Cost-effective prevention strategies to prevent weight gain and the development of obesity should be based on appropriate knowledge of the determinants of weight gain. The body of evidence on the dietary determinants of weight gain is, however, fragmentary at best, partly because inappropriate research methods are used to study the determinants of obesity under normal circumstances. Evidence from studies using experimental diets have shortcomings because of their short duration and selection of highly-motivated subjects and because the outcomes can be easily influenced by the choice of foods to be used in the intervention. Of the observational studies, many have severe methodological shortcomings, e.g. ecological studies, cross-sectional surveys and classical cohort studies in which the baseline diet is linked to subsequent weight development over long periods of time. Longitudinal studies with repeated measurements in which changes in diet and physical activity are linked to changes in weight are probably the most informative, but these studies are relatively rare. There is a great interest in interventions that are effective and efficient for the prevention of obesity. Many countries and research funding agencies seem to show a strong tendency to develop interventions for children and adolescents exclusively. It can be easily shown, however, that intervention programmes are much more likely to be cost-effective in older adults than in children, which indicates that adults should not be neglected as target populations for obesity prevention. Obesity prevention should follow a life-course approach, as currently recommended for non-communicable diseases in general by the WHO.