The increase of doctoral programs has generated the need to ensure their quality and with it the effectiveness of the mechanisms used. For this purpose, the objective is to generate a quantitative model for the meta-evaluation of doctorates taught in Chile. We used a sample of doctorate accreditation reports, with their criteria and sub-criteria, applying quantitative association techniques, visual representations such as networks, dendograms and probabilistic modeling. With this, a stochastic model was constructed to estimate the probabilities and the degree of objectivity of the accreditation. The characteristics and resources of each program is the criterion that most influences the accreditation results. The model establishes the probabilities of accreditation, as well as the random and subjective nature of doctorate accreditations, and reveals that doctoral accreditation emphasizes processes rather than resources and impacts. This metaevaluative model allows feedback to the doctoral accreditation system and can be replicated in quality assurance processes in other areas and systems.