A latent growth moderated mediation model of math achievement and postsecondary attainment: Focusing on context-invariant predictors.

• Autores: R. Sergio Guglielmi, Nancy Brekke
• Localización: Journal of educational psychology, ISSN-e 1939-2176, ISSN 0022-0663, Vol. 110, Nº. 5, 2018, págs. 683-708
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • Math proficiency is a strong predictor of postsecondary attainment, yet cross-national comparative assessments show that U.S. students consistently score below many of their peers, especially those in East Asia. Similar math achievement disparities exist across ethnic/racial groups within the U.S., where Asian American students routinely outperform their White and Hispanic counterparts. The stability of math achievement rankings within and across countries suggests that several predictors are transportable and context-invariant, as they forecast math achievement even following striking shifts in contextual characteristics, such as country of residence, language, or school system. We evaluated a multiple mediator model of math achievement and postsecondary attainment that incorporates several of these predictors, using cross-sectional international data (PISA-2012) and longitudinal U.S. data (NELS and ELS). In all samples and data sets, the model explained substantial portions of the variance in educational outcomes. Parents’ educational aspirations for their children were strongly associated with students’ own aspirations which, in turn, were powerful predictors of math achievement and postsecondary attainment—both directly, and through their influence on important mediators such as math self-perception. Longitudinal relations identified in an older national dataset (NELS) were generally cross-validated in ELS. Moreover, in both data sets, construct relations were found to be largely invariant across racial/ethnic groups. Although additional research is needed to clarify the mechanisms underlying group disparities, the outstanding fit of the model (across samples and data sets) provides the empirical foundation for educational interventions that are likely to help all students, regardless of their ethnicity or national origin. (PsycINFO Database Record (c) 2018 APA, all rights reserved)