In a sample from the general population of school children of 15 years of age, we studied whether receiving information about the prevalence of headaches had any effect on their subsequent headache report. Sixty children in the fourth year at four secondary schools were allocated at random to two conditions: a biased condition emphasizing the high prevalence of headaches and a neutral condition. Subjects in the biased condition reported more headaches but they did not report more other physical symptoms than the subjects in the neutral condition. The results are discussed in terms of Pennebaker's theory on reporting symptoms. It is concluded that epidemiological research using the general population should deal more explicitly with the way in which subjects are motivated to participate.