Article
Family selection and child care experiences: implications for studies of child outcomes

https://doi.org/10.1016/S0885-2006(00)00072-7Get rights and content

Abstract

Studies of the impact of child care experiences on child outcomes must consider family selection factors because children from more advantaged families tend to attend higher quality child care and are more likely to be in center care than children from less advantaged families. Although this issue is widely recognized, developmentalists and economists have used different statistical methods when testing whether child care experiences are related to child outcomes and have drawn different conclusions from their analyses. This paper discusses some of the family selection issues that should be considered in child care research and provides empirical evidence demonstrating why each issue should be considered. These issues include whether causal inferences can be made from observational studies and the impact on conclusions from regression analyses that include highly correlated measures of child care experiences, nonrepresentative samples, and family covariates with bi-directional effects on child care quality.

Section snippets

Implications for child care studies

Family selection factors are almost always considered in current child care research. Although the initial child care studies published in the 1970s and 1980s by developmentalists often did not consider family selection factors, all studies now must include family selection factors as control variables in their analyses to be credible (Lamb, 1998). Developmentalists identify family selection variables to include as control variables by identifying family characteristics that are significantly

CQO

The first dataset is from the Cost, Quality, and Outcomes Study (CQO), a large-scale study conducted in child care centers in California, Connecticut, Colorado, and North Carolina (for more details see Helburn 1995, Peisner–Feinberg and Burchinal 1997). These states varied widely in both economic climate and the stringency of state child care regulations. For example, regulations regarding staff:child ratios for 4-year-olds varied from 1:10 in Connecticut to 1:20 in North Carolina. The initial

Inferring causality

From our perspectives as developmentalists, much of the controversy between developmentalists and economists regarding selection effects is because of fundamental differences in analytic traditions. Psychologists regard the experimental design with random assignment to treatment and control groups as the gold standard against which all other designs are compared (Kirk, 1982). Most psychologists believe that it is not possible to infer cause and effect from observational studies, with the caveat

Discussion

Child care researchers must attend to family selection factors. It has been clearly demonstrated that children from families with higher incomes and more education tend to experience somewhat higher quality child care than children with fewer advantages. These family selection effects on child care quality appear to be relatively modest, but must be considered when asking whether children’s development could be influenced by their child care experiences. How to appropriately adjust for family

References (43)

  • M.R. Burchinal et al.

    Quality of center infant care and infant cognitive and language development

    Child Development

    (1996)
  • F. Campbell et al.

    Effects of early intervention on intellectual and academic achievementA follow-up study of children from low-income families

    Child Development

    (1994)
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1988)
  • L.J. Cronbach et al.

    How should we measure “change”—or should we?

    Psychological Bulletin

    (1970)
  • R. Davidson et al.

    Estimation and inference in econometrics

    (1993)
  • H.E. Doran

    Applied regression analysis in econometrics

    (1989)
  • Dunn, L. M., & Dunn, L. M. (1981). Peabody picture vocabulary test–revised. American Guidance Service: Circle Pines,...
  • E.M. Foster et al.

    An illustration of the use of instrumental variablesDo neighborhood conditions affect a young person’s chance of finishing high school?

    Psychological Bulletin

    (1996)
  • W.H. Greene

    Econometric analysis

    (1997)
  • T. Harms et al.

    Early childhood environment rating scale

    (1980)
  • W.L. Hays

    Statistics for the social sciences

    (1973)
  • Cited by (41)

    • Rural families’ use of multiple child care arrangements from 6 to 58 months and children's kindergarten behavioral and academic outcomes

      2017, Early Childhood Research Quarterly
      Citation Excerpt :

      Because children cannot be randomly assigned to the experience of multiple arrangements, we recognized the importance of accounting for factors that are related to selection into child care as well as child outcomes. Selection effects have been defined as characteristics of children and families related to entrance into non-parental child care as well as the degree to which children experience higher or lower quality care, more or fewer hours in care, formal or informal care, and single or multiple arrangements (Bratsch-Hines, Mokrova, Vernon-Feagans, & Family Life Project Key Investigators, 2015; Burchinal & Nelson, 2000; Dearing, McCartney, & Taylor, 2009). Selection effects often include characteristics such as race, household income, home environment quality, and maternal depression (Liang, Fuller, & Singer, 2000; Morrissey, 2008, 2009).

