of intellectual capital to the research that cannot be obtained in isolation of practice. Ideally, relationships generate a bidirectional flow to the work, with the research informing practice while craft knowledge and practical wisdom enrich the research. In some cases, important research cannot be conducted without this collaboration. These partnerships are not always easily formed, and often take long periods of time to establish. But they are often essential to develop the trust that is necessary for researchers to perform their jobs adequately and to engage education professionals in a mutually enriching dialogue about the role of research in practice. A current National Research Council effort is attempting to build the capacity of infrastructure for such long-term partnerships to foster research that is useful to practice (see National Research Council, 1999d), and others have suggested that research serve as a basis for long-term communications between researchers and practitioners (Willinsky, 2001). We argue in Chapter 6 that a federal education research agency should help broker such partnerships as part of its investment in strengthening the education research infrastructure.
Another way that some field-based researchers have recently attempted to bring educational practice closer to the research process is by embedding inquiry in “sites of practice” (National Research Council, 2001a). For example, to better understand the knowledge that teachers need to teach third grade mathematics effectively, researchers have grounded their work in concrete examples from teaching practice (e.g., samples of student work solving mathematical problems). Focusing research on these representations of the process of education in practice can generate important insights about the interactive nature of teaching and learning in classrooms (Ball and Lampert, 1999). Engaging in this kind of research, of course, depends on the willingness of school-based practitioners to participate and the establishment of relationships to facilitate it.
As we argue in Chapter 2, with some exceptions, U.S. society has not developed an appetite for using education research as a tool for improving teaching, learning, and schooling (Campbell, 1969). This posture exacerbates the difficulties establishing the relationships necessary to conduct research. The problems with conducting randomized trials attest to this fact: there is little expectation that educational programs or interventions should be subjected to rigorous research (Cook 2001; Burtless, in press). In