The purpose of this study was to determine whether strength deficits could be detected in individuals with and without shoulder impingement, all of whom had normal shoulder strength bilaterally according to grading of manual muscle testing. Strength of the internal rotators and external rotators was tested isokinetically at 60 degrees /s and 180 degrees /s, as well as manually with a handheld dynamometer (HHD) in 17 patients and 22 control subjects. Testing was performed with the shoulder positioned in the scapular plane and in 90 degrees of shoulder abduction with 90 degrees of elbow flexion (90-90). The peak torque was determined for each movement. The strength deficit between the involved and uninvolved arms (patients) and the dominant and nondominant arms (control subjects) was calculated for each subject. Comparisons were made for the scapular-plane and 90-90 positions between isokinetic and HHD testing. Despite a normal muscle grade, patients had marked weakness (28% deficit, P < .01) in external rotators at the 90-90 position tested with the HHD. In contrast, external rotator weakness was not evident with isokinetic testing at the 90-90 position (60 degrees /s and 180 degrees /s, 0% deficit, P = .99). In control subjects, greater internal rotator strength in the dominant compared with the nondominant arm was evident with the HHD at the 90-90 position (11%, P < .01) and in the scapular plane (7%, P < .05). Using an HHD while performing manual muscle testing can quantify shoulder strength deficits that may not be apparent with isokinetic testing. By using an HHD during shoulder testing, clinicians can identify weakness that may have been presumed normal.