Three broad statements summarizing nine axioms underlie a conceptualization of the character of behavioral and social research: (1) social and behavioral science is inherently reflexive, (2) ethics and epistemology are inseparable, and (3) the inherent complexity of behavioral and social phenomena must be reflected in the models developed to guide the support of productive educational research and development (R&D). The idea of differing epistemologies suggests that ignorance of or failure to recognize the worth of others' work leads to conflict between practitioners and researchers, research administrators and politicians as well as to slow adoption of research. A promising interaction model for educational research focuses on the outcomes of research and those who have a stake in the outcomes. Relationships between "stakeholders" and outcomes vary. The form of an outcome is related to accessibility to stakeholders; stakeholders may have several relations in respect to a research outcome--primary, secondary, or tertiary. Policy implications flowing from the model lie in three main areas: the legitimization of a broader array of R&D activities, functions, and actors; decentralization of authority and responsibility for educational R&D; and the demand for much more careful analysis and attention to policy in its own right. (YLB)