To analyse the association between nutritional and familial factors and obesity in boys and girls. Randomized, cross-sectional population study. Province de Luxembourg, Belgium. One thousand and twenty-eight boys and girls in age strata 6-8, 8-10 and 10-12 y, comprising 70.3% of primary cohort. Examinations included anthropometric measurements and questionnaires covering familial, socioeconomic and psychosocial factors. A three day dietary record was obtained in 955 children. This was analysed in relation to the anthropometric data. In comparison with similar studies from other regions and recommended allowances, the intakes of total energy, fat, particularly saturated fat and cholesterol, were high, while consumption of carbohydrate and fiber was low, as well as the polyunsaturated/saturated ratio of fat. Total energy intake showed no or weakly significant correlations with anthropometric factors. However, total fat (P=0.045) and saturated fat (P=0.0005) intake showed consistent positive correlations with body mass index (BMI, kg/m2) and skinfold thickness, with corresponding negative relationships to carbohydrate intake (P=0.034) in boys. Such relationships were also found when calculated as energy density. These associations were not statistically significant in girls. The high fat, low carbohydrate pattern of the nutritional status seemed to be more pronounced in families where the father had a low level of education (lipids, boys, P=0.0007), and where both parents were obese (saturated fat, boys, P=0.023), suggesting involvement of socioeconomic and familial factors. The lack of correlation between factors indicating obesity and total energy intake suggests that the positive energy balance causing obesity is due mainly to a low energy output. However, since energy intake measurements are imprecise, overeating can not be excluded, particularly since elevated consumption of food with high contents of fat, found in these children seems to be poorly regulated.