The passage of No Child Left Behind in 2001 (NCLB) ushered in a new era of standards-based reform strategies. A centerpiece of this reform strategy is its use of “high stakes testing” practices— the use of student test scores to hold students and educators accountable for performance. Critics identify three adverse consequences on students that may result from various teacher and school responses such as “teaching to the test” and targeting students close to test score cut-offs at the expense of lower performing and other academically vulnerable students. These include: (1) increased exposure to poor quality academic environments, (2) decreased attention to the most vulnerable students, and (3) decreased student engagement in school. Little empirical research has directly assessed these claims, especially from the perspective of students. This paper uses survey and test score data generated from the Chicago Public Schools between 1994 and 2001 to investigate student response to high stakes testing practices. In 1996, the Chicago system, which serves a population of mostly minority (African-American and Latino) and low-income students, coupled a school-level accountability program with high-stakes consequences for students. This effort required third, sixth, and eighth graders to meet minimum test score cutoffs in reading and mathematics on the Iowa Tests of Basic Skills in order to be promoted to the next grade. Low performing schools received significant intervention and pressure to improve. The Consortium on Chicago School Research has been conducting biannual surveys of students in the system since the early 1990s. Students report on the nature and quality of their academic environments and supports. This paper compares the survey responses from four cohorts (1994, 1997, 1999, 2001) of sixth and eighth graders. The 1994 cohort serves as the pre-policy comparison group. Data are analyzed using a three-level hierarchical linear model (measurement model nested in students nested in schools) to examine trends in student reports of the academic rigor and supports (including programmatic, teacher, and parent) they received as well as their overall engagement in school, controlling for their own and their schools' demographic and performance characteristics. The analysis includes a total of approximately 60,000 students in 200 elementary schools. The results find that the lowest performing students in the system reported the most positive change in academic supports over time and that this was most apparent in the lowest performing schools in the system. However, these supports were not accompanied by positive instructional change. In other words, while these students perceived more relational supports from teachers and parents, they did not perceive that their teachers held higher expectations for their academic work. Interestingly, higher performing students reported increased levels of boredom and disengagement over time. Two sets of implications will be discussed. First, these data represent the complex consequences of large-scale policy efforts such as these. Betts & Costrell (2002) argue that too few efforts plan for these multiple, and often contradictory, consequences. Second, it discusses the heterogeneity of student response to the policy and the implications for subsequent academic and supportive programming.