To test the predictive validity of a selection system for Senior House Officers (SHOs) and registrars in anaesthetics, 140 doctors short-listed from 635 applications between 1980 and 1987 were assessed by a semi-structured interview assessed and a personality questionnaire (Cattell 16PFQ-form C). The 62 doctors selected were followed up for between 3 and 8 years. Future performance was predicted from the psychological tests and by the interviewers. Academic, clinical, behavioural, and overall performance were used as criteria of outcome. Correlation coefficients between prediction and outcome measures were statistically highly significant (P < 0.01). Using multiple regression, equations could be derived from five of the Cattell personality factors to predict overall performance. Personality measures discriminated significantly between the best and poorest performers. Interview predictions were also statistically significant (P < 0.01). The method provides a blueprint for the effective selection of junior anaesthetists. Wastage in terms of those leaving the specialty was 16%.