This study aims to investigate and map references to online video resources in scientific publication and provide an overview of the relatively unexplored topic. Scientific articles containing references to videos uploaded on YouTube are examined and the methodology used to observe and describe video citation is influenced by previous research on citation analysis. This study uses a combination of context analysis and content analysis of the citations and their references. The location of video citation in the articles and the assessment of relevance indicates that referencing YouTube is meaningful for the scope of the citing articles and the most common purpose for citing uploaded video is to support and substantiate claims or assumptions. There are two key findings in the study. The first is the generated video citation, which is a unique type of citation only to be found when citing video. A generated citation is defined as a sort of self-citation, produced and uploaded by the author. In the case of generated citations, the author uses the multimodality and pedagogical properties of moving visuals to present their research. The other finding concerns the citation context and is the medium integral citation –a citation behavior where the type of medium with which the cited information is recorded, is mentioned within the citation context.