to be male, Latino, have a mobility impairment, have had a higher birth weight, and have received WIC. Obese children were also significantly more likely to reside in a lower-income household and a household with fewer children younger than 18 years. The age of the youngest child in the household was also significantly and positively related to the target child's likelihood of being obese. Obese children were significantly less likely than nonobese children to be white, have a mother who was employed part-time, have a married mother, and have a mother with a bachelor's degree. Obese children were significantly more likely to reside in the Northeast region of the country and less likely to reside in the Midwest.
Table 2 presents the results of 2 logistic regression models that, after controlling for child and family characteristics associated with obesity and choice of child care, tested the association between primary child care arrangement in the year before kindergarten and the likelihood of being obese at the start of kindergarten. Model 1 does not include interaction terms. Model 2 includes interaction terms between child care type and Latino ethnicity, the only high-risk group that showed significant interactions with child care.
In model 1, children in family, friend, and neighbor care were significantly more likely to be obese at the start of kindergarten than children in parent care (the reference category). The odds ratio was 1.22, which represents a 22% change in the odds of being obese or a 1.9% percentage-point difference in the likelihood of being obese for children in family, friend, and neighbor versus parent care calculated by using postestimation commands.43 The size of this effect is similar to the significant effect of watching ≥2 hours of television per day on kindergartners' overweight status found in another study using a later round of the ECLS-K data.36 Comparisons between family, friend, and neighbor care and the other categories of primary care did not reveal significant differences in obesity. (Results are available on request.) Furthermore, we did not find significant associations between other types of care and parent care.
We conducted several supplemental analyses to better understand these results. First, we tested the model excluding children whose primary care arrangement might be considered “nanny care” (paid, in-home, nonrelative care), a distinct category of unlicensed family, friend, and neighbor care. We found the same pattern of results, which indicates that our findings were not biased by including children whose primary care may be provided by nannies. Next, we tested whether other types of family, friend, and neighbor care might have different effects on obesity. We found that relative care in the child's home had the most significant positive association with obesity compared with parent care. The directions of the effects for nonrelative and out-of-home care, although not significant, were also positive, indicating a general trend toward a positive association with obesity. Whether the care was paid or unpaid made less of a difference. Finally, to determine whether the total number of weekly hours in the primary care arrangement was associated with obesity, we ran models with a continuous variable for weekly hours in the primary care arrangement for all children in nonparental care at least 10 hours/week. We found a small but significant positive association between hours in family, friend, and neighbor care and children's likelihood of being obese. We did not find significant associations with hours in other types of care. In other words, for the care category significantly associated with the likelihood of being obese (family, friend, and neighbor care), the likelihood increases slightly with additional hours in care. Results for these analyses are available on request.
Asian, Latino, and Native American children were significantly more likely than white children to be obese. Children's birth weight and number of weeks premature were significantly and positively related to the likelihood of being obese. Male children, children with a mobility impairment, and those who participated in the WIC program were also significantly more likely to be obese than their counterparts. Children's likelihood of being obese was negatively and significantly associated with both household income and the number of children in the household younger than 18 years old. Finally, children whose parents perceived them to be more active than other children were significantly less likely to be obese. In general, these associations conform with established research or, in the absence of other findings, are in the direction we would expect.
Model 2 includes the interactions between Latino ethnicity and primary child care arrangement, controlling for other characteristics. The pattern of results was very similar to those in model 1. Non-Latino children in family, friend, and neighbor care were significantly more likely to be obese than non-Latino children in parent care. In addition, non-Latino children in Head Start were significantly more likely to be obese than non-Latino children in parent care. This finding is somewhat surprising because Head Start is the only type of child care with a regulated nutrition component. However, because it only serves low-income children, other aspects of the child care environment may also be different.
Examining the interaction effects, Latino children in family, friend, and neighbor care and Head Start were significantly less likely to be obese than Latino children in parent care. So, for Latino children, who were significantly more likely to be obese overall, certain types of child care were associated with lower risks of obesity. We used similar postestimation commands to calculate the predicted probabilities of obesity for Latino children in different types of child care for the significant comparisons, holding all other variables at their mean.43 Latino children in parent care had a 15.7% probability of being obese compared with a 14.9% probability for Latino children in family, friend, and neighbor care and an 11.4% probability for Latino children in Head Start.
Although most preschool-aged children spend significant amounts of time in nonparental child care, we know very little about the influence of child care participation on children's weight. This study provides a first look at how different types of child care may influence children's likelihood of being obese. Still, some limitations apply. We cannot definitively establish causation between child care participation and obesity. We do not know the mechanisms by which child care environments contribute or protect against obesity. Also, because the data did not include children's BMI before preschool, we were unable to control for earlier influences on childhood obesity.
