Under the No Child Left Behind Act, states have been required to set minimum proficiency standards that virtually all students must meet by 2014. Using longitudinal, student-level test score data from seven states between 2002-03 and 2005-06 this paper addresses the following research questions: (1) Has NCLB increased achievement among lower-performing students? ; (2) Have these gains come at the expense of students who are already proficient or far below the proficiency target? Identification is achieved by exploiting the fact that in the early years of NCLB, not all grades counted for purposes of determining AYP. Thus the estimate of the NCLB effect is based on a comparison of outcomes in high-stakes vs. low-stakes years. The authors find consistent evidence of an achievement trade-off in the hypothesized direction, though the effects on any given student are not large. They find mixed evidence that students far below the proficient level have been harmed by NCLB; in fact at higher grade levels they appear to have benefitted. Effects of NCLB on efficiency, while positive, appear to be modest.