The purpose of this study is to investigate the extent that there is a typology of high schools based on their orientation toward science, technology, engineering, and mathematics (STEM), as well as the extent to which school-level demographic variables and student high school outcomes are associated with subgroup membership in the typology, by analyzing data from a large nationally representative sample of high schools (n = 940) from the High School Longitudinal Study of 2009 (HSLS:09) using latent class analysis (LCA). We used a three-step LCA approach to identify significantly different subgroups of STEM-oriented high schools, what covariates predict subgroup membership, and how subgroup membership predicts observed distal outcomes. We find that there are four significantly different subgroups of STEM-oriented high schools based on their principal's perceptions: Abundant (12.3%), Support (23.3%), Bounded (10.1%), and Comprehensive (54.3%). In addition, we find that these subgroups are associated with school demographics, such as the percent of students eligible for free and reduced-price lunch, school locale, and control (public or private). Subgroup membership is also associated with student outcomes, such as postsecondary program enrollment and intent to pursue a STEM degree.

Early Warning Systems (EWS) and Early Warning Indictors (EWI) have recently emerged as an attractive domain for states and school districts interested in predicting student outcomes using data that schools already collect with the intention to better time and tailor interventions. However, current diagnostic measures used across the domain do not consider the du& issues of sensitivity and specificity of predictors, key components for considering accuracy. We apply signal detection theory using Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) analysis adapted from the engineering and medical domains, and using the pROC package in R. Using nationally generalizable data from the Education Longitudinal Study of 2002 (ELS:2002) we provide examples of applying ROC accuracy analysis to a variety of predictors of student outcomes, such as dropping out of high school, college enrollment, and postsecondary STEM degrees and careers.