Background: Circuit resistance training (CRT) should promote favorable kinematics (scapular posterior tilt, upward rotation, glenohumeral or scapular external rotation) to protect the shoulder from mechanical impingement following paraplegia. Understanding kinematics during CRT may provide a biomechanical rationale for exercise positions and exercise selection promoting healthy shoulders. Objective: The purposes of this study were: (1) to determine whether altering hand position during CRT favorably modifies glenohumeral and scapular kinematics and (2) to compare 3-dimensional glenohumeral and scapular kinematics during CRT exercises. Hypotheses: The hypotheses that were tested were: (1) modified versus traditional hand positions during exercises improve kinematics over comparable humerothoracic elevation angles, and (2) the downward press demonstrates the least favorable kinematics. Design: This was a cross-sectional observational study. Methods: The participants were 18 individuals (14 men, 4 women; 25-76 years of age) with paraplegia. An electromagnetic tracking system acquired 3-dimensional position and orientation data from the trunk, scapula, and humerus during overhead press, chest press, overhead pulldown, row, and downward press exercises. Participants performed exercises in traditional and modified hand positions. Descriptive statistics and 2-way repeated-measures analysis of variance were used to evaluate the effect of modifications and exercises on kinematics. Results: The modified position improved kinematics for downward press (glenohumeral external rotation increased 4.5° [P=.016; 95% CI=0.7, 8.3] and scapular external rotation increased 4.4° [P<.001; 95% CI=2.5, 6.3]), row (scapular upward rotation increased 4.6° [P<.001; 95% CI=2.3, 6.9]), and overhead pulldown (glenohumeral external rotation increased 18.2° [P<.001, 95% CI=16, 21.4]). The traditional position improved kinematics for overhead press (glenohumeral external rotation increased 9.1° [P=.001; 95% CI=4.1, 14.1], and scapular external rotation increased 5.5° [P=.004; 95% CI=1.8, 9.2]). No difference existed between chest press positions. Downward press (traditional or modified) demonstrated the least favorable kinematics. Limitations: It is unknown whether faulty kinematics causes impingement or whether pre-existing impingement causes altered kinematics. Three-dimensional modeling is needed to verify whether “favorable” kinematics increase the subacromial space. Conclusions: Hand position alters kinematics during CRT and should be selected to emphasize healthy shoulder mechanics.