Using data from two cohorts of graduates, this article examines three aspects of over-education. First, using three new measures, we present an estimate of graduate over-education in the UK. We find that the scale of over-education varies with measurement techniques, with weak correlations between the three measures. Second, across the three measures we estimate the effects of over-education on earnings and job satisfaction. The effects of over-education on earnings and job satisfaction are similar, not withstanding the measures identifying different individuals as being over-educated. One finding is that the effects of being over-educated are more significant for female graduates than male, although it is ambiguous which gender is more prone to over-education. Third, we examine another source of ambiguity regarding over-education, namely that firms upgrade the tasks they allocate to their employees who appear to be over-educated. We find that, for graduates, job quality for the over-educated is not converging to that of the appropriately educated.