Overweight is an important public health problem affecting around 50% of the population of Wales, resulting in increased risk of illness, premature disability and premature death. The aim of this study was to examine critically the accuracy of self-reported data in describing the prevalence of overweight in Wales. A sample of 1622 adults aged 18 to 64 years was taken from the Welsh Heart Health Survey 1985. In that survey weight and height data were collected on a self-completed questionnaire and by clinical measurement. Mean differences between self-reported and measured weight and height were used as indicators of bias, and the accuracy of BMI and the prevalence of overweight based on this data were analysed. Weight was reported without significant bias in men, but women under-reported their weight by an average of 1.1 kg. Height was over-reported by 1.4 cm in men, and 0.7 cm in women, on average. More than two-thirds of subjects reported to within 2.3 kg and 2.5 cm of their actual weight and height. Reporting was more biased in older and overweight groups. The calculation of body mass index resulted in amplification of bias and underestimation of the prevalence of overweight and obesity in the study sample of 4.5% in men and 6.7% in women. The results have important implications for the use of self-reported data for the scientific measurement of the prevalence of overweight, especially in longitudinal studies, and suggest that further research should be conducted into the stability of reporting bias over time.