The intention of this research was to describe behaviors and characteristics of General Chemistry I peer leaders using a pedagogical reform method referred to as Peer-led Guided Inquiry (PLGI), and to discuss the ways in which these peer leaders created productive whole-class discussions. This reform technique engaged students to work on guided inquiry activities while working cooperatively in small groups, led by undergraduate peer leaders. These sessions were video recorded and transcribed. The data was evaluated using grounded theory methods of analysis. This study examined the dialog between students and peer leaders, paying specific attention to question types and observed patterns of interactions. The research took shape by examining the kinds of questions asked by peer leaders and the purposes these questions served. In addition to looking at questions, different kinds of behaviors displayed by peer leaders during their small group sessions were also observed. A close examination of peer leader questions and behaviors aided in developing an answer to the overall research question regarding what factors are associated with productive whole-class discussions. Five major categories of peer leader behaviors evolved from the data and provided a means to compare and contrast productive whole-class discussions. While no category single-handedly determined if a discussion was good or bad, there was a tendency for peer leaders who exhibited positive traits in at least three of the following categories to have consistently better whole-class discussions: Procedural Practices, Supervisory Qualities, Questioning Techniques, Feedback/Responses, and Interpersonal Skills. Furthermore, each of the major categories is tied directly to Interpersonal, Communication, and Leadership skills and their interactions with each other. This study also addressed applications that each of these categories has on instructional practices and their need in peer leader training. In addition, a scale was developed for rating the relative effectiveness of whole-class discussions in terms of student participation. This study provides a tool for measuring productive whole-class discussions, as well as practical applications for peer leader (or teacher) training.