Formalised peer-to-peer collaboration and the use of web-enhanced materials that are consistent with course objectives, graded assessments and learning outcomes is well known in educational practice. This study compared the academic gains of college students enrolled in Physics using Reciprocal Peer Tutoring (RPT) with others using non-Reciprocal Peer Tutoring (non-RPT). Structured procedures were developed, supported by on-line mentoring activities, and used by students arranged into pairs. Prior to an RPT activity, each student prepared practice exercises in multiple choice tests complete with answers and explanations based on the scope of the topic as indicated in study guides and modules provided to them through an established website. Each student then administered the practice test to the other, so switching roles happened between tutee and tutor within the same RPT session and prior to formal class examination. In the non-RPT group, only the students appointed as tutors were assigned to prepare exercises. One student was always the tutor and the other was always the tutee. In both groups, a class monitor and the peer mentor for each group were appointed to give feedback about peer tutoring activities. Both types of peer tutoring produced significantly larger academic gains than traditional classroom instruction and RPT resulted in marginally larger academic gains than non-RPT.