The use of concept maps (CMs) as an evaluation tool of students' learning was explored during the Natural Science course assigned to all 1styear higher education students at School of Arts, Sciences and Humanities (University of São Paulo). The richness and diversity of propositional networks produced by the students made the assessment task more complex than the evaluation of multiple-choice tests or open questions. A procedure for comparing the evaluation made by the professor (P), three experts in CMs (E) and the students (S) was used to appraise a set of CMs (n = 109). CMs' focal question explored the conceptual relationships among astronomic observations made during the 16$^{th}$ century, the birth of modern science and, the increase of our understanding about the universe. Correlation graphs were plotted to compare these evaluations in pairs (PxE, PxS and ExS). The results confirmed the complexity of the MCs evaluation, which asks for changing the procedures frequently used in the classroom. The possibility of including the students in the evaluation process is confirmed by the correlation found between the assessments made by them and the professor.
concept maps; learning evaluation; higher education; meaningful learning; peer review
