Research in robotics is becoming an ever more applied science. Roboticists acknowledge the existence of a role for experiments in their research, but whether the results of such experiments provide useful information to the intended industry or profession remains somewhat ambiguous. In this paper, we particularly consider experiments relating to robotic wheelchairs. There are many prototype robotic wheelchairs, but what level of performance must they achieve before being accepted into mainstream society and how do we verify the reliability of such performance? How can researchers evaluate their systems effectively? We compare and contrast the metrics used by medical practitioners to gauge the mobility status of a patient with those that are popularly used in academia to evaluate robotic wheelchair performance. We conclude that to design and execute successful experiments with robotic wheelchairs, researchers must draw not only on the experience of the intended end users, but also on the expertise of the medical practitioners who assess and support the patients in the day–to–day use of their wheelchairs.