A key requirement of successful initial teacher education is the development of professional vision, which includes shifting attention to features of the situation relevant to the specialized goals of teaching. Existing research hints at the value of targeted video-based courses in the development of professional vision, but often raises questions about the sources of shifts in the pattern of attention. We argue that existing work makes it difficult to distinguish whether shifts seen across video interventions are the results of the intervention, teaching experience, or methodological issues with the unbalanced use of videos in the data collection in these studies. Our first study suggests pre-service teachers' teaching practice experience does not notably affect attention, but that choice of video does. Our second study addresses the methodological issues and suggests that we may discount different or unbalanced videos as a source of the shifts in the pattern of attention. Finally, by introducing a new synthesis of the results in the literature, we identify a previously hidden key distinction between studies and suggest reasons why different studies have shown different results in this area.