What is most important: social factors, health selection, and adolescent educational achievement

Published: September 30, 2011
Category: Bibliography > Reports
Authors: De Rocquigny J, Edgerton J, Hiebert B, MacWilliam L, Manivong P, Roos LL, Walld R
Countries: Canada
Language: null
Types: Population Health
Settings: Government, PCP

Soc Indic Res 110:385-414.

Department of Community Health Sciences, Faculty of Medicine, Manitoba Centre for Health Policy, University of Manitoba, Winnipeg, MB, Canada

This paper explores the relative importance of social factors and health measures in predicting educational achievement in early and late adolescence using population-based administrative data. The sample was made up of 41,943 children born in Manitoba, Canada between 1982 and 1989 and remaining in the province until age 18. Multilevel modeling nests each individual (level 1) within a family (level 2) residing within a neighborhood (level 3). Most important in predicting adolescent achievement were a broad socioeconomic status index (and a narrower measure of household income), being on social assistance, mother’s age at first birth, gender, residential mobility, the presence of ADHD/Conduct disorders, and measures of family functioning (child taken into care or offered protection services and family structure history). Family size, birth order, and newborn characteristics (birthweight, APGAR, gestational age) were statistically significant but of little importance in explaining the outcomes. Both examining regression coefficients and systematically omitting variables showed social factors (often emphasized by epidemiologists) to have markedly greater effects than the combination of health measures (often stressed by economists) in predicting achievement. However, mental health in childhood is identified as among the important predictors. Record linkage across population datasets from health, education, and family services ministries allowed: tracking health and educational attainment at different times in a child’s life, following a large number of cases across childhood, considerable sensitivity testing, controlling for unmeasured family and neighborhood effects, generating an extensive list of predictors, estimating effect sizes, and comparing Manitoba results with those of well-known American studies.

Population Markers,Aged,Gender,Outcome Measures,Canada,Socioeconomic gradient,Multilevel modeling,Record linkage,Life course models,Childhood health

Please log in/register to access.

Log in/Register

LinkedIn Facebook Twitter

© The Johns Hopkins University, The Johns Hopkins Hospital, and Johns Hopkins Health System.
All rights reserved. Terms of Use Privacy Statement

Back to top