Using high school and district economic variables to predict engineering persistence

Prior research has shown that Peer Economic Status (PES), a socioeconomic indicator based on a school’s free lunch participation, is predictive of enrollment in engineering, first-year GPA, and engineering degree completion. In that study, PES was calculated as an average over the entire time period (1987-2004). To further explore the utility of this variable two new time-variant forms will be used, computed at the school-level and the district-level. Academic variables are drawn from the Multiple Institution Database for Investigation of Engineering Longitudinal Development (MIDFIELD) database and high school codes are used to link data from the National Center for Education Statistics (NCES). The time-variant PES is calculated from the four years each student is expected to have been in high school. Additionally, a new algorithm for the treatment of missing values is utilized. The district economic status (DES) is computed in a similar fashion. A series of logistic regression models is used to determine the impact of school- and district-level economic status variables on six-year degree completion. Results show that the time-averaged measures are stronger indicators of engineering persistence than the time-variant measures and that school-level variables are better predictors than district-level variables. Additionally the importance of context in interpreting socioeconomic variables is highlighted.