2019 ASHS Annual Conference

Classic student perception metrics, like surveys and course evaluations, offer limited insight into student content retention and student perception of individual course elements. Recently-developed algorithms facilitate analysis of text data, but their application to education research is still limited. This study used text mining and sentiment analysis to gauge student perception and content retention in a protected agriculture course. Students participated in seven different faculty-led hands-on activities. After each activity, students submitted a short essay summarizing their experience. Student submissions were de-identified and reformatted. Then, word frequencies, lexical depth, lexical density, and sentiment analysis were extracted from the text using the emotion and opinion lexica. Student essays included terms associated with joy, trust, surprise, disgust, and fear, but not sadness. “Exciting” was the most common word related to positive sentiments throughout the course, while “difficult” was the most common word related to negative sentiments. “Learned”, “experience”, “activity”, “plants”, and “protected agriculture” were present at similar frequencies in all essays. Mean sentiment scores were greater than +2.0 in all activities, suggesting that hands-on activities had a positive impact on student experience. Essay lexical depth and density increased over time, suggesting that students incorporated technical terminology to their horticulture practice. This was one of the expected student learning outcomes for the course. Altogether, text mining student essays provided insight into student perception and content retention in the course.