This paper presents a controlled experiment in which the goodness of using Theory Belbin roles for the integration of software development teams is explored. The study takes place in an academic environment with students from the Engineering of Software and compares the quality of the readability of the code generated by integrated teams with roles Compatible -according to the Theory of Belbin- and traditional teams, in our case, integrated teams with students selected randomly. The results provide positive evidence on the use of this theory and motivate researchers to continue studies in other activities related to the software development process; on the other hand, from a pedagogical perspective, the results obtained with the experiment, allow proposing as an alternative to integration teams, learning scenarios related courses Software Engineering, Theory of Roles of Belbin.
