Apports et limites de l'analyse des données

Exploratory data analysis applied to social sciences has become fashionable in France during the seventies. Three facts seem to have worked in that way : the advances in computing art, the relative indifference toward probability criteria for inference, and the flexibility of application. The latter may be somewhat risky because the success of the approach depends on the knowledge and ability of the investigator. As any descriptive statistical method, data analysis precedes the procedures of formal statistical inference. Actual economic examples based oh correspondence analysis and cluster analysis are used for illustrative purposes.