Background/Context: In recent decades, federal policymakers have pushed for education to be a more "scientific" endeavor. While scholars have considered the implications of this orientation for educational researchers, less attention has been given to its impact on educational practitioners. Purpose/Focus of the Study: By focusing on the local interpretation and implementation of a multitiered system of support (MTSS) model in one Midwestern school district, this study documents the translation of a comprehensive reform initiative meant to make educational practice more data driven and scientific. With particular attention to interactions between district and building administrators, classroom teachers, and a group of outside consultants, we also consider the consequential effects of principal-agent relations in determining how learners (should) learn and teachers (should) teach. Research Design: Using ethnographic methods over a period of five months, this study emerged from a larger project examining the work of educators in a rural district that includes 18 schools and serves approximately 7,600 students from racially, culturally, and linguistically diverse backgrounds. With MTSS as the unifying agenda across multiple interactions that involved a cross-section of the district's staff, administrative leaders, and outside consultants, we analyzed fieldnotes generated from participant observation during MTSS-specific meetings and semi-structured, individual interviews conducted with key implementation principals and agents. Other fieldnotes and interviews provided confirmation of our primary analysis, as well as supplementary perspectives from building and classroom contexts. Findings: Through our analysis, we found that implementation leaders presumed the infallibility of the MTSS model; relied exclusively on certain forms of quantitative data; standardized the individual needs of learners, processes of learning, and roles of teachers; and insisted on fidelity of intervention as an end in itself. Conclusions/Recommendations: Implementation leaders invoking research to inform practice can sometimes silence practitioners rather than foster their substantive involvement and understanding. This marginalizes certain types of knowledge that can contribute to understanding students' needs, and it forces practitioners to be data-deferent rather than data-driven. The concept of implementation fidelity also needs to be reconsidered--not as an absolute good but with the necessary flexibility afforded to practitioners who are (1) educated in the essential components of available interventions, (2) able to become fluent through practice, and (3) allowed to exercise their professional expertise and judgment as appropriate.