Dualistic models are intuitive and often highly evocative. However, they posses no a priori greater theoretical interest than models with a different structure and their validity should be always empirically confirmed. This paper shows that we must proceed beyond dualistic modelling in the building of an actor model for economics and social sciences. We tested the performance of different actor models using experimental data. The empirical performance of dualistic models was actually lower than the one of plausible alternative models. This result suggests that the preference given by part of the sociological literature for dualistic model has (at least regarding actor modelization) little ground and that the empirical validation represents indeed a fundamental step in any model-building process.

Giangiacomo Bravo, in "SOCIOLOGIA DEL LAVORO " 105/2007, pp. 51-63, DOI: