The American Recovery and Reinvestment Act of 2009 (ARRA) targeted substantial School Improvement Grants (SIGs) to the nation's "persistently lowest achieving" public schools (i.e., up to $2 million per school annually over 3 years) but required schools accepting these awards to implement a federally prescribed school-reform model. Schools that met the "lowest-achieving" and "lack of progress" thresholds within their state had prioritized eligibility for these SIG-funded interventions. Using data from California, this study leverages these two discontinuous eligibility rules to identify the effects of SIG-funded whole-school reforms. The results based on these "fuzzy" regression-discontinuity designs indicate that there were significant improvements in the test-based performance of schools on the "lowest-achieving" margin but not among schools on the "lack of progress" margin. Complementary panel-based estimates suggest that these improvements were largely concentrated among schools adopting the federal "turnaround" model, which compels more dramatic staff turnover. (Contains 12 figures, 12 tables, and 19 footnotes.)