Context  Since 2000, problem-based learning (PBL) seminars have been introduced into the curriculum of medical studies at the University of Liège. We aimed to carry out a cross-sectional investigation of the maturational increase in biomedical reasoning capacity in comparison with factual knowledge retention throughout the curriculum.

Methods  We administered a factual knowledge test (i.e. a true/false test with ascertainment degree) and a biomedical reasoning test (i.e. an adapted script concordance test [SCT]) to 104 students (Years 3–6) and a reference panel. The selected topic was endocrinology.

Results  On the SCT, the students obtained higher scores in Years 5 and 6 than in Years 3 and 4. In Year 3, the scores obtained on SCT questions in a new context indicated transfer of reasoning skills. On the true/false test, the scores of Year 3 students were significantly higher than those of students in the other three year groups. A positive correlation between SCT scores and true/false test scores was observed only for students in Years 3 and 4. In each group, the ascertainment degree scores were higher for correct than for incorrect responses and the difference was calculated as an index of self-estimation of core knowledge. This index was found to be positively correlated to SCT scores in the four year groups studied.

Conclusions  Biomedical reasoning skills are evidenced early in a curriculum involving PBL and further increase during training. This is accompanied by a decrease in factual knowledge retention. The self-estimation of core knowledge appears to be related to reasoning capacity, which suggests there is a link between the two processes.