We present a model for assessing the evolution of grades in higher education students. Sequential assessment encourages pupils to improve their marks in a general framework of feedback process. We have implemented the system in a Psychology degree annual subject two consecutive years. The contents of the subject were divided up according to the official program into theoretical and applied parts. First, theoretical part was assessed with four mid-course examinations which mean the 66.67% of the overall grade. On the other hand, the practical aspects (33.33% of the overall grade) were assessed by asking students to write an assay on two short-term research projects and by marking reading discussion in several sitting class. We have found a rise on the average marks as the program moves on and the process has been modelled and validated with a naïve Bayesian classifier. The Bayesian net is able to predict student grades and it improves the predictions as a function of the number of assessments. Our results are interesting in order to develop more efficient assessments methods and it could be used for designing automated intelligent tutoring systems.