Student Dropout Prediction

In the first year of my master`s degree, I worked as a research assistant at the Methods Center at the University of Tübingen. There, I developed a model to predict student dropouts in STEM classes using multivariate time-series analysis, specifically the Kalman Filter, to enhance the educational environment. We combined a psychometric approach, utilizing latent characteristics such as ability, motivation, and stress levels from questionnaires for prediction. The project received support from the Ministry of Education of Baden-Württemberg.