In peer tutoring, the learner is taught by a colleaguerather than by a traditional tutor. This strategy has been shownto be effective in human tutoring, where students have higherlearning gains when taught by a peer instead of a traditionaltutor. Similar results have been shown in child-robot interactionsstudies, where a peer robot was more effective than a tutorrobot at teaching children. In this work, we compare skillincrease and perception of a peer robot to a tutor robot whenteaching adults. We designed a system in which a robot providespersonalized help to adults in electronic circuit construction.We compare the number of learned skills and preferences ofa peer robot to a tutor robot. Participants in both conditionsimproved their circuit skills after interacting with the robot.There were no significant differences in number of skills learnedbetween conditions. However, participants with low prior domainknowledge learned significantly more with a peer robot thana tutor robot. Furthermore, the peer robot was perceived asfriendlier, more social, smarter, and more respectful than thetutor robot, regardless of initial skill level.