La misurazione della corruzione attraverso le sentenze della magistratura: una proposta metodologica con strumenti di text mining

The definition of the methods of measurement of corruption is a central topic in the scientific debate and in multilateral contexts where the policies of contrast are developed at a global level. The Authors start from the recognition of the methods used up to now, highlighting the limits and criticalities that characterize the so-called subjective or perceptive measuring instruments. On this premise, the article proposes an innovative methodology, of an interdisciplinary nature, through which to achieve anobjective representation of the corruption phenomenon, as emerges from the judicial sources. In particular, the method presented is based on the application of text mining and statistical techniques to an objective and stable information base, represented by a significant sample of sentences handed down by the Supreme Court of Cassation, in the period between 2015 and January 2020, regarding corruption offenses (arts. 317, 318, 319, 319-quater, 321 of the penal code). This article offers a preliminary exposition of the shown method and highlights its cognitive and applicative potential, which are not limited to the (quantitative) measurement of the observed phenomenon. The automatic analysis, through the text mining, allows to thoroughly interrogate the judicial source (judgments) in order to obtain extremely significant information from the scientific perspective of the qualitative description of the variegated corrupting phenomenology. Finally, the objective-based method presented represents a promising tool for the development of evidence-based policy