From the data of the Third Regional Study Comparative and Explanatory, effect sizes of direct indirect relations of four indices generated on the test score in science (sixth grade) were evaluated. Generated indexes were: willingness to learn, teacher attendance, willingness to teach perceived by the student, achievement expectations and index monitoring, considered critical in the generation of a learning and assessment context. The results show that the proposed model accounts for 5.8% of the variance, highlighting the contribution factor teacher attendance (perceived by the student). Colombia data as an example case of consistency of the overall results are presented. The results are discussed in terms of theoretical importance of all evaluated relations and methodology. Limitations of existing information identified in the study TERCE as to the expectations of students and educational actions (teacher) that do not allow the general assessment of the interaction between the agents in educational system.