We have reported 3 main findings. First, when all else was held constant, children in family, friend, and neighbor care were more likely to be obese than children in parent care. This finding warrants additional investigation into the nutritional and physical activity features of family, friend, and neighbor environments associated with obesity in young children. Little is known about this type of child care, in part because it is typically unlicensed and the caregivers are difficult to find and survey. Until such data are collected, we have some hypotheses that may help explain these findings. Most family, friend, and neighbor caregivers are relatives (grandparents in particular) whose own income and educational backgrounds generally mirror those of the child's parents.44 Yet, our findings show that caregiver influence on children's physical development may be different from that of parents. For example, grandparents may be less physically active than parents. Another possibility is that indulgent grandparents may be less likely than parents to restrict their grandchildren's television viewing or access to energy-dense favorite foods, believing that “grandparent time” should be a time when children's whims rule.
Second, we found that non-Latino children in Head Start were more likely to be obese than children in parent care, even after controlling for differences in household income. One possible explanation is that we were unable to control for all the characteristics of families using Head Start. Some of these characteristics could also be associated with obesity, or, despite Head Start's nutrition component, which likely varies in its implementation, other aspects of the environment may contribute to children's obesity. For example, if teachers perceive families to be suffering from food insecurity, they may impose fewer restrictions on the amount of food children consume during Head Start. Although our results are not directly comparable with other studies because of differences in child care definitions, this article contributes to the emerging literature on the influence of an important environmental setting on children's physical development.
Finally, we found a significant moderating effect of child care on obesity in Latino children. Although Latino children were significantly more likely than white children to be obese in the first place, the magnitude of this difference was reduced when Latino children were primarily in Head Start or family, friend, and neighbor care. In light of the main effect for family, friend, and neighbor care, we were surprised to find the slightly protective effect of this type of primary care for Latino children. One explanation is that grandparents who care for Latino children may be more likely than the child's parents to prepare and serve traditional Latino food, which may be protective. The protective effect of Head Start was less surprising given the greater emphasis on nutrition and physical activity in Head Start compared with other types of early childhood programs. However, the protective effect applied only to Latino children. Recent research about differences in Latino home environments compared with other ethnic groups may provide some explanation as to why certain types of child care may be protective. For instance, food insecurity, watching television during dinner, and putting a child to bed with a bottle have been found to be more common in Latino homes.21,45 Spending time in other settings may help mitigate these effects to the extent that they are associated with obesity.
Our research points to the need to better understand how the specific features of child care environments may promote or protect against the development of obesity. For Latino and other populations of children at greatest risk of obesity, a crucial task is to understand how social, cultural, and economic influences interact with features of child care environments in the development of obesity. This understanding can then lead to the design and development of targeted interventions to reach children and families in child care settings. The opportunity for targeted intervention to prevent obesity is ripe, because preschool-aged children spend so much time in nonparental child care.
Pediatricians can play a role in educating parents about aspects of child care settings that are critical for promoting healthy physical development. There is a current American Academy of Pediatrics policy statement on quality early care and education that highlights pediatricians' role in promoting access to high-quality environments.46 This statement could be revised to include features of quality environments that promote healthy nutrition and physical activity. These recommendations need to be tailored to the unique attributes of family, friend, and neighbor care, and messages should be crafted to reach caregivers and encourage parent and caregiver communication about these issues. Research suggests that parents, child care center directors, and child care health consultants all believe that health promotion in child care settings can be most effective with the direct involvement of health care professionals.47 Pediatricians and other health professionals clearly have a strong role to play here.
This study makes several contributions to the emerging literature on the influence of child care, one of the most significant and protracted experiences of a child's early life, on children's physical development and points to new targets for intervention. Future research can help us understand the specific influences of these environments, particularly in family, friend, and neighbor care. Although highly prevalent among young children, relatively little is known about this type of care because it is unlicensed and unregulated. The more we understand about the influences of child care on children's obesity, the more effective targeted interventions in the prevention of obesity will be. Child care settings offer an almost ubiquitous opportunity for policy and program intervention to address children's adiposity in the early years when significant change in developmental trajectories may still be possible.
This work was supported by a grant from the Robert Wood Johnson Foundation.
We thank Yingying Zhou for editorial and data assistance.
- Accepted December 12, 2007.
- Address correspondence to Erin J. Maher, PhD, Casey Family Programs, 1300 Dexter Ave N, Third Floor, Seattle, WA 98109. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
Few studies have examined the influence of child care on obesity, and none have used a nationally representative sample with direct assessments of children's weight with detailed child care categories aligned with those used in the early care and education field.
What This Study Adds
This study makes several contributions to the emerging literature on the influence of child care (one of the most significant and protracted experiences of a child's early life) on children's physical development and points to new targets for intervention.
