Pilot Studies   During the fall of 1998, Museum Learning Collaborative researchers conducted pilot studies in several museums, including the Birmingham Civil Rights Institute, the Pittsburgh Children's Museum, and the Senator John Heinz Pittsburgh Regional History Museum. The purpose of these pilot studies was to begin to address some of the methodological concerns that we need to resolve in preparation for designing our comprehensive study on learning in museums. We were interested in exploring the problems and issues of doing research on particular populations in museum settings, and in evaluating the data-gathering capabilities of our instrumentation and equipment. We also wanted to begin the process of developing and refining coding schemes we will ultimately use for interpreting visitor behaviors and for measuring the effects of a museum visit on peoples' conceptions of particular topics. We interviewed, observed and sometimes intervened with museum visitors, varying the level of researcher intrusion into their visits, as a way of gaining an understanding of the varying kinds of data that can be obtained under different conditions. To give a flavor of how these issues played out in the context of one of these studies, we share our experiences conducting a study entitled, "Looking Through the Glass: Case Studies in Museum Learning as Revealed through Visitor Conversation." For this study, researchers held and then closely examined extensive conversations with a set of carefully selected adult visitors before, during, and after a tour of the "Glass: Shattering Notions" permanent exhibit at the Heinz History Center. The visitors varied in their knowledge and appreciation of glass, and in their experience of local history and of museums. They ranged in age from 20s to 80s, and in education from high school to graduate school. This study had a deliberately intrusive design. The conversations between and among visitors and researchers that occurred during interviews before and after the museum visit included probes for detailed information both about the visitors' personal identity and about the displays encountered in the exhibit. Similarly, conversations during the museum visit were prompted and supported by researcher comments and queries. All interviews and tours were audiotaped. The intensive data collection on each subject or pair of subjects generated about 100 pages of typed transcript for each set of visitors. As a result of soliciting the kind of information that would yield a rich portrait of each visitor, we also have a comprehensive record of how visitors' perceptions and understandings about the glass industry in Pittsburgh emerged in a museum visit. We used the transcripts to practice a variety of possible coding dimensions, trying out different ways of coding in order to weigh the pros and cons of different approaches. Ultimately we need to develop coding that will best characterize the content and structure of visitor conversation, regardless of whether the conversation is between a visitor and a researcher or among members of a group of visitors with a researcher present. The data also allowed us to explore the relationship between aspects of visitor identity and engagement with the exhibit.