There is an accumulating body of evidence suggesting that being raised in a non-intact family may adversely affect child outcomes across a number of developmental domains. There are nevertheless questions re: interpretation for a number of methodological reasons ranging from conceptualization of exposure to analytic approach. Given the dynamic nature of family life for some children today, changes in family structure may need to be captured in a more comprehensive manner. The scientific question is how capturing dynamics of family structure can be achieved. Data were drawn from the National Longitudinal Survey of Children and Youth (NLSCY), an ongoing longitudinal population survey following cohorts of Canadian children from infancy to adulthood across a number of domains of health and well-being. The sample comprised individuals aged 11-14 in cycle five of the NLSCY. We examined different approaches to conceptualizing and modeling the effect of the family structure effect based on previous studies in this area and from the lifecourse epidemiological literature. We examined the effects of current family structure, previous family structures, trajectories of family structure, and change in family structure on externalizing and internalizing behavior in pre-adolescence. We found that the validity of each approach was largely dependent on the specific research question at hand, with each one having its own advantages and disadvantages. Main methodological findings included confounding of the current family structure effect by previous experience, collinearity among family structure main effects, and low analytic power for trajectories. Guided by a conceptual diagram of the underlying causal structure, we also considered the time-varying nature of family income and employment status of the primary household respondent using inverse probability weighting to estimate the causal parameters of a marginal structural model. In one of our most sophisticated conceptualizations of family structure effects, we found that recent change in family structure had a statistically significant effect on the odds of externalizing behavior, OR (95% CI) = 2.95 (1.73-5.02). Overall, our substantive findings were tempered by methodological caveats, which have important implications for future studies in the area and for broader issues related to data collection, study design, and analysis.
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.115849 |
| Date | January 2009 |
| Creators | McDonald, Sheila, 1971 Nov.17- |
| Publisher | McGill University |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Format | application/pdf |
| Coverage | Doctor of Philosophy (Department of Epidemiology and Biostatistics.) |
| Rights | All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. |
| Relation | alephsysno: 003134114, proquestno: AAINR66605, Theses scanned by UMI/ProQuest. |
Page generated in 0.002 seconds