The effectiveness of clay modeling to written modules is examined to determine the degree of improvement in learning and retention of anatomical 3D relationships among students with different learning preferences. Thirty-nine undergraduate students enrolled in a cadaver dissection course completed a pre-assessment examination and the VARK questionnaire, classifying learning preference as visual, auditory, read/write, or kinesthetic. Students were divided into clay, module, and control groups with preference for learning style distributed among groups. The clay and module groups participated in weekly one-hour classes using either clay models or answering written questions (modules) about anatomical relationships, respectively. The control group received no intervention. Post-assessment and retention examinations were administered at the end of the semester, and three months later, respectively. Two variables (Δ1, Δ2) represented examination score differences between pre- and post-assessment and between post-assessment and retention examinations, respectively. The Δ1 for clay and module groups were each significantly higher than controls (21.46 ± 8.2 vs. 15.70 ± 7.5, P ≤ 0.05; and 21.31 ± 6.9 vs. 15.70 ± 7.5, P ≤0.05, respectively). The Δ2 for clay and module groups approached but did not achieve significance over controls (−6.09 ± 5.07 vs. −8.80 ± 4.60, P = 0.16 and −5.73 ± 4.47 vs. −8.80 ± 4.60, P = 0.12, respectively). No significant differences were seen between interventions or learning preferences in any group. However, students of some learning styles tended to perform better when engaging in certain modalities. Multiple teaching modalities may accommodate learning preferences and improve understanding of anatomy. Anat Sci Educ. © 2013 American Association of Anatomists.