69 patients without shoulder pain were randomized to thoracic spinal manipulation therapy (SMT), sham SMT, or to a sham-ultrasound group to determine the best sham comparator for thoracic SMT. They were told they were assigned to either manual therapy, range of motion, or ultrasound, in order to measure the perceived effects of the treatments on shoulder motion, pain and function. SMT and sham-SMT had similar believability, but fewer patients believed in sham-ultrasound treatment. The treatment perception was similar and no change was observed in shoulder flexion in all groups. SMT improved shoulder internal rotation AROM, compared to both shams. These results suggested that sham-SMT was a more believable comparator to thoracic SMT than sham-ultrasound.