In this study, 21,085 tweets from university students over a period of twelve weeks were analyzed. The aim was to investigate the number and evaluative content of students' tweets related to their study, and to know whether the number of negative tweets decreases when teachers follow students on Twitter. About thirteen percent of all students' tweets relate to their study. This paper categorized these tweets in student-centered (10%) or curriculum-centered (3%), and divided both categories in negative (4%) or positive (9%) tweets. Teachers became followers of their students after six weeks, which led to a significant decrease of negative curriculum-centered tweets (from 2.2 to 0.7%), although a small decrease was observed even before the intervention started. Contrary to the teachers' initial fears, it is unlikely that students 'infect' each other with negative evaluations about the curriculum, since the number of negative tweets is very low overall.