The importance of a sound evidence base for health promotion is recognized. Criteria for establishing evidence have, in the past, been heavily influenced by the bio-medical agenda. The problems in using experimental and quasi-experimental approaches and their limited relevance for evaluating health promotion interventions are examined. The multifactorial nature of the health promotion enterprise in relation to both inputs and outputs is recognized and a range of direct, indirect and process indicators are identified. The relevance of combining quantitative and qualitative methods for data collection is discussed and the paper concludes by advocating a 'judicial principle' for assessing evidence.