We tested the validity of the paddle method for measuring both the kinesthetic and visual-kinesthetic perception of inclination. In three conditions, subjects performed three different tasks: (1) rotating a manual paddle to a set of verbally given inclinations (blindfolded subjects), (2) rotating a manual paddle to the same set of verbally given inclinations after specific kinesthetic training (blindfolded subjects), and (3) rotating the paddle to a set of fixed visual inclinations after the kinesthetic training. The results showed a high degree of accuracy and precision in the second and third task but not in the first one. When subjects were asked to rotate a manual paddle to a set of verbally given inclinations, they used three main anchors (0°, 45°, 90°). Furthermore, the paddle method is biased by a kinesthetic deficiency, namely a rotational problem of the wrist that can be corrected by means of specific training.