The current study investigated patterns of growth in academic achievement trajectories among American high school students (N = 12,314) that were obtained from a nationally representative, public-use dataset (the High School Longitudinal Study of 2009) in relation to key demographic information (i.e., gender, grade level, socioeconomic status [SES] in ninth grade, and ethnicity) and a distal outcome (i.e., applying for college). Unconditional growth mixture model showed that the three-class model was most appropriate in capturing the latent heterogeneity (i.e., low-achieving/increasing, moderate-achieving/decreasing, and high-achieving/slightly increasing). Two covariates (i.e., gender and SES in ninth grade) were positively associated with the intercept growth factor (i.e., initial GPA) in two of the three achievement classes (i.e., high-achieving and moderate-achieving). In contrast, two other covariates (i.e., Hispanic and African American) were negatively associated with the intercept growth factor in all of the achievement classes. The multinomial logistic regression coefficients identified an increase in the likelihood of belonging to the following achievement classes: (1) Moderate-achieving, if the students were male or African American and of low SES, (2) Low-achieving, if the students were male and of low SES, and (3) High-achieving, if the students were female and of an ethnicity other than African American and high SES. The probability of not applying for college was higher among the low-achieving and the moderate-achieving classes compared with the high-achieving class (223 words).
Trajectories of Academic Achievement in High Schools: Growth Mixture Model
