Abstract
It has been claimed that mothers' drinking during pregnancy may affect the neurodevelopment of around 1% of all children. If this is true, then prenatal alcohol exposure represents an important risk factor for neurodevelopmental problems, giving rise to a large burden of disability which could be potentially preventable. Evidence to support this idea has come from animal experiments and human observational studies. However, such findings need to be supported by more robust research designs. Because randomized controlled trials in this area are neither feasible nor ethical, suggestions are made for further research making more use of natural experiments.