Acromial morphology is one of the important factors related to rotator cuff pathology, especially if there is acromial encroachment onto the supraspinatus tendon. This is the rationale for many surgeons performing an acromioplasty, based on preoperative radiographs, in their treatment of rotator cuff disease. However, geometric measurements of the acromion are dependent on the radiographic view, and suboptimal radiographs of the scapula might bias the acromial morphology on the radiographs. Our objective was to study the effect of suboptimal lateral-view radiographs of the scapula on 6 acromial angles, which have been correlated with rotator cuff pathology, by use of a computational method. Anatomic landmarks were digitized on 12 cadaveric scapulae via a spatial linkage system. A coordinate system based on the scapula was then created that allowed projection of the anatomic landmarks on the sagittal plane and simulated the optimal lateral-view radiograph of the scapula. By rotating the coordinate system in the coronal plane (cranial and caudal projections) and axial plane (anterior and posterior projections) with 5 degrees, 10 degrees, 15 degrees, and 20 degrees increments, 16 suboptimal lateral-view radiographs of the scapula were simulated. Significant differences in the 6 acromial angles (P < .01) between the optimal and suboptimal lateral-view radiographs of the scapula were found in almost every cranial and caudal projection. Rotating the projection in the axial plane did not significantly affect the acromial angles (P > .05) up to 20 degrees of rotation in the anterior and posterior direction. Of the 6 different acromial angles, the acromial slope was least affected when the projection was varied. In contrast, the anterior slope of the acromion was highly affected by errors in the projection.