Purpose: The purpose of this paper is to examine the way in which low-performing schools and their district define, acquire, use, and diffuse research-based evidence. Design/methodology/approach: The mixed methods case study builds upon the prior research on research evidence and social networks, drawing on social network analyses, survey data and interview data to examine how educators in low-performing schools and across the district use evidence (including which types and for what purposes), as well as the relationship between network structure and evidence use for school improvement. Findings: Educators had narrow definitions of, and skepticism about, evidence, which limited its acquisition and use for school improvement. The authors found a lack of diffusion of evidence within schools and districtwide as a result of sparse connections among and between educators. Evidence was used in an instrumental, yet superficial, manner leading to weak interpretation of evidence and resulting in limited understanding of underlying problems and available solutions. Research limitations/implications: The paper suggests the importance of using social network analyses to examine the diffusion of evidence, as well as the need to better understand how evidence is defined and used. Practical implications: It is necessary to pay greater attention to how educators acquire evidence, as well as the ways in which it is used to impact school-based decisions in low-performing schools and districts. Moreover, the work suggests the influence of the district office on school-level reform. Originality/value: The paper contributes to the research on low-performing schools and accountability policy by examining the larger districtwide context and integrates social network, survey, and interview data. (Contains 1 table and 6 figures.)