While increasing teachers' scientific knowledge base has been identified as a challenge for teacher education (e.g., NCES, 1996), the skills used to identify a need for knowledge and the skills necessary to search for that knowledge have been less discussed. Yet, the ability to learn for oneself is really the goal of lifelong teacher education. In this paper, four class periods of video data from a problem-based educational psychology curse were examined and an eight-minute segment was chosen to analyze in depth. The preservice science teachers grapple with science concepts of electricity and atomic structure as they analyze a video case of a physics classroom and devise ways to redesign instruction in order to enhance students' learning. Three analysis were undertaken: 1) categorizing the scientific discourse to determine how student teachers identify a need for knowledge and how they solve it; 2) analyzing the resources used to build understanding, including how student teachers present themselves as resources, which outside resources are used and why, and the status of those resources in the community; and 3) investigating learning-what did student teachers seem to understand based on the current data sources and how does this compare to an expert view of electricity? The results presented here indicate our current understanding of many rich sources of information, rather than a final analysis. Our discourse analysis of the video segment showed the student teachers identified a need for knowledge in the form of a direct question, or two types of inquiring statements. Most often, they attempted to answer the science question among themselves. The science mentor, World Wide Web, and facilitator were also helpful in building new knowledge. Student teachers made few reflective remarks during problem solving. They engaged in extended scientific reasoning during the video segment, constructed concepts related to charge imbalances, and generated difficult questions, according to science experts. Changes in the role of the participants, facilitator, and other resources are recommended. Coded transcript, transcript conventions, excerpt from final paper, and page of interactions with ment0r are appended. (Contains 44 references.) (Author/DDR)