Objective: The objective of this study was toanalyze bibliometric data from ISI, NationalInstitutes of Health (NIH)–funding data, andfaculty size information for Association of AmericanMedical Colleges (AAMC) member schools during1997 to 2007 to assess research productivity andimpact.Methods: This study gathered and synthesized 10metrics for almost all AAMC medical schools(n5123): (1) total number of published articles permedical school, (2) total number of citations topublished articles per medical school, (3) averagenumber of citations per article, (4) institutional impactindices, (5) institutional percentages of articles withzero citations, (6) annual average number of facultyper medical school, (7) total amount of NIH fundingper medical school, (8) average amount of NIH grantmoney awarded per faculty member, (9) averagenumber of articles per faculty member, and (10)average number of citations per faculty member.Using principal components analysis, the authorcalculated the relationships between measures, if theyexisted.Results: Principal components analysis revealed 3major clusters of the variables that accounted for 91%of the total variance: (1) institutional researchproductivity, (2) research influence or impact, and (3)individual faculty research productivity. Dependingon the variables in each cluster, medical schoolresearch may be appropriately evaluated in a morenuanced way. Significant correlations exist betweenextracted factors, indicating an interrelatedness of allvariables. Total NIH funding may relate morestrongly to the quality of the research than thequantity of the research. The elimination of medicalschools with outliers in 1 or more indicators (n520)altered the analysis considerably.Conclusions: Though popular, ordinal rankings cannotadequately describe the multidimensional nature of amedical school’s research productivity and impact. Thisstudy provides statistics that can be used in conjunctionwith other sound methodologies to provide a moreauthentic view of a medical school’s research. The largevariance of the collected data suggests that refiningbibliometric data by discipline, peer groups, or journalinformation may provide a more precise assessment.