We systematically reviewed and analyzed the literature to find randomized controlled trials (RCTs) that employed robotic devices in upper-limb rehabilitation of people with stroke. Out of 574 studies, 12 matching the selection criteria were found. The Fugl-Meyer, Functional Independence Measure, Motor Power Scale, and Motor Status Scale outcome measures from the selected RCTs were pooled together, and the corresponding effect sizes were estimated. We found that when the duration/intensity of conventional therapy (CT) is matched with that of the robot-assisted therapy (RT), no difference exists between the intensive CT and RT groups in terms of motor recovery, activities of daily living, strength, and motor control. However, depending on the stage of recovery, extra sessions of RT in addition to regular CT are more beneficial than regular CT alone in motor recovery of the hemiparetic shoulder and elbow of patients with stroke; gains are similar to those that have been observed in intensive CT.