OBJECTIVE. MRI and sonography are both used to evaluate patients with painful shoulders. This study was conducted to compare patients' perceptions and satisfaction with both tests. SUBJECTS AND METHODS. One hundred eighteen patients with shoulder pain and a clinically suspected rotator cuff tear underwent both MRI and sonography and filled out satisfaction surveys after both tests. Patients were asked the following questions: Did the test cause pain? If it did, they were asked to grade the pain on a scale of 1-10 (1, minimal pain; 10, severe pain). Did the test take too long? Would they be willing to undergo the test again? How would they grade their overall satisfaction with the test (1, poor; 2, fair; 3, good; 4, very good; and 5, excellent)? Which test did they prefer if both were equally accurate? RESULTS. Sonography caused pain above the baseline in 39 patients, with an average pain score (mean ± SD) of 5.4 ± 2.3, and MRI caused pain above the baseline in 40 patients, with an average pain score of 6.1 ± 2.7 (p = 0.36). Two patients thought the sonography examination took too long, and 28 patients though the MRI examination was too long (p < 0.001). The average satisfaction level for sonography was 4.3 ± 0.7 and for MRI, 3.6 ± 1.2 (p < 0.001). The satisfaction score was higher for sonography in 54 patients, higher for MRI in 13 patients, and the same for both in 50 patients (p < 0.001). All patients were willing to repeat the sonography, but 10 patients were unwilling to repeat the MRI (p = 0.002). Ninety-three patients preferred sonography, eight patients preferred MRI, and 17 patients had no preference (p < 0.001). CONCLUSION. Most patients with shoulder pain prefer sonography to MRI.