ideas may be found to exert influence. As noted earlier, developing such a chain of evidence and inference will usually require multiple, coordinated studies.

An important caution: The Framework provides only a conceptual scheme for considering research claims, not all the information needed to assess research-based claims fully. A full analysis must include a host of “technical” considerations, such as standards of evidence, quality of measurement, and appropriateness of the research design. All of these concerns must be addressed in deciding whether particular research conclusions are trustworthy and rigorous.

Assuming comparable technical quality, several hypothetical examples illustrate how the Framework can help determine the soundness of research-based inferences.

• A study of standards-based classroom practice in mathematics. Imagine an investigation of standards-based mathematics teaching practice in a high socioeconomic environment that supports this kind of instruction, using a curriculum that embodies the principles of the NCTM standards. Assume that teachers have been well trained in this form of teaching and are committed to it. If, after sufficient time passes for the curriculum to have affected student learning across grades, a well-designed study documents indifferent or poor student results on assessments keyed to NCTM standards, it would be reasonable to infer that national mathematics standards contributed little to student learning—or might even have detracted from it. That inference could be further substantiated if other school settings less committed to NCTM standards produced more favorable results with comparable students.

• A study of district investment in teacher professional development aligned with nationally developed content standards for technological literacy. In a group of districts heavily emphasizing the principles and themes of ITEA content standards in their professional development and support programs, assessments of teacher knowledge