Purpose. To compare evaluation methods for identifying third-year medical students whose funds of knowledge are marginal. Method. The written evaluation forms and comments from a formal evaluation session for 124 students in the inpatient medicine clerkship in 1992-93 at the Uniformed Services University of the Health Sciences F. Edward Hebert School of Medicine were reviewed. The written and verbal ratings of a student's general fund of knowledge were converted to similar five-point categorical scales. For each evaluation method, chi-square analysis was used to compare the students identified as having marginal funds of knowledge with those who scored ≤300 on the end-of-clerkship NBME subject examination. Results. Sixteen students scored ≤300 on the NBME subject examination. For the checklist descriptors, ratings of 'marginal' identified three of these 16 students (a sensitivity of 19%). For the written comments on the evaluation form, ratings of 'marginal' identified four of the 16 (a sensitivity of 25%). For the formal evaluation session, however, ratings of 'marginal' identified seven of the 16 (a sensitivity of 44%). The specificity of a 'marginal' rating was excellent (≤95%) for all three evaluation methods. Chi-square analysis was significant for each method (p ≤ .01). Conclusion. Although the ability of housestaff and faculty to identify students with weak funds of knowledge may be less than ideal, it may be improved by the routine use of a formal evaluation session. If done during the clerkship, this would allow for a specific plan of remediation to be designed with the instructors for the students at risk.