    • Child care subsidies and the stability and quality of child care arrangements

      2017, Early Childhood Research Quarterly
      Citation Excerpt :

      Fixed effects methods have been used in assessing the impact of child care choices in other settings, for instance in examining the relationship between multiple child care arrangements and children’s behavior (Morrissey, 2009) or the effect of income on child development (Blau, 1999). The main downside of using fixed effect methods is likely to be reduction in the power to test hypotheses and larger standard errors (Burchinal & Nelson, 2000). Fortunately, in our results below, we have sufficiently precise estimates to test our hypotheses of interest about the impacts of subsidy receipt.

    • Child care instability from 6 to 36 months and the social adjustment of children in prekindergarten

      2015, Early Childhood Research Quarterly
      Citation Excerpt :

      In the literature on child care, researchers place emphasis on understanding the selection effects associated with young children's child care experiences. Selection effects have been defined as characteristics of families and children associated with the likelihood that they will enter into nonparental child care as well as the degree to which they will experience higher or lower quality care, more or fewer hours in care, formal or informal care, and enter into child care earlier or later (Burchinal & Nelson, 2000; Dearing, McCartney, & Taylor, 2009; Hirshberg, Huang, & Fuller, 2005). Furthermore, family characteristics may influence whether or not children experience child care instability.

    • The quality of toddler child care and cognitive skills at 24 months: Propensity score analysis results from the ECLS-B

      2014, Early Childhood Research Quarterly
      Citation Excerpt :

      Following Kontos, Howes, and Galinsky (1996) and Phillipsen, Burchinal, Howes, and Cryer (1997), low-quality care is categorized as a score on the ITERS or FDCRS below 3, medium-quality care as a score greater than or equal to 3 and less than 5, and high-quality care as a score of 5 and above. In using this approach to measure overall child care quality, we borrow from prior research (Burchinal & Nelson, 2000) that treated scores from different child care quality measures equally in analyses that accounted for potential differences in mean scores and variability across the two instruments. To ensure the quality construct is appropriate, the regression models described below allow for the combined quality score to vary by type of care (by including a type-of-care-by-quality interaction term) to account for any differences that might be measure based.

    • New evidence on the validity of the Arnett Caregiver Interaction Scale: Results from the Early Childhood Longitudinal Study-Birth Cohort

      2013, Early Childhood Research Quarterly
      Citation Excerpt :

      Whitebook et al. (1989) found small partial correlations (.24 and below), Howes et al. (1998) did not report effect sizes, and Vernon-Feagans and Manlove (2005) reported large effect sizes, although their findings were specific to nine children with chronic ear infections who were cared for in low-quality contexts. The tendency for prior studies to create composite measures of child care quality from the Arnett CIS and other scales reflects the fact that quality measures tend to be highly inter-correlated (Burchinal & Nelson, 2000). For example, Burchinal and Cryer (2003) reported correlations between .74 and .91 among the Arnett CIS, the Early Childhood Environmental Environment Rating Scale-Revised (ECERS-R) and another quality measure.

    View all citing articles on Scopus

    The secondary data analysis was supported by a grant from Office of Educational Research and Improvement. The Cost, Quality, & Outcomes study was funded by grants from the Carnegie Corporation of New York, the William T. Grant Foundation, the JFM Foundation, the A. L. Mailman Family Foundation, the David and Lucille Packard Foundation, the Pew Charitable Trusts, the USWEST Foundation, and one anonymous foundation, and was conducted by a team of researchers including Donna Bryant, Peg Burchinal, Richard Clifford, Debby Cryer, Mary Culkin, Suzanne Helburn, Carollee Howes, Sharon Lynn Kagan, H. Naci Mocan, John Morris, Leslie Phillipsen, Ellen Peisner–Feinberg, and Jean Rustici. The Carolina Otitis Media Project was supported by the Maternal and Child Health Bureau, Health Resources and Services Administration, Department of Health and Human Service and by the Spencer Foundation. We are grateful to the child care staffs, the children and their families for their participation in these projects. Portions of the paper were presented at the 1999 Biennial Meeting of the Society for Research in Child Development.

    View full text