La subjetividad y la estadística en la investigación médico-social

Abstract

The most important task of statistic is to give objective quantitative alternatives to simple personal judgements. The significance tests constitute the better recognized expression of such a desire of objetivity. These tests are showed as if they could generate conclusions independently of the persons performing the research. However, this approach has important conceptual and practical limitations, and it is unable to carry out its original task. Anyway, subjectivity is not completely avoidable therefore to have a method able to combine the subjective background knowledge with the objective evidence of the particular experiment can provide better results for scientific analysis. The bayesian methods offer this possibility.