We explore the relationship between earnings, education, and fixed-term contracts using data from the 1997 British Social Attitudes Survey. We find that the log hourly wage of workers employed under such contracts is approximately 13% lower than that of their 'permanent' counterparts, even after controlling for a plethora of personal and job characteristics. Standard decompositions indicate that the vast majority of this differential (more than 70%) is attributable to price effects, which may reflect discrimination on the part of the employer. Such findings may, therefore, suggest that employment protection is appropriate for individuals employed under fixed-term contracts. n University of Leicester nn University of Bath 1 We interpret the term 'fixed-term' as an employment contract of definite duration. 2 The proportion of workers on fixed-term contracts in France, Italy, Spain and the UK increased respectively from 4. 7%, 4.8%, 15.6% and 7.0% in 1985 to 13.1%, 8.2%, 33.6%, and 7.9% in 1997. Approximately 12% of the 150 million strong European Union workforce were employed on fixed-term contracts in 1997 (European Commission, 1999).