GED Student Achievement Trajectories and Predictors: Using the Latent Class Growth Model to Understand Heterogeneous Learning Patterns

Abstract

Each year, 500,000 adults take the General Education Development (GED) exam, the gateway to a high school equivalency diploma. Nearly 40 percent fail. Little research exists on the relation between preparation experiences and exam success. This thesis investigated heterogeneous learning patterns among women at a GED preparation program in Detroit, Michigan during their first ten months of enrollment. Using Latent Class Growth Modeling, patterns among repeated measures on math and on reading were best accounted for by three latent trajectory classes per subject. Classes followed unique and non-overlapping achievement trajectories. Five factors were assessed as potential predictors of trajectory class membership. Younger age, non-English native language, and employment predicted a higher-achieving math trajectory, while being a native English speaker and being a single head of household predicted a higher-achieving reading trajectory. Last school grade completed was not predictive of achievement in either subject domain. Associations between math and reading achievement trajectories were also analyzed, with students generally exhibiting similar levels of achievement across subjects, though high math achievement trajectory membership was more predictive of high reading achievement trajectory membership than vice versa. Finally, membership in classes following higher achievement trajectories was determined to be associated with a greater likelihood of success on the GED exam.