Editor’s Note: Our Books, Articles, and Items of Academic Interest feature is at present being written by guest authors. The perspectives and opinions presented are those of the authors and will of necessity vary from issue to issue.
Civil Rights Issue?
Racial gaps in academic achievement constitute “the greatest civil rights issue of our time,” according to the subtitle of The Black-White Achievement Gap, co-authored by Rod Paige (with Elaine Witty), former secretary of education, ed school dean, and Houston schools superintendant (American Management Association, 2010). Perhaps that is an exaggeration—or perhaps it’s true, and it demonstrates how much American race relations have improved in the last 150 years. But these gaps certainly exist, and they certainly matter.
One thing the book does exceedingly well is to explain the racial gaps of which it speaks. The authors devote chapter 2 to “the facts of the matter,” tracing black-white gaps from addition-and-subtraction proficiency in spring of first grade (82 percent of white children are proficient, compared with 59 percent of black children) through high-school graduation rates (78 percent for whites, 56 percent for blacks) and average SAT scores (1065 for whites, 856 for blacks).
Standardized tests have long been dogged by accusations that they are culturally biased, including in one recent study of the SAT (“Unfair Treatment,” by Maria Veronica Santelices and Mark Wilson, in the Spring 2010 Harvard Educational Review). Some argue that the gap in test scores does not imply a gap in real skills. But Paige and Witty aren’t buying it: “[T]his is ludicrous. Performance disparities show up on virtually every measure of student achievement. It’s obvious that the problem is real, it’s undeniable, it’s big, it’s not going away, and, in some cases, it’s getting worse.”
The existence of a testing gap raises two questions: How are colleges and employers responding to the gap? And what can public policy do to close it? The last several years have seen a lot of academic work on these topics.
The biggest recent development in college admissions research is the publication of No Longer Separate, Not Yet Equal: Race and Class in Elite College Admission and Campus Life, by Thomas J. Espenshade and Alexandria Walton Radford (Princeton University Press, 2009), reviewed by Russell Nieli in the Fall 2010 AQ. For this study, several anonymous elite universities gave researchers access to their admissions data—including GPAs, test scores, National Merit Scholarships, extracurricular activities, etc. By analyzing which students were accepted and which weren’t, the academics could calculate how much weight the colleges assigned to each factor.
Their results reveal that in response to racial academic gaps, elite universities have instituted severe affirmative action. At the private schools they analyzed, relative to being white, being black is worth 310 SAT points (on a 1600-point scale, with a possible score of 200 to 800 per section), being Hispanic is worth 130 points, and being Asian costs 140 points. At public schools, Espenshade and Radford present their results in ACT points out of 36: being black is worth 3.8, Hispanic 0.3, Asian −3.4.
In addition, the study showed that affirmative action is indeed a response to racial gaps rather than an attempt to help the disadvantaged in general. All else being equal, poor blacks get more help than rich blacks, but for whites, being poor actually hurts one’s chances of admission.
The study’s major limitation is its focus on elite colleges; other studies, such as those by the Center for Equal Opportunity (which has a massive collection of the research it has conducted at its website, www.ceousa.org), reveal that lower-tier colleges don’t use preferences nearly as strong. But No Longer Separate, Not Yet Equal is the most detailed study of racial preferences to date.
Many private employers also test their applicants—and unlike colleges, employers often face lawsuits if their tests reveal gaps. In “Disparate Impact Realism” (William and Mary Law Review, Fall 2011), law professor Amy L. Wax assesses the legal ramifications and psychological validity of standardized employment tests.
“Disparate impact” is a legal guideline that has its roots in the 1971 Supreme Court case Griggs v. Duke Power. It holds that whenever a test screens out a disproportionate number of minorities, courts can presume it is illegally discriminatory; at which point the employer can be dragged into expensive legal proceedings to prove that the test is job-related. The specific cutoff is four-fifths: If a minority group doesn’t pass the test at least 80 percent as often as whites do, the test has “disparate impact.”
A 2009 decision, Ricci v. de Stefano, found that it was illegal for a fire department to throw out the results of a test simply because not enough minorities passed; this practice discriminates against the applicants who did pass. However, the decision left the four-fifths rule in place, putting employers who test their applicants in a litigation catch-22. If a test screens out too many minorities, they can be sued, and if they throw out the test to avoid a lawsuit, they can be sued for that, too.
In her article, Wax draws on research from a variety of fields, including “industrial and organizational psychology,” to demonstrate that the gaps these tests measure are real, not merely artifacts of the tests themselves. In fact, employment tests, and especially tests that measure intelligence, are the best-known predictors of job performance, and they seem to work equally well for all racial groups.
Because significant gaps exist, and because it’s quite difficult to measure ability without revealing them, almost any well-designed test will risk running afoul of the four-fifths rule. Businesses thus face what’s called the “validity-diversity tradeoff”: the better a test is at predicting job performance, the more likely it is to screen out too many minorities and get the employer in legal trouble. Efforts to screen employees in a way that is highly predictive but does not reveal racial gaps—such as testing personality rather the intellectual ability, relying heavily on job interviews, or having applicants carry out tasks in a “job simulation”—have not managed to elude this tradeoff.
