Using longitudinal data from the No Child Left Behind (NCLB) era, I applied regression techniques and found a positive association between school failure to reach "adequate yearly progress" in mathematics and subsequent changes in the quality of middle grades mathematics instruction in districts where district leaders adopted robust theories of action for improving mathematics instruction. The positive association was robust to multiple sensitivity tests and may reflect a causal relationship. The evidence suggests that educational leaders in similar contexts can use school failure to reach accountability standards as measured by standardized assessments to promote instructional quality in mathematics.