Multiple linear regression (MLR) was used to test our hypotheses
that psychosocial and behavioral variables are predictive of
high school GPA after controlling for traditional predictors
(school-level, demographic, and academic achievement variables).
Because of the nesting of students within schools, hierarchical
linear modeling (HLM), where the regression coefficients could
vary by school, was also considered. As detailed later in the
Results, the parameter estimates from MLR were not tangibly
different than the HLM-generated estimates. Furthermore, because
the MLR model permits dominance analysis (description forthcoming),
the MLR model was chosen over HLM.
Multiple linear regression (MLR) was used to test our hypothesesthat psychosocial and behavioral variables are predictive ofhigh school GPA after controlling for traditional predictors(school-level, demographic, and academic achievement variables).Because of the nesting of students within schools, hierarchicallinear modeling (HLM), where the regression coefficients couldvary by school, was also considered. As detailed later in theResults, the parameter estimates from MLR were not tangiblydifferent than the HLM-generated estimates. Furthermore, becausethe MLR model permits dominance analysis (description forthcoming),the MLR model was chosen over HLM.
การแปล กรุณารอสักครู่..