To explore the problem, we undertook a collaborative, qualitative research and development effort under the general umbrella of the constructivist paradigm. Constructivism as a research paradigm in the social sciences has the following positions: (1) realities are apprehensible as the multiple, sometimes conflicting, products of human intellects, which may change as their constructors become more informed or sophisticated; (2) knowledge is created in interaction among investigator and respondents; and (3) individual constructions can be elicited and refined through interaction between and among the investigator and respondents (Guba & Lincoln, 1994).

We used facilitated teacher study groups as a mechanism to engage teachers in an examination of their choices about curriculum, instruction, and assessment. From the beginning, the project was a collaborative effort between the project staff, educational consultants in the field, and teachers at the research sites. The notion of instructional coherence has not been well developed in the research literature. We had some ideas about coherence, but were convinced that we had to learn from and with the teachers in order to understand the issues, develop ways of thinking about coherence, and create approaches to support teacher learning that would be useful in schools. Our work was influenced by the research team of Cochran-Smith and Lytle (Cochran-Smith, 1994; Cochran-Smith & Lytle, 1990, 1992, 1999; and Lytle & Cochran-Smith, 1992, 1994), Cole (1989), Cole and Knowles (1993), and Day (1991), among others.

Five diverse school and district sites (rural, suburban, and urban) were selected. A study group of 12-18 individuals (primarily teachers) participated at each site. The teachers were "ordinary" or average classroom teachers (Elbaz, 1990), rather than teachers who had been identified as master or exemplary teachers or who were selected based on specific criteria such as writing ability. Teacher-participants volunteered, and they received a small stipend to compensate for meeting after school or on weekends. Each group met regularly to talk about teaching practice. These meetings were facilitated by a project researcher or her partner (a consultant from a local university or service center who was interested in our approach to teaching improvement, had experience in professional development, and had some relationship with the site or some of the teacher-participants at the site).

We used records of conversations as data to understand the experiences of teachers and the meanings they attached to these experiences (Carson, 1986; Clandinin & Connelly, 1994; van Manen, 1990). The groups met for 16 hours or more during each of the two phases of the study. Field notes from the meetings provided one source of data. Each teacher reacted to the study group conversations in a personal journal, and the journals provided a second source of data. We had conversational interviews with most participants, which were taped and transcribed and formed the third source of data. Additionally data sources included field notes from classroom visitations and informal conversations with teachers.

During the study, we met with the consultants as a group several times to discuss what was happening at the five sites, discuss problems encountered at any of the sites, refine strategies for working with the study groups, review data, and develop deeper understanding of coherence. The conversations moved all of us - researchers and consultants - to a clearer and more unified vision of coherence in practice as we shared our different perspectives on the issue. This group collaboration was a critical aspect of the study, as were individual reflective conversations with our consultant partners and the debriefing sessions with study group teachers.

The study was conducted in two distinct phases. The exploratory phase, or first phase, of the study (winter 97 to spring 98) helped us learn about the challenges and issues facing teachers in bringing together curriculum, instruction, assessment, external mandates, and community context in a meaningful way to support improved student learning. Each group started with a consideration of student learning. Participants reflected on issues raised in the video, The Private Universe, as they began to explore how children learn. Through the open-ended conversations about learning and teaching, issues of improving their practices emerged. We collected data on what they talked about and how they talked about it. The findings were used to create a vision of coherence and develop tools and strategies that could support teachers in improving teaching practices.

The second phase of the study (fall 98) was an investigation of the usefulness of the tools and strategies developed to address a focal question at each site. The study group sessions were somewhat structured as participants engaged in guided conversations, activities, and classroom actions. However, each group followed a unique path. The final product of the project is a collection of tools, strategies, and resources that were found useful in supporting teachers' work toward constructing more coherent practices.

The purpose of the study was to identify issues of coherence and create and evaluate tools to enable conversations among practitioners about these issues and related ideas and practices. It was expected that engaging in reflective and thoughtful conversations would lead teachers to improve practice as they developed deeper understanding of coherent practice; however, the study did not attempt to analyze the progress or amount of improvement made by individual teachers. We do have anecdotal evidence that teachers who participated in the study did become more thoughtful about their practice and most participants reported making specific changes in their classrooms to better focus on learning and the learner. This paper reports on what we learned from the study.