Log in  |  Search

Closing the Harlem-Scarsdale Score Gap

Caroline Hoxby’s updated report on New York City’s charter schools uses a provocative construct: she finds that Harlem’s charter students are making standardized test score gains that put them on track to substantially close their achievement gap with Scarsdale.

Hoxby, a Hoover Institution fellow and Stanford professor who has published extensively on charter schools (favorably) and teacher unions (unfavorably), looked at students who won admittance by lottery to certain New York City charters and compared their performance to students who applied but were not admitted.

As Jonathan Gyurko writes in an earlier post, she found an incremental scale-score improvement of 2.4 to 3.6 points (on a 325-point scale) more per year  in reading and math for charter pupils over those who lost the lottery and did not attend a charter. But she then projects that a Harlem student who attended charters from K-8th grade would make the same gains every year and could  narrow his or her achievement gap with Scarsdale students by 86% in math and 66% in ELA.

What Hoxby did was take this point difference, this “charter effect,” and present it as a persistent, undiminishing causal effect that can work educational miracles over eight years on the same student. And it’s unlikely she has test scores for very many students who’ve been continuously enrolled in a charter for eight years.

But she did have a nice construct, one that would make anyone sit up and take notice. Scarsdale is one of the top performing school districts in New York State, even in the United States. Its campus-like schools boast rich electives, high-tech labs and music rooms, green rolling playing fields, helicopter parents and relaxed, highly-paid teachers. Below, the Scarsdale High School library. Under that, a Harlem Charter School.

Scarsdale High School

Harlem Charter School

The gaps between Harlem and Scarsdale students are about far more than test scores, and closing the academic ones will take a lot more than test prep

SOME HARLEM-SCARSDALE GAPS

INDICATOR HARLEM SCARSDALE
Median household income $23,150 $122,234
SCHOOLS
Free/reduced lunch eligible 78% 0%
Percent black and Hispanic 96% 5%
% Teachers w Masters+30 or PhD 29% 67%
Average class size grade 8 math 27 19
Mean scale score G4 math 661 705
Mean scale score G8 ELA 638 688

To perpetuate the fiction that if they could just attend charter schools, Harlem’s struggling students would morph into Scarsdale over-achievers (if indeed they even wanted to) is a disservice.

Hoxby’s study was hailed by the Wall Street Journal on page 2 and greeted at the final word on charter superiority in a Daily News editorial the next day.

But researchers know better. The black-white test score gap has been shown to persist even between middle-class blacks and whites, for deep and complex reasons. Ronald Ferguson, a Harvard professor who studied the racial achievement gap in Shaker Heights, Ohio  (Hoxby’s home town), says it would take at least 25 years to close the racial achievement gap, even between students of from families with similar incomes.

Ferguson found that blacks scored on average nearly 100 points below whites on SATs. Even in Shaker Heights the average grade for a black senior was C+ versus B+ for whites. The gap is not a result of effort — he found blacks studied harder than whites — but of the persistence of poverty’s ills even after incomes had equalized.

So by all means let’s equalize the resource gaps with Scarsdale. But it’s not right to generalize that a few points average gain on a standardized test means high-needs students will continue to make the same gains year after year. Nor that those decimal points are all it takes to close America’s racial achievement gaps.

Print

1 Comment:

  • 1 Connie Van Brunt
    · Oct 17, 2009 at 11:36 pm

    I am working to launch a peer-reviwed Journal of Charter Schooling because it is important to have valid and reliable studies and reviews of studies, including methods, procedures, and findings. This is one function served well by a peer reviewed journal of the field and not served well by newspaper level discussion.