Policy makers and others charged with driving economic growth often assume a link between entrepreneurship education and business start-up. However, there is little by way of supporting literature in this regard, with few studies exploring impact measures that relate to actual venture creation. This paper considers two routinely collected data sets in the UK that directly relate to graduate self-employment and business start-up over a five-year period--the Higher Education-Business and Community Interaction survey (HE-BCI), and the survey of Destinations of Leavers of Higher Education (DLHE)--and explores whether the data can be used to assess impact. There is some evidence that HE-BCI may be affected by changes in data collection requirements, but it remains the most complete, extensive and useful longitudinal data set on graduate business start-up. DHLE would appear to be the best source of data for measuring the impact of both enterprise education and start-up support initiatives at an institutional level.