The theory of action upon which high-stakes accountability policies are based calls for systemic reforms in educational systems that will emerge by pairing incentives for improvement with extensive and targeted technical assistance (TA) to build the capacity of low-performing schools and districts. To this end, a little discussed and often overlooked aspect of the No Child Left Behind Act (NCLB) mandated that, in addition to sanctions, states were required to provide TA to build the capacity of struggling schools and Local Education Agencies (LEAs, or districts) to help them improve student achievement. Although every state in the country provides some form of TA to its lowest performing districts, we know little about the content of these programs or about their efficacy in improving student performance. In this paper, we use both quantitative and qualitative analyses to explore the actions taken by TA providers in one state--California--and examine whether the TA and support tied to California's NCLB sanctions succeeds in improving student achievement. Like many other states, California requires that districts labeled as persistently failing under NCLB (in Program Improvement year 3, PI3) work with external experts to help them build the capacity to make reforms that will improve student achievement. California's lowest performing PI3 districts are given substantial amounts of funding and are required to contract with state-approved District Assistance and Intervention Teams (DAITs), whereas the remaining PI3 districts receive less funding and are asked to access less intensive TA from non-DAIT providers. We use a five-year panel difference-in-difference design to estimate the impacts of DAITs on student performance on the math and English language arts (ELA) standardized tests relative to non-DAIT TA during the two years of the program intervention. We find that students in districts with DAITs perform significantly better on math California Standards Tests (CSTs) averaged over both treatment years and in each of the first and second years. We do not find evidence that students in districts with DAITs perform higher on ELA CSTs over the combined two years of treatment, although we find suggestive evidence that ELA performance increases in the second year of treatment relative to students in districts with non-DAIT TA. Ordinary least squares (OLS) regressions that explore the association between specific activities fostered by DAITs and changes in districts' gains in achievement over the two years of treatment show that DAIT districts that report increasing their focus on using data to guide instruction, shifting district culture to generate and maintain high expectations of students and staff, and increasing within-district accountability for student performance, have higher math achievement gains over the course of the DAIT treatment. In addition, DAIT districts that increase their focus on ELA instruction and shift district culture to one of high expectations have higher ELA achievement gains than do DAIT districts that do not have a similar focus.