Une enquête sur la fécondité en Algérie. Application de l'analyse factorielle des correspondances

SUMMARY This is a methodological article whose aim is to emphasize the interest of proximity analysis for small sample surveys which collect a a lot of variables to be used for an overall and simultaneous analysis. Although this method is only a descriptive one, it is more satisfactory than correlation analyses between two variables whose interpretation is not always reliable and which do not permit an available classification. The progress is therefore obvious : the point is to emphasize and classify the relations between variables according to their intensity. The use of this method in a fertility survey in Algeria shows that education and mass media are the main variables which determine a traditional or modern general attitude. Education is the most " discriminant " variable followed by income, occupation, urban or rural residence, region, age at marriage and also by some general attitudes concerning female activity, women participation to political life or in connection with birth control. The sample is more homogeneous for fertility than for the other variables, i.e. Algerian couples are less different by their fertility than by their other economic, social or cultural characteristics. The influence of other variables such as religion or frequency of sexual intercourses upon general behaviour or fertility seems to be very weak. The analysis of the first projection plan makes it possible to establish a more complex classification of the correlated variables.