Closing the Achievement Gap: Have We Flat-Lined?
The recent controversy over No Child Left Behind waivers has drawn attention to the intransigence of group differences in performance on NCLB-style standardized tests. As unrealistic as the NCLB goal of 100 percent proficiency was for advantaged students, it was trebly more so for poor and minority students.
And yet the design of NCLB emphasized the illumination of subgroup performance in the definition of “Adequate Yearly Progress” toward the goal, in the hopes of limiting the ability of schools and districts to hide patterns of low performance by vulnerable subgroups amidst the successes of more advantaged students.
Florida, Virginia and Washington, D.C. are among the jurisdictions that have sought waivers to allow different proficiency-rate targets for different racial/ethnic groups. Critics worry that targets differing by race will legitimate a kind of institutional racism, in which schools and districts might let up on efforts to close racial and ethnic achievement gaps. But if the timetable for eliminating gaps is unrealistic, race-based targets may be warranted: challenging but attainable goals foster more motivation to succeed than unrealistic ones.
Sean Reardon of Stanford University and Andrew Ho of Harvard University are preparing an analysis of the effects of NCLB on the magnitude of racial/ethnic gaps in student achievement; keep an eye out for their work when it is circulated. In the meantime, here’s a very simple demonstration of the challenge we face in closing the achievement gap that separates white and Asian students from Black and Latino students, using an approach developed by Reardon and Ho.
New York City has seen some of the more far-reaching educational reforms over the past decade, as Mayor Michael Bloomberg and Chancellor Joel Klein set in motion an array of market-based reforms. Both Bloomberg and Klein argued vigorously that the New York City schools had substantially closed the achievement gap, pointing to a shrinking difference in the percentage of white students and Black and Latino students classified as proficient on the New York State English Language Arts and mathematics assessments administered in grades 3 to 8.
Many scholars have demonstrated, however, that differences in proficiency rates are potentially misleading, and especially so if the tests have inflated scores reflecting predictable and easier questions. Has the achievement gap in New York City decreased over time? What happened to the achievement gap when the state of New York, recognizing the flaws in its testing system, raised the “cut scores” defining proficiency on its tests in 2010?
To gauge trends in New York City’s achievement gaps, I present a hypothetical scenario. Suppose that we randomly select a white or Asian student in one of the grades that is tested and look at that student’s test score. Then, we randomly select a Black or Latino student and look at his or her test score. What’s the probability that the score of the white or Asian student is higher than that of the Black or Latino student? If that probability decreases over time, that’s evidence that the achievement gap separating white and Asian students from Black and Latino students is shrinking. On the other hand, if that probability remains constant over time, that’s evidence the achievement gap remains unchanged.
The figure below shows these probabilities from 2006 to 2012, separately for English Language Arts (the blue line) and mathematics (the red line). The solid line at 50 percent indicates parity; at that level, a randomly chosen white or Asian student would have a 50 percent chance of a higher test score than a randomly chosen Black or Latino student—a toss-up.
In fact, though, the probabilities favoring white and Asian students are much higher than 50 percent, indicating the relative advantage of white and Asian students over Black and Latino students. In 2006, the probability that a white or Asian student would have a higher score than a Black or Latino student was 72 percent for English Language Arts, and 74 percent for mathematics.
Both the blue and the red lines are relatively flat, indicating little change over time in the magnitude of the test score gap separating white and Asian students from Black and Latino students in New York City. In 2012, white and Asian students had a 71 percent probability of a higher English Language Arts score than Black and Latino students, and a 75 percent probability of a higher math score.
My conclusion? There’s been no shrinkage in the test score gap between 2006 and 2012, a period in which many of Bloomberg and Klein’s reforms have begun to reach maturity. If the only purpose of their reforms were to close the achievement gap, this flat-lining would indicate that the reforms were dead on arrival.
That’s probably too harsh a verdict for a complex package of reforms, some of which may prove beneficial in the long run. And the point here is not a referendum on what’s happened in New York City as much as it is a demonstration that racial/ethnic group differences in test performance are stubborn, even in the face of efforts intended to minimize them.
We are about to enter an era with a new set of Common Core curricular standards and new assessments designed to measure students’ mastery of those standards. The combination of a more challenging set of standards, a lag in the development of curriculum and the professional development that teachers need to teach to those standards, and assessments that are widely proclaimed to be more difficult than existing NCLB-style tests will likely result in plummeting rates of student proficiency in English and mathematics in the near future. Significant closure of the achievement gap may be beyond the grasp of educators who will be struggling simply to keep their heads above water in the next five years.
This blog post has been shared by permission from the author.
Readers wishing to comment on the content are encouraged to do so via the link to the original post.
Find the original post here:
The views expressed by the blogger are not necessarily those of NEPC.