The Supreme Court’s four-to-three decision in Fisher v. University of Texas to uphold the school’s affirmative action policies is not likely to help African Americans much. That’s because the fight over whether racial preferences in university admissions are constitutional is more about how to treat all students fairly than how to help the millions most hurt by bad schools, bad luck, and prejudice. The Constitution is silent on this graver question.
If the noblest goal of affirmative action is to reduce the disparity of fortune between the races, that effort has clearly fallen short. The gap in college completion rates between whites and blacks aged 25 to 29 grew nearly six-fold between 1930 and 2013, data from the federal Digest of Education Statistics show. Indeed it has widened even during the affirmative-action era. From 1996 to 2007, the percentage of white students completing degrees within six years of graduating high school grew 4.8 percent (to 43 percent); for blacks the rate grew just 1.9 percent (to 20.8 percent).
This growing disparity correlates with growing differences in academic achievement. The percentage of whites aged 17 who scored 300 or higher on the National Assessment of Educational Progress grew from 45 percent in 1988 to 47 percent in 2012. It declined during that period for blacks, from 25 to 22 percent. Students scoring 300 or higher on NAEP are able to find, understand, and explain complex informational material—skills essential to succeed in college.
The black-white achievement and graduation gaps relate in turn to differences in poverty rates. One in four blacks lives in poverty, versus one in ten whites, according to the U.S. Census.
Why Aren’t More Black Kids Learning?
If the surest road to prosperity leads through college, and if decades of affirmative action policies have failed to make the traffic on that road any less white, do we not owe slavery’s heirs something more—something beyond the wishful thinking of universities and the fatalism of late Justice Antonin Scalia’s truism that black students should go where they are best qualified to go? Do we not owe them a searching analysis of the roadblocks to academic achievement and how to eliminate them?
Consider three landmark studies. The first, the billion-dollar federal Follow Through project from the 1970s, tracked more than 75,000 students in kindergarten through third grade in what remains the largest, most careful experiment ever to compare the effectiveness of different approaches to teach poor children. Twenty-two model designers participated. All 22 could point to places that showed promising results with their models. Twenty-one failed to replicate their success at scale. All twenty-one suffered the same problem—the inability to provide effective teacher training. Models teachers had created fared worst.
The second study, published in 1988 by psychologist Galen Alessi, reviewed the cases of 5,000 students who were evaluated by school psychologists to determine why they were doing poorly in class. All 5,000 evaluations attributed the student’s problems to deficiencies in the child or child’s family. Not one linked the student’s problems to faulty teaching or curricula.
Follow Through showed two things: most teachers don’t know how to design instruction that works for poor kids, and most curriculum writers don’t know how to design programs that work for teachers. Alessi showed how educators respond to these failures: they blame the victim. Policymakers unaware of these problems are unlikely to figure out how to fix them.
The third study helps explain the findings of the first two. In their 1995 book, “Meaningful Differences in the Everyday Experience of Young American Children,” researchers Betty Hart and Todd Risley quantified a language deficit in children from welfare families so vast that it’s hard to conceive how teachers might overcome it. By age three, the authors found, children of parents who were professionals had heard, on average, more than 8 million more words than children from welfare families. The kids themselves had spoken more than 4 million more words than the welfare children.
The oral vocabularies of the professional-family kids exceeded those not just of the children, but of the parents of the welfare families. This language gap has grim consequences: follow-up studies showed it correlates closely with large deficits in vocabulary and reading ability at age nine—which, in turn, correlate with large deficits in the reading ability and prosperity of adults.
Still, Schools Can Help
Blame the home then, and pity the school? Not so fast. Educators could narrow this gap, but they have widened it by clinging to a set of romantic fallacies about how children learn. Anyone serious about making affirmative action work cannot escape reckoning with these fallacies and how they impede progress, especially for the poor.
Central to the prevailing view in education is that children learn naturally, and that they learn most when they are allowed to direct the pace and content of their own learning. The ideal teacher is thus not a teacher at all but “a guide on the side” who facilitates the child’s personal growth and knowledge creation. From these premises flow a host of others. Teachers are taught that the mind matures in distinct developmental phases that they cannot do much to accelerate: thus if a child fails, it’s not because the teaching is faulty, it’s because the child is either developmentally disabled or not yet developmentally ready to learn.
