Rehabilitation robots are increasingly being tested and promoted for clinical neurorehabilitation. Compared with conventional and manual methods, robots allow for a variety of advantages, particularly in the areas of interventional control and the ability to provide a high volume of facilitated movement. Since 1997, there have been more than 60 clinical trials reporting the use of two dozen different robots for neurorehabilitation. Although there are a number of smaller pilot studies, there are only few larger clinical trials. There may be a number of reasons why pilot robot studies do not materialize into larger studies. Beyond devices that failed to perform as intended, what are the clinical design issues that have limited these studies? Some basic considerations include randomization, inclusion of a control group, power calculation based on a clinically meaningful outcome, and finally, reproducible descriptions of the intervention being tested. Although many of these issues are general challenges presented for all rehabilitation studies, there are clinical design features that would likely greatly improve interpretation of results and better position robot devices toward the next clinical trial step. On the other hand, the absence of these elements, even in the setting of a pilot study, may significantly hamper the interpretation of results and not yield sufficient information on treatment effects, adverse event rates, dropout rate, and so on, to allow further testing to proceed to follow-up Food and Drug Administration phase II and III studies. Development of rehabilitation robots for clinical use needs to occur hand in hand with well-conducted clinical trials to provide evidence of efficacy while also taking into account costs.