President Bush said the following about education in the State of the Union address:
Spreading opportunity and hope in America also requires public schools that give children the knowledge and character they need in life. Five years ago, we rose above partisan differences to pass the No Child Left Behind Act, preserving local control, raising standards, and holding those schools accountable for results. And because we acted, students are performing better in reading and math, and minority students are closing the achievement gap.
Now the task is to build on the success, without watering down standards, without taking control from local communities, and without backsliding and calling it reform. We can lift student achievement even higher by giving local leaders flexibility to turn around failing schools, and by giving families with children stuck in failing schools the right to choose someplace better. We must increase funds for students who struggle — and make sure these children get the special help they need. And we can make sure our children are prepared for the jobs of the future and our country is more competitive by strengthening math and science skills. The No Child Left Behind Act has worked for America’s children — and I ask Congress to reauthorize this good law.
Having recently hopped into the skeptic column on NCLB, I decided to do a quick check on the President’s empirical assertion that students are performing better in reading and math, and that minority students are closing the achievement gap.
Since the year 2000, Math NAEP scores are up a bit. Reading scores are flat and mixed, up a bit in 4th grade, down a bit in 8th grade. Nothing much to get excited about.
On the achievement gap front, things look better. I tend to focus on 4th grade NAEP reading results above all else. Kids who don’t learn how to read in the early grades are put a severe disadvantage, and the pool of 4th graders who can’t read overlaps with the pool of subsequent 9th graders who begin dropping out in droves 5 years later. Needless to say, students have a hard time learning any sort of academic subject if they can’t read their textbooks.
Nationally, 4th grade reading scores do show some closing of the achievement gap between Anglos and African Americans, and between Anglos and Hispanics. Furthermore, they show this narrowing in the best possible way: the gap is closing while Anglo scores are themselves improving.
All aboard the NCLB bandwagon then! Well, not quite. Even the current trend of progress were to be sustained indefinitely (extremely unlikely) the Black-White achievement gap in 4th grade reading would close somewhere around the year 2035.
Better gradual progress than no progress, some might say. Some states, however, have been at this standards-based accountability reform agenda longer than others, however, and the news there is not very encouraging on the achievement gap. Texas and North Carolina, for example, were early leaders in testing, and neither of them has seen even a sliver of gap closing between Anglos and African Americans on 4th grade reading.
More worrying still, however, is the fact that the seeds of NCLB’s destruction as a credible reform strategy were sown at the outset of the bill. NCLB puts the federal government in the position of eventually requiring all students to pass tests (thus the grandiose name) but leaves test content and passing requirements up to the states. The result: a race to the bottom whereby states have already cut their passing levels to absurdly low levels.
In 2004, the Atlanta Journal Constitution reported that the questions on the state’s third-grade reading test were essentially “speed bumps on the road to fourth grade.” The Georgia Department of Education released data showing that 16 of the 40 questions on the third-grade reading test were “easy,” with 75 percent or more of the students getting them right. Not coincidentally, students only had to get only 17 of the 40 questions right to pass the exam and advance to the fourth grade.
Georgia students are not alone in needing to answer less than a majority of answers correctly in order to “pass” a state accountability exam. In Texas, for example, students needed to correctly answer only 29 of 60 questions in order to pass the mathematics section of the accountability exam, prompting one Texas columnist to express surprise that the education bureaucracy did not simply require a single correct answer, noting that “then they could have had 100 percent passing rates.”
Mind you, this is happening with much lower NCLB passing requirements than we will be seeing in a few short years. In short, it is going to get worse. The race isn’t over yet. Absent a change of course, states will be demonstrating “100% proficiency” on tests that you are declared proficient on for signing your name.
Ah well, the jaded cynic might sigh, yet another education fad headed for the ash-heap of failed education reforms. Frightfully, this situation is far more serious than that. The standards movement is on a collision course with an even more important education trend: transparency for educational results.
Transparency is the baby in the NCLB bathwater. In the not so distant past, there was next to nothing available about student performance in public schools. Today, there is quite a bit of information available online. Websites like greatschools.net have succeeded in taking turgid state reports and turning them into user-friendly repositories of data.
This system is far from perfect today, but light years ahead of where we were 15 years ago. The former head of the Texas House Education committee, for instance, has related to me the story of his contacting his local district and asking for the district’s scores on the state accountability exam in the early 1990s. The district’s response was “I’m sorry, we don’t release that information.”
We’ve come a long way, baby.
The biggest problem with NCLB is that it is on track to compromise this transparency. As the cut score bar continues to be lowered to avoid federal sanctions, the value of the data becomes increasingly worthless.
What is, for instance, a parent, a voter, or a policymaker to make of pass rates on the Georgia exam referenced above? You only need 17 correct answers to pass, and 16 of them are easy. What value will the data have when you only need 13 correct answers, or 10, or 5? How much pressure will there be on schools to improve? The correct answers are (in order) “not much,” “nothing” and “zero” respectively.
The cost of NCLB, therefore, will be staggeringly high. A large percent of parents send their students to schools other than their assigned neighborhood school- through charter schools, magnet schools, open enrollment, etc. Parents won’t be able to make a rational choice without reliable academic data.
Likewise, state lawmakers cannot really aspire to make rational changes to education policy without reliable outcome data. True, they flew blind for decades, but look at the results. We don’t want to go back there. The same is true for the voters who elect school boards, the school boards who make policies, and the administrators that school boards hire. Without meaningful output data, we will be reduced to making policy by anecdote.
The limited promise of standards-based reform simply isn’t worth throwing the transparency baby out with the bathwater.
What to do then? Remove the threat of federal sanctions by allowing states to develop their own accountability regimes. It is abundantly clear that the federal government isn’t going to bully states into truly improving their scores. States had already started standards-based reform before NCLB. This could be achieved through so called “charter state” provision, being pushed by the Heritage Foundation.
Can we expect states to rectify their already compromised tests once federal sanctions have been removed? Ummmm, no not so much. Some sort of (sanction free) national norm-reference testing should be required of schools, or absent that, more information provided on state tests. For example, the meaningfulness of the Georgia test above would increase greatly if the public new the percentage of correct answers by school and grade level, instead of just the percent passing.
Finally, we need to be focusing on more substantial reforms than NCLB. Parental choice needs to be expanded, and the entire system of public school human resource development need to be revamped, for starters. Getting reliable data is just the beginning of the fun. We’ve got to keep the data coming and make good use of it.
Trial and error in education will be much worse if we find ourselves unable to know when we’ve made an error.
Dr. Matthew Ladner, a former director of state projects for the Alliance for School Choice, is vice president of research for the Goldwater Institute.