Pre-K and kindergarten teachers are warned that it is developmentally inappropriate to seat children at desks, make them work to learn the alphabet, letter sounds, and math, or assess their academic skills. Reading teachers are told children will become better readers if they can choose what they read. Math teachers are taught that kids will be better at math if they are made to figure out their own strategies to solve problems, rather than learn standard procedures from the teacher. All teachers are told that different children learn the same thing in different ways, requiring different teaching methods for each, and that it is wrong to group students by skill level to instruct them in skills (because children learn best from each other).
These ideas may seem reasonable—as geocentrism did before Copernicus—but they have not held up when tested in the classroom. They endure because the public they harm does not know enough to contradict them, and because educators—unlike, say, surgeons and cell-phone-makers—do not face enough pressure to compel them to replace old ways with new ones that work better. The worst consequence, as Follow Through showed, is that educators chronically underestimate what it takes to teach the poor, and overestimate the ability of teachers to design effective instruction for them.
We’ve Found One Thing that Works
The one model that did succeed in Follow Through and that continues to succeed today—Direct Instruction—avoids both traps and explodes the constellation of myths that led to them.
DI is the brainchild of Siegfried Engelmann, an advertising man turned educator who got into the field teaching his twin four-year-old sons. Key to Engelmann’s approach is the scrupulous use of the scientific method to design programs that work for all students, even those who are hard to teach, and that most teachers can be trained to use. Engelmann has written more than 100 such programs, covering all the major subjects from preschool to high school. His disciples and emulators have written many more.
Thousands of field-tested details go into making a DI program. The most conspicuous are concise teacher scripts and choral student responses. The scripts specify the precise sequence of examples, exercises, and wording that teach the subject quickly and clearly. The whole-class unison responses maximize every student’s opportunity to practice and tell the teacher right away if everyone is keeping pace or if some need help. (A teacher trained in DI methods will hear when students answer late or incorrectly—just as an orchestra conductor hears a violinist who comes in late or off-key.)
Lessons are scripted to include as many as 15 teacher commands and choral student responses per minute—many times the teaching rate of most programs. Students are grouped by ability. Procedures to correct mistakes and praise good behavior are clearly specified. Mastery tests every five to ten lessons make sure no child’s struggles go unaddressed.
The system works. More evidence validates DI’s effectiveness than any other approach to instruction. More than 200 studies and meta-analyses since Follow Through have confirmed its findings, and found that DI programs accelerate learning in older students, children with above-average IQs, different racial and ethnic groups, students with disabilities and in special education, and students in urban, rural, and suburban schools. The research has also found that DI lowers rates of grade-retention, discipline problems and referrals to special ed—all benefits that save money. No scientific study has ever found negative effects from DI. Such consistency of results across populations and settings is rare in the social sciences.
But DI is also the ugly duckling of education. No other educational reform strays further from accepted theory, differs more from accepted practice, or draws such brutal slander from those who would hide and deny its achievements. Few colleges teach it, many states won’t fund it, and barely 2 percent of teachers use it.
DI is unpopular for three reasons: it puts more responsibility on educators to achieve results, it gives teachers less freedom to do what they want in the classroom, and it defies a vast system’s vast stake in the conventional wisdom. Few educators have been willing to abandon their beliefs—or the profits from textbooks, training, and curricula that express them.
This Is What Real Affirmative Action Looks Like
The only place that does use DI at scale, IDEA Public Schools, a charter network serving low-income families in Texas, has sent an impressive 50 percent of its students through college since its founding in 2000. That’s four times the normal college graduation rate for the poor—a sign of how great an opportunity educators who put desires above science have missed.
IDEA is the first stop for anyone who wants to see what it takes to do meaningful affirmative action. It enrolls 24,000 students and expects to have 100,000 by 2022. Four of five are poor. Three of four are the first in their families to go to college. Every high school ranks in the top 1 percent of the nation’s most challenging high schools, as ranked by the Washington Post and U.S. News.
Why not attack the language deficit at its roots, and prioritize helping the poor at home? Because no parenting project of any size has ever been shown to increase college graduation rates significantly. The models in Follow Through that focused on family intervention were failures. Besides, bemoaning a poor child’s home life and demanding it be fixed is what educators do to pass the buck.
The evidence is clear: well-designed instruction, delivered by well-managed schools, is the only proven way to narrow the achievement gap. Adversaries in the fight over university admission policies should fight harder to fix K-12 education instead, so that the policies they’re fighting over become unnecessary.
The goal is not to put DI in every classroom, but to put science—DI’s stern mistress—into a profession that ignores it. Pry educators from their idols, and turn teaching into a profession that embraces evidence as its guide. Help a dozen DIs to bloom. That would be affirmative action—for the poor, and any student who struggles to learn.