Record Details
Early Predictors of Academic Achievement and Externalizing Problems for Children in Low-Income Families
ScholarsArchive at Oregon State University
| Field | Value |
|---|---|
| Title | Early Predictors of Academic Achievement and Externalizing Problems for Children in Low-Income Families |
| Names |
Duncan, Robert J. (Robert Joseph)
(creator) McClelland, Megan M. (advisor) |
| Date Issued | 2015-06-10 (iso8601) |
| Note | Graduation date: 2016 |
| Abstract | Empirical studies and theory have identified many early predictors of children's academic achievement and externalizing problems. Moreover, research shows that children's early cognitive abilities and behavioral problems are strong predictors of later academic achievement and externalizing problems. The current dissertation studies extend previous work and unpack how early predictors relate to children's academic achievement and externalizing problems in a low-income sample. Both of the dissertation studies are secondary data analyses of the Early Head Start Research and Evaluation Study, which evaluated the effectiveness of Early Head Start (EHS) from 1996 through 2010. Study I uses inductive, sequential partitioning analyses, to explore the effects of family income and needs, child and family characteristics, child health, and early care and education (i.e., from roughly the time of the child's birth) on children's prekindergarten and 5th grade academic achievement. The study also explores subgroup effects of EHS based on sequential partitioning analyses. Study I had three key findings. First, much of the explained variance in prekindergarten academic achievement was accounted for by children's ethnicity and maternal educational attainment. However, ethnicity explained substantially more variance than maternal educational attainment for 5th grade academic achievement. Second, common-domain skills (e.g., early math for later math) explained much of the link for 5th grade math and vocabulary performance. However, early math and vocabulary skills, and not early literacy, were predictive of 5th grade literacy. Third, EHS was a selected split (in the intended direction) for a subgroup of children (i.e., Black/Hispanic, relatively lower levels of maternal educational attainment, and not in deep poverty) in the prekindergarten vocabulary model, but not a selected split in any other model. Follow up subgroup analyses of EHS effects showed mostly non-significant results on children's prekindergarten and 5th grade academic achievement. Study II examines the effects of economic hardship, child and family characteristics, and early care and education (i.e., from roughly the time of the child's birth) and prekindergarten behavioral problems on children’s externalizing problems at 5th grade. The study uses two analytical approaches to better understand these relations: a deductive, logistic regression model and an inductive, sequential partitioning model (i.e., SEARCH). Across both models, being male, maternal unemployment and not living with the husband at roughly the time of the child's birth, and high levels of prekindergarten aggression (i.e., FACES subscale) were predictive of a greater likelihood of externalizing problems at 5th grade. However, three differences in results were found between the two analytical approaches. First, variables related to economic hardship (e.g., depth of poverty, welfare receipt) were selected in the SEARCH model, but were not significant in the logistic regression model. Second, there was a significant beneficial subgroup effect of EHS (i.e., females not in deep poverty and who had mothers employed or in school at roughly the time of the child's birth) in the SEARCH model, but it was not significant in the logistic regression model. Third, prekindergarten hyperactivity was not selected in the SEARCH model but was significant in the logistic regression model. The two dissertation studies uniquely contribute to research on early predictors of children’s academic achievement and externalizing problems. Study I shows how inductive analyses can be used to connect back to previous research and guide new hypothesis driven research, whereas Study II shows the usefulness of taking deductive and inductive analytical approaches for developmental research |
| Genre | Thesis/Dissertation |
| Access Condition | http://creativecommons.org/licenses/by-nc-nd/3.0/us/ |
| Topic | children's development |
| Identifier | http://hdl.handle.net/1957/56268 |
