Course publication lists are often used to illustrate the benefits of research-oriented teaching, yet the number of papers does not identify how many students contributed or what they learned. We examined the publication and assessment records of a graduate oncology course to establish which conclusions the published material supports. The analysis covered five annual enrollment entries, 18 publication entries, a presentation rubric, and a grading schedule. A DOI-linked register separated publication identity, course year, printed reference year, research activity, and topic. We then compared alternative counting rules and derived conservative bounds on the number of enrolled authors. The 18 publications had only 16 distinct author-year labels. Ten were primary investigations and eight were literature syntheses; 16 focused on cancer. Annual counts were 2, 4, 4, and 8 by course year, compared with 2, 4, 2, and 10 by printed reference year. Synthesis remained at two works per course year, so investigations accounted for the entire six-work increase between the first and last rows. The pooled count was 32.14 works per 100 registrations, falling to 24.39 when the in-progress 2024 course was excluded. These quotients could not identify the proportion of students who published, and the assessment material did not support a calculation of learning gains. The case shows why publication identity and counting rules need to be reported explicitly, while claims about student participation require verified authorship links and claims about learning require direct assessment.
