A growing number of studies are showing that internship-style undergraduate research experiences (UREs) and course-based UREs (CUREs) can influence students’ persistence in science. This has led to rapid growth in CUREs/UREs with limited insight into the features of research experiences that are necessary and sufficient to influence student persistence. We are addressing this by adapting methods from clinical psychology and natural language processing, namely Ecological Momentary Assessment (EMA) and topic modeling (a statistical model for identifying recurring patterns of words), to characterize features of CUREs/UREs as learning environments.