This research analyzed linked high-quality state data from K-12, adult education, and postsecondary state datasets in order to better understand the association between student demographics and successful completion of a postsecondary program. Due to the relatively small sample size compared to the large number of features, we analyzed the data with Nearest Shrunken Centroid (PAM), a statistical method developed for cancer and genomic research. Our findings conclude that there are five features that are predictive of an adult education student's successful completion of a postsecondary program.