[Methodology for household surveys: the case of two investigations carried out (example of Bandafassi, Senegal)].

The survey area of Bandafassi, in Eastern Senegal, was the subject of two studies each allowing to measure the risk of bite and death consecutive to a snakebite in a rural area. Two different methods were being used to carry out these investigations. Although both were based on the visit of households, the sample of the first study was designed to be representative of the population (the zone of study of Bandafassi is the subject of a regular demographic follow-up since 1970) while the second study was carried out by interviewing the concession (groups of households) heads. Therefore, the analysis of these surveys must take into account the sampling discrepancies. In spite of the different methodologies, similarities in the results obtained validate both methods and permitted to explain the small variations observed and to measure bias relative to this type of investigations.