The Lincoln Plawg - the blog with footnotes

Politics and law from a British perspective (hence Politics LAW BloG): ''People who like this sort of thing...'' as the Great Man said

This page is powered by Blogger. Isn't yours?
Wednesday, July 07, 2004

The bizarre and troubling world of high school graduation statistics

The recent further comments of Bill Cosby on the fecklessness of (some) blacks yield an op-ed piece which mentioned a study on the pitifully low rates of high school graduation for blacks and Hispanics I vaguely remembered from earlier in the year.

Getting hold of the study (PDF), Who Graduates? Who Doesn't? by Christopher Swanson (from the Urban Institute), which looks at the Class of 2001, I find it's full of interesting stuff. Or, at least, stuff that this layman finds in need of explaining.

The headline numbers show there is no room for inter-racial Schadenfreude - whilst graduation rates (GRs) (p28a) for blacks (50%) and Hispanics (53%) are bad, that for whites (75%) is nothing to shout about.

The study does regression analysis to tease out (p39ff) the correlation of various factors with variations in GRs: no surprise that poverty (as measured [1] by the proportion of recipients of free and reduced-price lunches (FRL)) was the largest (p40).

But it is a surprise - to me at least - that the effect of poverty varies by ethnic group (p43); graphing the correlation found, as the FRL rate moves from 0% to 100%, the GR for whites declines from around 85% to 75%; the GR for Hispanics from 58% to 53%; but the GR for blacks goes from 68% to 50%.

Unfortunately, the study doesn't seek to explain this discrepancy.

Some of the most interesting material comes in the state-by-state and district-by-district analysis (p50ff).

For example, California (p55): Los Angeles (FRL 74%) graduates 68% of its whites, whilst San Bernardino (FRL 75%) only manages 45%.

Sacramento graduates 64% of blacks, but only 59% of whites. And Oakland (FRL 54%) manages only 24% of blacks and 57% of whites.

And Florida (p60), with a rate of white graduation of just 58% - Asians have a 22 point advantage in the Sunshine State. But there are no crossovers as in Sacramento, which is no doubt a consolation. Even in Pinellas County, with a 36% FRL, whites only manage a 50% rate.

As an illustration of the poverty effect on black graduation, in Gwinnet County, Georgia (p61), with 21% FRL, the black and white GRs are almost tied (71% and 73%) - in Atlanta (FRL 76%), they're 39% and 65%. (But it goes right against the general finding on Hispanics: Gwinnet graduates 55% of Hispanics but Atlanta only 25%.)

I've no way of explaining these numbers: perhaps there's an interesting story behind each of those numbers, perhaps they're merely artefacts, perhaps a mixture.

There is another study, from 2003, from the Manhattan Institute for Policy Research, Public High School Graduation and College Readiness Rates in the United States which overlaps with the UI study and more or less seems to track its findings.

(There is, it seems, no universally recognised statistic for high school GRs. The UI study looks at the construction of its stat on p12ff.)

I'm a little puzzled why the study did not test correlation with the per capita amount spent on schools in the state or district...

You may be wondering whether the UI study make much of splash, media-wise, when it came out in February? Can't tell without actual Nexis: Google for "urban institute" "high school" graduation suggests it got some coverage.

And, I see, the presumptive John Kerry includes a reference at the end of his paper John Kerry's Plan for One Million More Americans to Graduate High School [2].

  1. Per capita personal incomes by state stats here.

  2. Kerry's paper flags the manipulation of statistics: in particular, the incentive of the No Child Left Behind Act for districts to employ push-out - to disappear the worst performing students from schools (or from the statistics, at least) to bump up test averages. Clearly, NCLB won't have affected the UI study - but state education initiatives, like that in Texas, might well have done.

    Kerry's 100 Days to Change America education page looks as if it's been hacked (check the links below the main text...).


A Ron Brownstein piece (May 5), saved for posterity, looks at Kerry's policy, under the hed Kerry Zeroes In on High School Dropout Levels.

I've no idea how likely Kerry's scheme is to be effective. But it's interesting that the paper that accuses Bush of cooking the books with education stats itself has a deeply misleading headline. Get into the text, and its clear that the million more Americans to graduate is not, as would be usual, an annual statistic, but a five years' total.

A particular confusion arises in that, as Brownstein says, the graduation
percentages translate into about a million students who drop out each year, the Kerry campaign calculated.

Brownstein continues:
Kerry said his new plan would reduce the number of dropouts each year by about 200,000, raising the overall graduation rate to about 80%.

Compared with the Kerry paper heading, it sounds a good deal less impressive.

free website counter Weblog Commenting and Trackback by