To use video in educational research has become more and more common in the last few decades, including methodologies where informants themselves use video-cameras for documentation. The purpose of the article is to discuss affordances and constraints of using video-diaries as a data generation method for investigating students' identity constitution. Video-diaries were recorded as part of a larger project, where the empirical data also included observations and videorecordings of teaching and semi-structured interviews. The noninterference of the researchers during the video-diaries was found to be both a strength, in that students more freely could tell their own stories, and a weakness, in that it put high demands on the students' ability to express themselves in monologue format. An important affordance of the video-diaries was that they contributed to a "thick" data set, both in that they informed our individual, semi-structured interviews and allowed us to quickly move on to in-depth conversations, and in that the students were able to utilize artifacts and show environments of importance to them.