- ↵Nader PR, O'Brien M, Houts R, et al. Identifying risk for obesity in early childhood [published correction appears in Pediatrics. 2006;118(5):2270]. Pediatrics. 2006;118 (3). Available at: www.pediatrics.org/cgi/content/full/118/3/e594
- ↵West J, Denton K, Germino-Hausken E. America's Kindergarteners. Washington, DC: US Department of Education, National Center for Education Statistics; 2000
- ↵National Center for Education Statistics. Household Education Survey Parent Survey. Available at: http://nces.ed.gov/nhes/dataproducts.asp#1999dp. Accessed November 1, 2007
- ↵Kuczmarski R, Ogden C, Guo S, et al. 2000 CDC growth charts for the United States: methods and development. Vital Health Stat 11. 2002;(246):1–190
- Robert Wood Johnson Foundation Medical Panel. Medical panel outlines measures to improve treatment, prevention of childhood obesity. Available at:www.rwjf.org/programareas/features/digest.jsp?c=EMC-ND138&pid=1138&id=5778. Accessed June 13, 2007
- ↵American Medical Association. Appendix: expert committee recommendations on the assessment, prevention, and treatment of childhood and adolescent overweight and obesity. Available at: www.ama-assn.org/ama1/pub/upload/mm/433/ped_obesity_recs.pdf. Accessed June 6, 2007
- ↵Whitaker R, Pepe M, Wright J, Seidel K, Dietz W. Early adiposity rebound and the risk of adult obesity. Pediatrics. 1998;101 (3). Available at: www.pediatrics.org/cgi/content/full/101/3/e5
- ↵Centers for Disease Control and Prevention. Using the BMI-for-age growth charts. Available at: www.cdc.gov/nccdphp/dnpa/growthcharts/training/modules/module1/text/page4btext.htm. Accessed October 10, 2006
- ↵Variyam J, Blaylock J, Lin B, Ralston K, Smallwood D. Mother's nutrition knowledge and children's dietary intakes. Am J Agric Econ. 1999;81 (2):373– 384
- ↵Pate R, Pfeiffer K, Trost S, Ziegler P, Dowda M. Physical activity among children attending preschools. Pediatrics. 2004;114 (5):1258– 1263
- ↵National Center for Education Statistics. Early Childhood Longitudinal Study Kindergarten Cohort, Base Year Public-Use Data File, Kindergarten Class of 1998–1999 [CD-ROM]. Washington, DC: US Department of Educational, National Center for Education Statistics; 2001
- ↵National Center for Education Statistics. Early Childhood Longitudinal Study, Kindergarten Class 1998–99. Washington, DC: US Department of Education; 2002
- ↵Dean AG, Arner TG, Sunki GG, et al. Epi Info: A Database and Statistics Program for Public Health Professionals. Atlanta, GA: Centers for Disease Control and Prevention; 2005
- ↵Blau DM. The Child Care Problem: An Economic Analysis. New York, NY: Russell Sage Foundation; 2001
- Capizzano J, Adams G, Ost J. Caring for Children of Color: The Child Care Patterns of White, Black, and Latino Children Under 5. Washington, DC: Urban Institute; 2006
- ↵Drewnowski A, Specter SE. Poverty and obesity: the role of energy density and energy costs. Am J Clin Nutr. 2004;79 (1):6– 16
- Strauss R, Knight J. Influence of the home environment on the development of obesity in children. Pediatrics. 1999;103 (6). Available at: www.pediatrics.org/cgi/content/full/103/6/e85
- ↵Danielzik S, Czerwinski-Mast M, Langnase K, Dilba B, Muller M. Parental overweight, socioeconomic status and high birth weight are the major determinants of overweight and obesity in 5–7 y-old children: baseline data of the Kiel obesity prevention study (KOPS). Int J Obes Relat Metab Disord. 2004;28 (11):1494– 1502
- Ness A. The Avon Longitudinal Study of Parents and Children (ALSPAC): a resource for the study of the environmental determinants of childhood obesity. Eur J Endocrinol. 2004;151 (suppl 3):U141– U149
- ↵Rose D, Bodor N. Household food insecurity and overweight status in young school children: results from the Early Childhood Longitudinal Study. Pediatrics. 2006;117 (2):464– 473
- ↵Stata Corp. Stata Survey Data Reference Manual. Release 9. College Station, TX: Stata Press; 2005
- ↵Salsberry PJ, Reagan PB. Dynamics of early childhood overweight. Pediatrics. 2005;116 (6):1329– 1338
- ↵Long JS, Freese J. Regression Models for Categorical Outcomes Using Stata. 2nd ed. College Station, TX: Stata Press; 2005
- ↵Brandon R, Maher E, Joesch J, Doyle S. Understanding Family, Friend, and Neighbor Care in Washington State: Developing Appropriate Training and Support. Seattle, WA: Human Services Policy Center, Evans School of Public Affairs, University of Washington; 2002
- ↵American Academy of Pediatrics, Committee on Early Childhood, Adoption, and Dependent Care. Quality early education and child care from birth to kindergarten. Pediatrics. 2005;115 (1):187– 191
- ↵Gupta RS, Shuman S, Taveras EM, Kulldorff M, Finkelstein JA. Opportunities for health promotion education in child care. Pediatrics. 2005;116 (4). Available at: www.pediatrics.org/cgi/content/full/116/4/e499
- Copyright © 2008 by the American Academy of Pediatrics