The population of manual wheelchair users is increasing notably faster than the general population. Manual wheelchair users have a disproportionately high prevalence of shoulder pain and injury, most likely due to the repetitive and load bearing nature of propulsion. Previous wheelchair propulsion studies have reported on the kinematics of propulsion using traditional time-series data (e.g., propulsion cycle time, shoulder ROM, et cetera) in an attempt to predict those at risk for developing shoulder pathology. However, a reliable method of prediction has yet to be attained using these metrics. Due to the repetitive nature of wheelchair propulsion cycle-to-cycle variability (i.e., intercycle variability) of the traditional kinematic indices may provide additional insight into the movement patterns of the shoulder during wheelchair propulsion. The purposes of these studies are to quantify baseline kinematic variability of the shoulder during manual wheelchair propulsion, quantify the effects of multiple speeds on shoulder kinematic variability, and quantify the effects of an exercise intervention on shoulder kinematic variability.