Problems of recurring corruption have stimulated major reforms in policing in many countries in the last 30 years. Considerable advances have been made in recruitment, training, complaints investigations and external oversight of conduct. However, continuing problems have prompted a search for more effective forms of misconduct prevention. This article examines the situation in Australia in relation to the emerging and controversial anti-corruption strategy of integrity testing. The study is concerned with `integrity tests' that simulate misconduct opportunities for serving officers-not pre-employment screening tests. The eight police agencies in Australia were asked to supply information on planned or implemented testing programmes as well as information on policy perspectives and debates. Only three jurisdictions were identified as conducting targeted testing. The success of these programmes in identifying misconduct suggests this may be an essential anti-corruption device and leaves a question mark over the adequacy of accountability in jurisdictions without this capacity. Two other agencies were planning to introduce targeted testing. While some agencies had given serious consideration to random integrity testing, legal, ethical and practical concerns have meant that no programmes have been introduced except for drug and alcohol testing. The latter also appears to be a useful tool to improve police conduct. The article concludes with a theoretical review of the possible benefits of randomized testing as a form of behavioural audit.