Research into the communication and interactions in classrooms need to take the multimodal nature of classrooms into account. Video cameras can capture the dynamics of teaching and learning, but the use of videos for research purposes needs to be well thought through in order to accommodate the challenges this tool holds. This article refers to three research projects where videos were used to generate data. It is argued that videos allow the researcher to hone in on the micro-details and, in contrast to other data generation tools, allows researchers who were not present at the time to view what has been witnessed. A video recording is a data source but not data by itself and the information that is discerned from a video is framed and shaped by the research paradigm and the questions asked.