Background and Purpose-Locomotor training (LT) using a treadmill can improve walking ability over conventional rehabilitation in individuals with hemiparesis, although the personnel requirements often necessary to provide LT may limit its application. Robotic devices that provide consistent symmetrical assistance have been developed to facilitate LT, although their effectiveness in improving locomotor ability has not been well established. Methods-Forty-eight ambulatory chronic stroke survivors stratified by severity of locomotor deficits completed a randomized controlled study on the effects of robotic-versus therapist-assisted LT. Both groups received 12 LT sessions for 30 minutes at similar speeds, with guided symmetrical locomotor assistance using a robotic orthosis versus manual facilitation from a single therapist using an assist-as-needed paradigm. Outcome measures included gait speed and symmetry, and clinical measures of activity and participation. Results-Greater improvements in speed and single limb stance time on the impaired leg were observed in subjects who received therapist-assisted LT, with larger speed improvements in those with less severe gait deficits. Perceived rating of the effects of physical limitations on quality of life improved only in subjects with severe gait deficits who received therapist-assisted LT. Conclusions-Therapist-assisted LT facilitates greater improvements in walking ability in ambulatory stroke survivors as compared to a similar dosage of robotic-assisted LT.