Individuals with disabilities arising from neurological injury require rehabilitation of the distal joints of the upper extremities to regain the ability to independently perform activities of daily living (ADL). Robotic rehabilitation has been shown to effectively conduct high intensity, long duration therapy and quantitatively assess the effects of therapy. This thesis presents methods and results for validating rehabilitation devices for training and assessment. Traditionally, methods for validating rehabilitation robots relied on robotic characterizations, which enables comparison of different designs' performance independent of a human user. An example of this method is presented here, in particular quantifying the torque output, range of motion, closed loop position performance, and high spatial resolution of two rehabilitation devices. However, these traditional validation methods do not assess the effect wearing the robot has on the user, and a new assessment method has been developed to address this shortcoming of traditional methods. A novel hand and wrist device was assessed through kinematic analysis of synergistic movements, as quantified by velocity- and position-dependent metrics. This experimental approach is promising for the characterization of multi-articular wearable robots as measurement tools in robotic rehabilitation. Together, the two methods presented can be used to validate rehabilitation robotic devices.