Wax offers several suggestions for reform. One is to convert the four-fifths rule to a sliding scale that jibes with statistical reality: the more selective a job is, the less courts should expect employers to create a balanced racial distribution. Another suggestion is to eliminate the four-fifths rule entirely; in Wax’s view, racial gaps are overwhelmingly the result of real differences rather than arbitrary discrimination, so the rule serves little purpose anyhow.
Closing the Gaps
Rather than papering over the gaps with preferences and overzealous laws, can we close them? The elephant in the room is the question of whether the gaps are partly genetic in origin: if they are, closing them entirely is impossible, unless we intentionally stunt the educational achievement of high-scoring groups. (No serious contributor to this debate contends that the gaps are entirely genetic.) This issue has inspired decades of research and a tremendous amount of emotion, but no definitive answer; until genome research advances considerably, it will remain highly difficult to tease apart the effects of genes and environment statistically.
The best summary of the evidence for a genetic component is “Thirty Years of Research on Race Differences in Cognitive Ability,” by J. Philippe Rushton and Arthur R. Jensen (Psychology, Public Policy, and Law, June 2005). In this paper, the (highly controversial) authors survey “10 categories of evidence” on the question, including everything from brain-size research, to statistical analyses of IQ scores and environmental factors, to transracial-adoption studies. One of the better arguments against a genetic component is available in Richard E. Nisbett’s Intelligence and How to Get It: Why Schools and Cultures Count (W.W. Norton & Company, 2010). Nisbett concedes that the gaps measured by IQ tests are real and important, but presents evidence that they are caused exclusively by environment. For example, American blacks with more European ancestry don’t seem to have higher IQs, and blacks who grow up in better family environments do tend to have higher IQs.
In the meantime, it makes sense to consider a more limited question: can we narrow score gaps by improving education? The optimistic answer is provided by teacher-effectiveness research, the state of which is neatly summarized in “What Studies Say about Teacher Effectiveness,” a report that Education Week reporter Stephen Sawchuk prepared for the Education Writers Association in June 2011. These studies look at whether students’ test scores rise or fall after spending a year with a particular teacher. They find consistently that some teachers do a better job than others—which implies that by training or firing low-performing teachers, we can raise students’ test scores.
What remains unclear, however, is how dramatic an improvement we will see, and what effect it will have on racial gaps. Minorities are disproportionately sent to schools with bad teachers, so it seems likely that widespread adoption of teacher-quality measures would help them more than whites. And some studies show that in a single year, having a great teacher instead of a poor one can make a significant difference in test scores. Many observers, including some academics, have assumed that this effect would simply accumulate year over year, so that gaps would progressively narrow with each year of decent schooling. But as Sawchuk notes, and as one would expect, some studies have found that the effects of a good teacher wear off over time.
Another source of optimism is research on charter schools. Some studies show that charter schools increase test scores and graduation rates. By far the sunniest report is How New York City Charter Schools Affect Achievement, by Caroline M. Hoxby, Sonali Murarka, and Jenny Kang (The New York City Charter Schools Evaluation Project, September 2009), which argues that children who win charter school lotteries in New York not only outpace their peers who weren’t selected, but continue to improve with each additional year of charter school education. This offers a way to close the “Scarsdale-Harlem achievement gap,” a marketing term the authors presented and news reporters quickly picked up on. Other scholars have, of course, disputed some of the findings—for example, Matthew Di Carlo, “A ‘Summary Opinion’ of the Hoxby NYC Charter School Study” (National Education Policy Center, July 2011)—but most see the study as proof that at least in some circumstances, charter schools can boost scores.
The pessimistic answer is provided by Charles Murray in Real Education (Crown Forum, 2008), reviewed by Wight Martindale, Jr., in the Winter 2008–09 AQ. Murray outlines the long history of failed educational interventions. Dating all the way back to the “Coleman Report,” Equality of Educational Opportunity (U.S. Department of Education, 1966), the most rigorous studies have found that once demographic factors are accounted for, school quality has very little effect on student outcomes. Big successes have almost always been the result of bad research design (and thus not really successes at all) or an intense, focused effort by highly talented educators (and thus not replicable in the country at large). For example, the Abecedarian Early Intervention Project in North Carolina has long been presented as evidence that early childhood education can dramatically improve academic achievement—but upon closer analysis, some researchers say the project wasn’t as effective as advertised, and efforts to replicate it have fallen short. And for decades, studies have shown that the federal government’s Head Start program boosts scores only slightly, and that even those gains often fade out over time.
Robert Weissberg’s Bad Teachers, Not Bad Schools (Transaction, 2010) devotes an entire chapter to the black-white achievement gap, and provides an even more detailed look at the utter failure of most attempts to close it with education reforms. Weissberg even wades into the uncomfortable truth also explored in Stuart Buck’s Acting White (Yale University Press, 2010) that busing and other integration efforts—making sure that blacks and whites have access to the exact same educational resources—don’t do the trick. To note just one example: in the 1970s, Pasadena, California, succumbed to political pressure and integrated its schools. Before the project, Pasadena whites scored at the fifty-ninth percentile nationally, blacks at the twenty-seventh. After some initial promising results, the score averages reverted to their previous levels—and whites left the school system.
These critics do not say we should give up on finding new and better ways to educate children, of course—but their work will help readers form reasonable expectations. And reasonable expectations are sorely needed whenever testing gaps are at issue.