The Government and its agencies have seriously overestimated the impact of the specialist schools programme on educational attainment. The substantially higher exam scores achieved on average by schools with specialist status are due primarily to sample selection bias and not to any benefits flowing from subject specialisation itself. A fixed effects model is used on the panel of maintained secondary schools in England covering the period 1992-2005 to obtain this result. It is found, however, that the specialist schools programme has had beneficial distributional consequences. There is evidence that schools with the highest proportion of pupils eligible for free school meals have experienced by far the biggest improvement in exam results as a consequence of acquiring specialist status.