In connection with the increasing importance of chronic diseases the estimation of the prevalence of overweight and obesity becomes more and more important. Today these estimations are usually done via the Body-Mass-Index (BMI). For economic reasons BMI is often obtained by means of questionnaires or interviews. These (subjective) BMI-data show great differences to measured (objective) data. The differences between subjective and objective data and their dependence on age, gender and residence were investigated. Subjective and objective data show significant differences. On the basis of subjective data too many persons classify themselves as underweight or normal weighted and fewer persons classify themselves as overweight and obese. Variance analysis shows significant influences of gender and age. Women underestimate their BMI more than men. With increasing age the differences also increase. The estimation of BMI based on subjective data is inaccurate. In this way the prevalence of obesity and overweight are underestimated. That is why subjective data are not useful for clinical and epidemiological research, but it is interesting against the background of health psychology.