I nuovi orizzonti dei metodi quantitativi

Abstract

New Perspectives in Quantitative Methods In this essay, new perspectives in quantitative methods are analyzed by comparing linear causal models, used in the social research of Paul F. Lazarsfeld, and mathematical adaptive models with the models of Artificial Neural Networks (ANN) that have been promoted actually in Italy by Massimo Buscema and his Semeion Group of Rome. This analysis of quantitative methods is examined by considering three paradigms (Kuhn) of sociological research: (a) the paradigm of objectivity, (b) The paradigm of action-research; (b) The paradigm of feminist methodology. The essay presents a map of the relations between the three paradigms in terms of specificity/generality and a study on educational issue with the application of an ANN which enables a reflection on the relations between explanation and prediction. The difference between linear causal models and Artificial Neural Network are explained as: differences in linear/non-linear mathematical models, different possibilities of using variables at different levels of scale, different paradigms in social research, differences in the explanation-prediction relation, and as different use of time in the research process.

Keywords

• action-research, artificial neural network, linear causal models, feminist methodology, paradigm of objectivity, metodologia femminista, modelli causali lineari, paradigma dell'oggettività, reti neurali artificiali, ricerca-azione