This article reports that a quiet debate is unfolding over proposals to tinker with No Child Left Behind Act's (NCLB) definition of what constitutes "scientifically based research" in education while other ideas for revamping it are taking center stage. Sprinkled through the federal education statute more than 100 times, the references to "scientifically based research" require educators to rely on such studies in choosing everything from approaches to reading instruction to anti-drug programs for students. But the legislative definition of "scientifically based research," which favors randomized or experimental studies over other kinds of research in determining what works in schools, has also been criticized for promoting a narrow view of educational scholarship. Leaders of the House Education and Labor Committee, in a draft proposal for reauthorizing the NCLB law, would tone down that emphasis on scientific experiments by stipulating that studies aimed at determining whether an educational program or practice works may include--but are not limited to--random-assignment experiments. However, the proposed change has raised concerns among some federal education research officials and research groups.