The No Child Left Behind Act was a well-intentioned law, but like federal education law generally, the reality of what it has likely accomplished has not lived up to its promise.
I’d like first to look at the evidence of the law’s academic achievement effects, especially on underserved populations. I do this with a few important provisos. First, standardized test scores generally only provide limited information about how children are performing, and NCLB focuses on reading and mathematics. Not only does that mean NCLB ignores art, social studies, physical education, and other academic subject areas, it also ignores character development, preparation to become active citizens, and other, broader educational goals. Second, test scores often tell us how well students were prepared to take certain kinds of tests, which does not always translate into useable skills or other desired educational outcomes. Finally, sundry variables influence academic outcomes — students’ health, home lives, motivation levels, district policies, state policies, etc. — and controlling for all of them in order to isolate the effect of a federal law is extremely difficult, if not impossible.
With those caveats in mind, what does the National Assessment of Educational Progress — the federal “Nation’s Report Card” — suggest about NCLB?1
The good news is that below high school level there were gains during the time of NCLB for underserved student populations on both the “main” NAEP, which has been administered since the early 1990s, and the Long-Term Trend NAEP administered since the 1970s.
On the main NAEP, for free and reduced-price lunch eligible 4th graders, math scores rose from 222 to 230 (out of 500) between 2003 and 2013, and for 8th graders, from 259 to 270. If you go back to 2000, the gains were greater, but it is very hard to know the extent to which NCLB could have driven this since the law was not passed until early 2002, and it took a long time to implement.
In reading less progress was made. Between 2002 and 2013 4th grade, free and reduced-price lunch eligible reading scores rose only from 203 to 207, and for 8th grade from 249 to 254.
The long-term trend results are similar. For African Americans we saw a big jump in 9‑year-olds’ math scores between 1999 and 2012, rising from 211 to 226. For Hispanics, the jump was from 213 to 234. But much of that happened between 1999 and 2004, and how much occurred before or after NCLB is impossible to tell. For 13 year-olds, African-American scores rose from 251 to 264, and Hispanics’ scores rose from 259 to 271, again with the 1999–2004 pinpointing problem in mind.
The long-term trend reading results show African-American 9‑year-olds’ scores growing from 186 to 206, and Hispanics from 193 to 208. For 13 year-olds, African American scores rose from 238 to 247, and Hispanic scores from 244 to 249. All of this, again, with the 1999 to 2004 problem.
These are generally good gains, but there are big problems with attributing them to NCLB. The first, of course, is pinpointing when to start the NCLB clock. The second is, again, that numerous variables affect test scores, and while controlling for income or race helps a bit, it does not tell us, for instance, how motivation to do well in mathematics and reading may have changed over time regardless of NCLB, or whether states set their tests to be more like NAEP without actually increasing useful learning. They also don’t tell us whether schools increasingly taught testing strategies rather than more effectively teaching course content.
Efforts to pinpoint the effect of NAEP empirically have suggested smaller effects, though they have difficulty controlling for hard-to-capture variables like motivation and culture. For instance, Dee and Jacob found that comparing main NAEP scores of states with and without accountability systems before NCLB suggests that NCLB improved math scores for the youngest and most disadvantaged students, but had no effect in reading. The study, however, had no control for potentially changing motivation or attitudes toward learning.2 They also found evidence that NCLB had encouraged schools to devote more time to “narrow test preparation activities.” Looking at Wisconsin, Chakrabarti examined the effect of adequate yearly progress failure on subsequent high- and low-stakes test scores, and found a positive effect for reading but no statistically significant effect for mathematics.3 In addition, Chakrabarti found no evidence of improvement for weaker student groups — which NCLB was specifically supposed to help — and perhaps some “deterioration in performance… especially in the Economically Disadvantaged group.” This may be the result of triaging students to focus on those on the margins of passing. Where Chakrabarti did find a powerful effect was in schools that faced a serious threat of students being able to move to other, well-performing schools, but those were rare overall.
It is worth noting that, looking at long-term trend results, we see several periods before NCLB that had larger gains for underserved populations. For instance, math scores for African-American 9‑year-olds rose 10 points between 1986 and 1994, or 1.25 points per year, but only 5 points between 2004 and 2012, or .63 points per year. Similarly, reading scores for Hispanics rose 9 points between 1994 and 1999, or 1.8 points per year, but only 8 points between 2004 and 2012, or 1 point per-year. Again, it is very hard to determine how much of these gains are attributable to federal policy, but they show that it is difficult to conclude that No Child Left Behind has been as a major difference maker.
Perhaps the most negative evidence we have for NCLB is test scores for, roughly, high school seniors, the school system’s “final products.” (The trends report uses age, and the main NAEP uses grade, so the scores are for 17-year-olds and 12th graders, respectively.) Here we see, first, overall stagnation since the 1970s, an indicator that, despite roughly a doubling of both real federal spending and overall per-pupil outlays, we have not made much progress, at least as measured by federal tests.4
What do the scores show for NCLB, especially on the groups most likely to struggle? On the long-term trends test, they show little, if any, improvement. Between 1999 and 2012, scores for African Americans rose from 283 to only 288 in math, and scores for Hispanics increased just one point. In reading, African-American scores rose only 5 points and Hispanic scores went up only 3 points.
The main NAEP does not provide scores before 2005 for mathematics, but in reading reveals only a 2 point increase between 2002 and 2013 in scores for children from homes in which students reported neither parent having completed high school, and a 2 point drop for students in homes where neither parent had completed education beyond high school.
Given all this, at best one can say that No Child Left Behind may have had some positive effect on underserved 4th and 8th graders, but no discernable effect by the time students neared the end of elementary and secondary education. That means we have no evidence of any lasting effect — by far the most important outcome — and some evidence of short-term effects for students when in grades four and eight. And none of this can be conclusively pinned to NCLB because numerous variables affect outcomes.
