Causal models in the social sciences

Abstract

This chapter deals with the review of causal models in the social sciences. Causal modeling is an umbrella term for a variety of techniques that are used to make causal inferences from statistical data. These techniques take many different forms and have a number of names—for example, regression, simultaneous or structural equations, factor analysis, use of path models, and much else as well. Such techniques are widely used in the social sciences, particularly in disciplines such as sociology and political science that lack powerful and generally agreed upon formal theories. They are also widely employed in certain areas of psychology and in some bio-medical contexts such as epidemiology. Causal modeling techniques are used for the causal analysis of experimental data, but much of their interest stems from the fact that they are also used to make causal inferences from non-experimental or “observational” data. Causal modeling techniques raise a number of philosophical and methodological issues that should be of interest both to philosophers of science and to users of those techniques.