Background
Although several studies (Anat Sci Educ, 8 [6], 525, 2015) have shown that computer‐based anatomy programs (three‐dimensional visualisation technology [3DVT]) are inferior to ordinary physical models (PMs), the mechanism is not clear. In this study, we explored three mechanisms: haptic feedback, transfer‐appropriate processing and stereoscopic vision.
Methods
The test of these hypotheses required nine groups of 20 students: two from a previous study (Anat Sci Educ, 6 [4], 211, 2013) and seven new groups. (i) To explore haptic feedback from physical models, participants in one group were allowed to touch the model during learning; in the other group, they could not; (ii) to test ‘transfer‐appropriate processing’ (TAP), learning ( PM or 3DVT) was crossed with testing (cadaver or two‐dimensional display of cadaver); (iii) finally, to examine the role of stereo vision, we tested groups who had the non‐dominant eye covered during learning and testing, during learning, or not at all, on both PM and 3DVT. The test was a 15‐item short‐answer test requiring naming structures on a cadaver pelvis. A list of names was provided.
Results
The test of haptic feedback showed a large advantage of the PM over 3DVT regardless of whether or not participants had haptic feedback: 67% correct for the PM with haptic feedback, 69% for PM without haptic feedback, versus 41% for 3DVT (p < 0.0001). In the study of TAP, the PM had an average score of 74% versus 43% for 3DVT (p < 0.0001) regardless of two‐dimensional versus three‐dimensional test outcome. The third study showed that the large advantage of the PM over 3DVT (28%) with binocular vision nearly disappeared (5%) when the non‐dominant eye was covered for both learning and testing.
Conclusions
A physical model is superior to a computer projection, primarily as a consequence of stereoscopic vision with the PM. The results have implications for the use of digital technology in spatial learning.