A robotic exoskeleton was designed for individuals with crouch gait caused by cerebral palsy with the intent to supplement existing muscle function during walking. The aim of this study was to evaluate how powered knee extension assistance provided during stance and swing phases of the gait cycle affect knee kinematics, and knee flexor and extensor muscle activity. Muscle activity and kinematic data were collected from four individuals with crouch gait from cerebral palsy during their normal walking condition and while walking with the exoskeleton under stance, swing, and stance & swing assistance. The exoskeleton was effective in reducing crouch by an average of 13.8° in three of the four participants when assistance was provided during the stance phase; assistance during the swing phase alone was ineffective. Peak knee extensor activity was maintained for all of the conditions during the stance and swing phases. Integrated (i.e. area under the curve) knee extensor activity decreased in two of the subjects indicating a more well-modulated activation pattern. Modest increases in peak and integrated antagonist knee flexor activity were exhibited in all participants; the subject without kinematic improvement had the greatest increase. While the exoskeleton was well tolerated, additional training with a focus on reducing knee flexor activity may lead to further improvements in crouch gait reduction.