In open prospective recording the participating doctors may be influenced by the study itself. There are two factors contributing to this influence: the doctor's status as an observer and as an observed subject. A model for analysing this influence is presented. In a Norwegian general practice study of the recording of conditions suspected of being cancer the changes owing to the recording itself have been studied by three different means: indicator comparison, time-trend analysis and post-registration enquiry. The study showed that the effect of the recording was small, and this was perhaps a result of prophylactic efforts by the doctors to minimize the influence. An influence on the doctors' actions and thinking during a period of registration should always be expected, and control methods to give a semi-quantitative expression of the consequences of this influence should therefore be carried out.