Modeling middle school students preconceptions and inclination to STEM-related careers using neural networks

Abstract

We conduct a survey of three distinct populations of Grade 8 female students to determine students' preconceptions toward STEM careers and their inclination to select a STEM career in the future. A set of 29 questions was given to each student and the responses are recorded and analyzed using standard one-way ANOVA. Furthermore, we model the students' responses to the survey using a multilayer feed-forward artificial neural network. We find that students prioritize STEM subjects in school to get into a STEM-related career in the future, indicating that initial personal interest highly contributes to students' decision to go to a STEM career.