Réflexions méthodologiques sur une enquête à passages répétés : l'EMIS de Bobo-Dioulasso

ObAiDOi Nassour and van de Walif Etienne ? Methodological reflection on multi-stage field surveys : EMIS survey of the region of Bobo-Dioulasso. The most difficult problem involved in multi-stage field surveys is caused by non-response for unknown reason In the EMIS survey carried out in the region of Bobo-Diou lasso between 1981 and 1984 to evaluate infant mortality rates, the reasons for bias could be traced First, birth problems were found in determining the place of delivery, particularly for women who has been moved from distant health centers, in the case of deliveries by Caresanan section, records of the delivery were often defective, particularly because record cards were discarded when the child died Secondly, women, particularly those in vulnerable categories, were lost in successive phases of the enquiry Although the questionnaire used in the survey was not designed to define the reason for non-response or to take into account temporary absences, it should be possible to evaluate differences in mortality by looking at those who did not respond at a particular stage of the inquiry separately.