Retrospective family videos of have been the main source for studies in early identification of ASD at 12 month. Despite its results there still is no consensus about which behavioral categories are reliable in identifying ASD at that age. Since retrospective family videos study methodology leads to evaluation of discrete behavioral categories, analysis of social interaction in early month has been defeated. The purpose of this study is to test a methodology to investigate early infant-adult interactions qualitatively. It was tested on family videos of two groups of infants: one with ASD diagnosis and other with infants that developed typically (DT). Findings indicate differences in social interaction between groups. Methodology was efficient in its purpose and also allowed quantitative analysis of behavioral categories that were compared to findings of others retrospective family videos studies. Implications about the use of social interactions as a measure for ASD identification at 12 month are discussed.