“Really,” Stanford education sociologist Sean Reardon told Education Week‘s Sarah Sparks last week, “there are very, very few school districts that serve a large proportion of poor students and that have achievement that’s even at the national average.  That suggests we may not be able to just ‘school reform’ our way out of that kind of inequality.”

A revolutionary statement in an era when, for fifteen years now, we have been punishing the school districts that serve our nation’s poorest students.

Ohio, my state, and many others have implemented a policy that grades school districts by their students’ test scores.  The school districts that serve poor children are getting “Fs” and for these districts, other kinds of punishments for school failure soon follow—closing schools, expanding charters, giving kids vouchers to leave so-called failing public schools altogether.

Sparks reports that in the interview, Reardon adds: “It’s not clear we’ve figured it out.  There’s some deep… problems that we as a society haven’t faced up to yet.”  Sparks explains that new research by Reardon found that Detroit, among the poorest cities, is one of very few poor school districts where there is virtually no achievement gap. Reardon explains: “Detroit is not the poster child for reducing the achievement gap.  The achievement gap is zero in Detroit largely because everyone’s doing really poorly, not because black students are doing particularly well.”

If we look at Reardon’s new paper and in the context of the new NAEP scores released last week that show dropping achievement and the achievement gap widening among high school seniors, it is evident that the new research begins to explore the issue that test-and-punish school accountability has failed to address: whether achievement gaps are created by the quality of schools themselves.  Reardon introduces the new paper: “One of the central sets of questions in the sociology of education for the last 50 years—since the publication of the Coleman Report—concerns the primary causes of racial and ethnic achievement gaps and disparities in educational outcomes more generally.  To what extent are these disparities the result of racial/ethnic differences in socioeconomic family background and circumstances, and to what extent are they the result of racial/ethnic differences in school quality?  Put differently, to what extent should racial/ethnic disparities in educational outcomes be attributed to institutional features of the US educational system…?”

In the NY Times, Motoko Rich describes Reardon’s data analysis: “The new analysis surveys data from about 200 million standardized math and reading tests given to third through eighth graders in every state between 2009 and 2012.  Although different states administer different exams, Mr. Reardon and his team were able to compare the state results with scores on federal tests known as the National Assessment of Educational Progress in order to develop a consistent scale by which to compare districts.  Mr. Reardon said the analysis should not be used to rank districts or schools.”

Reardon is also at pains to insist that the new study documents correlations, not cause and effect, and he cautions that the new research raises more questions than it answers: “Our findings should not be taken as causal estimates; as we argue here, the forces producing racial/ethnic inequality in educational outcomes are complex, interactive, and self-reinforcing….”  That being said, Reardon does identify several conclusions:

1. “Most of the variation in district achievement gaps lies within, rather than between, states.  State-level processes do not appear to be a dominant force in shaping patterns of racial/ethnic academic achievement gaps.”
2. “(O)f the several thousand school districts we analyze, which enroll almost 90% of all black and Hispanic students in the US, there are but a handful where the achievement gap is near zero.”
3. “(R)oughly half the variance in local achievement gaps can be explained by racial/ethnic disparities in socioeconomic status.”
4. “There are clearly factors other than racial/ethnic socioeconomic disparities at play in generating academic achievement gaps.  Chief among these factors is racial segregation… In particular, achievement gaps are larger, all else equal, in places where black and Hispanic students attend higher poverty schools than their white peers.”
5. “Achievement gaps are larger, on average, in districts and metropolitan areas with higher levels of parental education…. (O)ne possible explanation for this is the possibility that socioeconomic disparities—and corresponding disparities in social capital, social networks, and access to school district leaders—are more salient in competitive, high resource communities.  Another possibility is that social psychological processes that inhibit minority students’ performance, such as stereotype threat, are particularly strong in the most affluent places….”
6. “(W)hile many of our measures of segregation are correlated with achievement gaps, the one that consistently remains statistically significant in our multivariate models is racial differences in exposure to poverty.  This is in line with the argument that race, per se, is not the causal factor linking segregation to worse outcomes for minority students.”
7. “(A)mong all the covariates included in our models, our measures of school quality explain the smallest amount of the variance in achievement gaps.”

In a companion paper, Reardon warns: “Test scores and academic performance more generally are shaped by many factors other than schools.  They are shaped by children’s families, their home environments, their neighborhood contexts, their child care and pre-school experiences, afterschool experiences, and by their schools.  Knowing that children in a particular community scored higher, on average, than those in another community does not tell us that the schools were better in that community.  Average test scores are more appropriately interpreted as a measure of the educational opportunities available to children living within a district.”

Keeping that caveat in mind, however, in the primary paper, Reardon explains: “(M)any school districts have achievement gaps that are larger or smaller than would be expected, given their socioeconomic conditions and segregation levels.”  One primary example cited in Reardon’s study is Union City, New Jersey, the subject of Berkeley public policy professor David Kirp’s fascinating book, Improbable Scholars. Reardon would not, of course, jump from his correlation study that highlights school achievement that beats the odds among poor children in Union City to offering a prescription for school improvement.  It is, however, interesting to look at David Kirp’s portrayal of the many ways Union City, a high achieving and very poor school district, has worked to beat the odds.

Kirp describes the strategies in Union City that he believes have turned around achievement as measured by test scores.  Union City empowered its teachers to re-shape a bilingual curriculum to serve its primarily Hispanic students. School administrators trusted teams of teachers to collaborate. Kirp summarizes reforms he observed in Union City:  “High-quality, full-day preschool for all children starts at age three. Word-soaked classrooms give youngsters a rich feel for language. Immigrant kids become fluent first in their native language and then in English. The curriculum is challenging, consistent from school to school, and tied together from one grade to the next.  Close-grained analyses of students’ test scores are used to diagnose and address problems.  Teachers and students get hands-on help to improve their performance.  The schools reach out to parents, enlisting them as partners in their children’s education.  The school system sets high expectations for all and maintains a culture of abrazos—caring—which generates trust.”  (Improbable Scholars, p. 9)

The emphasis in Union City has been on support-and-improve, not test-and-punish.