analysis, these associations should not be interpreted causally. The studies described above generally argue that NCLB might have affected achievement levels by introducing “consequential accountability” systems. Moreover, they argue, the effects of NCLB in a given state might have been moderated by the stringency of the accountability system adopted by that state. Although these are plausible arguments for how NCLB might have affected achievement levels , it is less clear that these features of NCLB would have affected achievement gaps . Rather, we reason that NCLB was most likely to affect achievement gaps through its subgroup- specific reporting and accountability requirements. By drawing attention to racial/ethnic differences in performance, and by holding schools accountable for the performance of each subgroup, NCLB may have led schools and districts to focus explicitly on reducing achievement gaps at the same time as they focused on increasing achievement levels. Absent the subgroup-specific nature of the NCLB accountability regime, we might not expect it to be particularly effective at narrowing achievement gaps. 4 Moreover, the subgroup- 4 That is, unless an accountability regime requires states and schools to focus on relative performance, we reason that it will shift the performance of all groups (to the extent that it is effective) without affecting gaps. It is also possible that an accountability regime would bring up the lower tail of the student achievement distribution and, to the extent that more minority than white students are clustered in that tail, narrow the difference in average test scores between the two groups. However, such an effect might not alter the ordinal ranking of students in the test score distribution, and so would not affect rank-order measures of
9 specific informational and accountability features of NCLB may be more effective in states where most minority students are in schools where their scores are reported and subject to consequences. This implies that overall accountability stringency may be a less important moderator of NCLB’s effect on achievement gaps than is the proportion of minority students in schools meeting the minimum subgroup size criterion. Our analytic strategy below allows us to test this hypothesis. Analytic Strategy and Hypotheses Our analysis proceeds in five parts. First, we describe the data and methods used to measure racial achievement gaps in a comparable way across states, grades, years, and test subjects. Second, we describe average within-state trends in white-black and white-Hispanic achievement gaps. Third, we estimate the average effect of NCLB on within-state achievement gaps, using a dose-response model that relies on the assumption that the effects of the NCLB accountability regime accumulate as students progress through school. Fourth, we test whether and how the effect of NCLB varies among states. Our hypothesis is that if NCLB operates via subgroup-specific informational and accountability mechanisms, it should most reduce gaps in states where a large proportion of black and Hispanic students are subject to these mechanisms.
You've reached the end of your free preview.
Want to read all 63 pages?
- Winter '18
- No child left behind Act, U.S. state, NCLB