Full Of Feller

Not everyone's happy about what he's been writing, but there's no doubt it's been a busy past week or so for AP's education writer Ben Feller and the education team at AP. First, Feller helped write the big story about the AYP loophole and the Federal response (Congress Leaders to Probe No Child Scoring, Spellings, too). We all know what a doozy that one's been. Then yesterday he filed a story (AP Poll: Teachers Dubious of 'No Child') about how parents are actually more confident about NCLB than teachers. Early Friday morning, he filed this one (For contractors, education law means money), which chronicles the rise in opportunities for vendors that NCLB has brought. (Before anyone goes crazy about the $22 billion number, remember that the entire K12 education venture in the US is estimated at $400-500 billion a year.)


Blogger Caroline said...

The AP reporters don't get Statistics 1A (this is probably a given, since journalists are overwhelmingly, famously innumerate). Here's a commentary on the issue adapted from the San Francisco Schools blog, www.sfschools.org :

By KC Jones
San Francisco

A recent Associated Press story breathlessly alleged that "States Omit Minorities' School Scores" (Chronicle, 4/18/2006). This story has circulated far and wide. Google News finds more than 500 publications running it.

Only one problem: The story gets it entirely wrong. The accusers simply don't know what they're talking about. It's time to set the record straight.

The "loophole" over which alarms are being raised is actually a
necessary limit on a dangerous and perverse part of the No Child Left Behind (NCLB) law. The kids who are supposedly not counted are, in fact, counted appropriately. All students' scores count towards their school's test score, so no child is left uncounted.

But here's what's happening: Under NCLB, if any identified ethnic subgroup fails to progress, the entire school is considered to be failing to progress. That fits with the goal of leaving no child behind, right? But when there are too few students of a particular ethnic group to form a statistically relevant cohort, NCLB doesn't require tracking that group's progress.

The real outrage here, which the article completely fails to address, is that the "all groups must progress" rule penalizes schools for being ethnically diverse. Consider two schools with the same number of students. One is segregated, with one or two identified ethnic groups in the school. The other is diverse, with five or more ethnic cohorts - like many schools here in the Bay Area. The diverse school will always be at greater risk of penalty under the rule, because the
groups are smaller and thus more statistically volatile. This is
simple statistics, and remains equally true no matter which
ethnicities are under the microscope.

Back to the loophole. NCLB recognizes that tracking very small cohorts would lead to penalizing schools for meaningless statistical noise -- variations that have no bearing on the school's performance or the students' progress. If you track the progress of, say, four American Indian kids in a school of 800 students, of course there will be far greater volatility for the American Indian kids' scores. It makes no sense to track their performance as an ethnic group, so NCLB doesn't require it.

In fact, dividing some diverse schools into small cohorts while not dividing more-segregated schools makes no sense either. It is truly perverse and wrong-headed. But at least the "loophole" mitigates this regulatory disaster. Far from causing outrage, this bit of sanity is a relief from an insidious NCLB regulation.

The other manufactured outrage is the allegation that this loophole is used to exclude minority kids. This charge confirms the reporters' complete uncomprehension. Of course minorities will tend to be "undercounted" in this scheme. Minorities, especially those not living in segregated neighborhoods, will tend to be dispersed in statistically insignificant numbers - because they are in the minority! Duh!

Majority kids - white kids or, here in San Francisco,
Asian kids - will be far less likely to divide into numbers small
enough to be statistically insignificant. Again, duh! This is no
sinister plot to hide the lack of progress of minority kids. It's
merely statistics in action. Clearly, this simple numeracy is too much to ask of journalists.

There are many reasons to question NCLB. There are many reasons to question the motives of some educators, who will tend to be racist in the same proportion as the general population. But this story gets it all wrong.

Tagline: KC Jones is a parent in San Francisco schools, a software engineer, and founder of the 760-member San Francisco schools e-discussion group sfschools@yahoogroups.com. He blogs at

1:15 AM  

Post a Comment

<< Home