Unidades de análisis y falacias inferenciales en ciencias sociales.

Abstract

A central problem of quantitative research in the social sciences is the "construction of the datum". This process comprehends, as one of its stages, the selection and adequate handling of the units of analysis and the relations among them. Using an example from the sociology of the family, the article examines different problems to which a less than rigorous handling of the units of analysis may lead and, at the same time, develops some basic con¬siderations for overcoming these problems. The methodological reflection is complemented by an operative example of the treatment of structured data bases, with information on surveys, in a statistical package designed for the social sciences.