Hierarchical Linear Modeling of Students’ Mathematics Self-Efficacy and School Effects on Mathematics Achievement

The purpose of this study was to investigate the relationship between mathematics self-efficacy and mathematics achievement of high school sophomores across the United States, and to examine the effects of gender, ethnicities, and school characteristics on students’ mathematics achievement using hierarchical linear modeling (HLM). The base-year data of the Educational Longitudinal Study (ELS): 2002 were used for analysis. Hierarchical linear models were developed from the one-way random effects ANOVA model, and the unconditional Model with mathematics self-efficacy in level 1, to the contextual models with variables in the both levels. Both fixed effects and random effects were estimated and interpreted for all the models