Propuestas metodológicas para la evaluación de programas públicos de formación. Una revisión crítica

Evaluation of public training programs has stronger developed in the last decades because of the use of causal inference methods. This paper shows an overview of the disposable methods of public training programs evaluation in the context of randomized experiments also in observational studies. Additionally, we evaluate the real possibilities of these methods. From a theoretical point of view, these methods can be developed in randomized experiments, although they are often designed over a small size of the population. From another point of view, public training programs evaluation can be implemented with observational methods. We can remark the three main observational methods. The first one is named selection on observables and includes three types of differentiated: subclasiffication, matching estimators and the propensity score. The second one is the difference in difference estimator. From our point of view, this is the more appropiate method when we evaluate public training programs. Finally, the last one use instrumental variables to develop the evaluation. This third method is out from this paper.