In order to plan, implement and monitor health interventions for the most deprived sector of the population, it is necessary to identify socioeconomic groups at risk. Multiple Correspondence Analysis was used to construct a socio-economic index based on data collected from a sample of 2698 households in South-West district of the Ugandan Republic in 1988. This study is a part of the baseline survey done by the Government of Uganda in collaboration with UNICEF. Its aim was to reduce the incidence of death of children below 5 years from diarrhea. Two factorial axes, representing respectively the socio-cultural and the anthropological conditions, explained more than 80% of the total variability. Among the 11 variables employed the most useful in characterizing the socio-economic classification were: father's occupation, parent's literacy, father's professional position and ownership of a radio. A classification in 7 levels was obtained. The first two levels are characterized as professionals and civil servants. The bottom two levels include households where both parents are illiterate and where father's primary activity is agricultural at a subsistence level. The three middle levels represent a transitional situation. In order to classify the family into the different levels, the other related variables, such as father's professional position or ownership of radio or father's religion or presence of latrine proved to be very useful. A flow chart which identifies which level a household belongs to was constructed. A general and valid observation is that families classified into the last two levels (6 and 7) constituted the population at risk for health conditions.