Secondary Data Analysis Strand and Statistical Modeling

Investigators: Margaret Burchinal (burchinal@unc.edu ) , Robert Bradley, Lynette Keyes

Project staff: Eloise Neebe, Karen Cai, Wei Jin, Dee Stewart, Donna Hudgins

Research questions, goals: The statistical modeling and support core provided data management and analysis for the major projects and sub-projects and conducts new research based on secondary data analysis of archived national and local data sets. Specific goals were to:

1. provide professional data management and analysis for the projects
2. demonstrate statistical approaches to analyzing longitudinal early childhood data based on ecological models
3. address important issues regarding quality of care and child outcomes and whether these pathways differ among children from diverse backgrounds
4. address issues regarding factors associated with successful transitions from infancy to toddlerhood, to toddlerhood to preschool, and preschool to school

Research achievements: Data management and analysis of other strands' data

• national surveys of kindergarten teachers and child care providers
• field testing of the study to develop a quality instrument for early intervention services for infant or toddler children with disabilities
• study of teacher continuity in infant child care
• second grade follow-up of children in the Cost, Quality, and Outcomes Study
• family literacy and behavior problems studies (first 3 years only)

New, major findings:

• Child care quality matters for children's development while they are in child care and as they transition into public schools. Children who experience higher quality care are more likely to have more advanced language, academic, and social skills. The child care effects are smaller than those of the family, with child care effect sizes being about half the size of the family effect sizes. Nevertheless, the child care effects are not trivial and can be maintained through second grade.
• Quality child care matters more for at-risk children. The long-term prediction of academic outcomes in second grade from child care quality was stronger for children whose mothers had less education.
• Standard measures of child care quality "work" for children of color. Analyses of data from the various child care projects indicate that the standard middle-class definition of quality actually predicts child outcomes known to predict school success for children of color at least as well as it predicts outcomes for white children.
• Center-based childcare interventions are more effective than family-based home visiting interventions. Analyses of child outcomes from clinical trial studies of center-based and family-based intervention programs for children from low-income families indicate that center-based interventions can have substantial and enduring effects and that home-visiting interventions, at best, have very modest, concurrent effects on child outcomes.
• Predictors of quality differ by child care setting. Whereas better adult-child ratios are consistently related to higher quality in child care centers, this association is inconsistently observed in child care homes. In contrast, provider education is clearly linked to quality in both settings.
• Child care training and education improve the quality of child care and child outcomes.
• Caregivers with more training, either on-site or formal training, provided higher quality child care and had children with better language and social skills.

Benefits of research: Findings from our studies and other child care studies have resulted in enhanced regulations in many states and greater interest in childcare at the state and national level. This should benefit both children and their families.

Policy, practice or professional development implications:

• Policies should ensure that all children, and especially children from less advantaged homes, have access to high quality child care.
• Professional development of child care providers enhances the quality of the care they provide.
• Interventions need to be sufficiently intense to change developmental outcomes for children from less advantaged homes. It appears that high quality center-based programs are substantially more successful than home-based programs.

Publications, products:

Bradley, R. H., Burchinal, M., & Casey, P. H. (2001). Early intervention: The moderating role of the home environment. Applied Developmental Science, 5, 1-7.

Burchinal, M. R. (1999). Statistical methods for describing developmental patterns. Early Education and Development, 10, 83-99.

Burchinal, M. R. (1999). Childcare experiences and developmental outcomes. Annals of the American Academy of Political Science. 563, 73-98.

Burchinal, M. R., & Nelson, L. (2000). Family selection and child care experiences: Implications for studies of child outcomes. Early Childhood Research Quarterly, 15, 385-412.

Burchinal, M. R., Peisner-Feinberg, E., Bryant, D. M., & Clifford, R. (2000). Children's social and cognitive development and child care quality: Testing for differential associations related to poverty, gender, or ethnicity. Applied Developmental Sciences, 4, 149-165.