Given the evidence of overall ineffectiveness, there is a good argument for eliminating No Child Left Behind completely. Indeed, the Constitution — which grants the federal government no authority to govern education outside of Washington, DC, federal installations and territories, and prohibiting state and local discrimination in provision of education — seems to require the end of laws such as NCLB.
The just re-introduced Student Success Act in the House, and the Every Child Ready for College and Career Act in the Senate, would stop well short of removing the federal government from elementary and secondary education, but would reduce the federal pressure in NCLB. The Student Success Act would still require states to set uniform standards (except for students with very significant disabilities), have assessments in grades three through eight and once in high schools, and have accountability systems that identify school performance and the performance of subgroups of students. It would end the use of Adequate Yearly Progress and a requirement of full proficiency by the end of a given school year. The Senate bill is very similar, except it has two potential testing options, one of which does not include annual testing.
These bills would be moves in the right direction, reducing federal micromanagement of education. But there is little evidence that any federal requirements should remain. The fact is standards-based reforms started in states in the 1980s, and a major effect of No Child Left Behind’s effort to impose such structures on all states may have actually been to put a damper on potential progress by threatening control of federal funding based on “proficiency” rates. That’s quite possibly why assessments of state proficiency definitions revealed not so much a “race to the bottom” as a very broad trend of either setting low standards or maintaining standards that were low to begin with.5 States did not push themselves because of the negative ramifications of setting high bars.
What about shining sunlight on achievement by subgroups? Many people credit NCLB with creating unprecedented transparency, and it may well have increased the general availability of data on test scores and other measures. But the nation has been discussing achievement gaps since at least the late 1950s, and NCLB has had the unintended consequence of openly labelling many children as “proficient” who were not close to proficient according to NAEP proficiency levels. And nationally normed tests that schools have been using for decades have shown individual students where they have stood compared to national averages.
When discussing federal policy, it is valuable to remember the idea that states are “laboratories of democracy,” as Justice Brandeis put it, which a crucial benefit of a federal system. That means one state can try a policy innovation, and if it works other states are free to adopt it. If it doesn’t, all states do not suffer the consequences of failure. And states, while all suffering from numerous problems that are inherent to policymaking at any governmental levels, have an incentive to find better ways of educating children because they compete with one another to attract new businesses, residents, etc. In other words, if standards-based reform really does work, states will have incentives on their own to use it, though they may modify it so as not only to use it, but to improve it. When it is imposed from above they have an incentive to game the reform, as we’ve seen with proficiency definitions, and the ability to compete by changing or improving a policy is quashed.
Of course, states are also welcome to try school choice on their own, which I believe — and evidence suggests6 — is the key to sustained innovation in education and empowerment of all families. But the federal government should not attempt to foster choice, including helping charter schools, as the Senate and House bills do. This too should be a decision made at the state level, and the laboratories should be allowed to work.
Unfortunately, efforts to reduce some of the prescriptiveness of NCLB by the Obama Administration, using waivers of dubious legality, has in some ways exacerbated the problem of uniformity impeding competition and innovation, especially by putting pressure on states to adopt nationally uniform standards and tests. This began with the Race to the Top, part of the American Recovery and Reinvestment Act, that de facto required states to adopt the Common Core curriculum standards and participate in one of two testing consortia to compete for funds. The Student Success Act and Every Child Ready for College and Career Act would move in the right direction by prohibiting such federal pressure, as well as similar pressure to move states toward more uniform evaluations of teachers. Again, a major benefit of a federal system is that states are free to try new and different things, and to adapt policies to their own needs. The Senate and House bills, while by no means going far enough, would move in the right direction.
Based on the federal government’s own tests, there is little evidence that the No Child Left Behind Act has spurred significant, lasting improvements in academic outcomes. It has, however, likely put a damper on the innovation that can come through federalism, while spurring a significant revolt against federally driven, high-stakes standards and testing. Informed by these results, and understanding that the Constitution gives the federal government no authority to control the nation’s education system, Washington should work to largely remove itself from elementary and secondary education.
Notes:
1 All National Assessment of Educational Progress Scores come from the NAEP website: http://nces.ed.gov/nationsreportcard/ Various website tools were used to pinpoint scores.
2 Thomas S. Dee and Brian A. Jacobs, “The Impact of No Child Left Behind on Students, Teachers, and Schools,” Brookings Papers on Economic Activity, (Washington, DC: Brookings Institution Press, 2010), pp. 149–194.
3 Rajashri Chakrabarti, “Incentives and Responses under No Child Left Behind: Credible Threats and the Role of Competition,” Federal Reserve Bank of New York Staff Reports, no. 525, http://www.newyorkfed.org/research/staff_reports/sr525.pdf, November 2011.
4 Federal expenditure (not including major, temporary increases of ARRA funds) from National Center for Education Statistics, Digest of Education Statistics 2013, Table 401.10, http://nces.ed.gov/programs/digest/d13/tables/dt13_401.10.asp?current=y…, and total per-pupil expenditure from Table 236.55, http://nces.ed.gov/programs/digest/d13/tables/dt13_236.55.asp?current=y….
5 National Center for Education Statistics, “Key findings from the 2009 State Mapping Analysis,” http://nces.ed.gov/nationsreportcard/studies/statemapping/findings.aspx, August 4, 2011.6 See, for instance, Greg Forster, “A Win-Win Solution: The Empirical Evidence on School Choice,” The Friedman Foundation for Educational Choice, http://www.edchoice.org/CMSModules/EdChoice/FileLibrary/994/A‑Win-Win‑S…, April 2013.
