Mathematical competence is a key requirement for development in different areas of society. However, very few 15 years old students in Ecuador manage to develop it adequately during their primary and secondary education. Therefore, it was decided to create a multilevel predictive model of academic performance in mathematics from the results of PISA-D 2018, which provides information on sociodemographic, economic and educational factors associated with performance. For this reason, the study follows a mixed approach that includes a systematic literature review, statistical modeling, focus groups, design, implementation and evaluation of an intervention proposal in a selected sample. The research is in the second phase, which includes the analysis of quantitative data from students and schools in Ecuador. Therefore, a plausible value of the Mathematics area has been randomly chosen to combine it with the contextual, individual, teacher and school variables. Subsequently, the same procedure will be carried out, but considering the 10 plausible values that PISA-D gives to the area.