A social marketing approach used both qualitative and quantitative methods to develop a hygiene behaviour intervention in rural north-east Thailand. Behaviours were preselected from a previous study and the intervention was designed to promote hand washing, especially before feeding a baby, cooking, eating, and after defaecation or cleaning a baby's bottom, and dish washing immediately after eating. A bacteriological indicator (enumerating faecal streptococci using a finger impression technique) was developed to measure changes in hand washing behaviour and observation (spot checks) of dirty dishes to indicate dish washing practice. There was a significant improvement in both behaviours and a significant reduction in diarrhoeal disease as a result of the intervention. Furthermore, both indicators were retrospectively found to be positively related to diarrhoeal disease incidence. However, receiving and being able to recall the intervention messages was not necessarily sufficient to ensure behaviour change, as some adults found it difficult to change old habits. Villages showing the greatest improvement tended to have a stronger sense of community than others and to have more people actively involved in the intervention.