I promise that this whole week won’t be deadly depressing, but Alexander Russo threw down the gauntlet about the media’s lack of attention to results in Chicago Public Schools under Arne Duncan. So I took a look at Chicago’s NAEP performance.
Have gaps separating white/black and white/Hispanic students in Chicago shrunk in the last 5-6 years?
Nah.
There are no statistically significant declines in these gaps in 4th or 8th grade reading or math. In many cases - for example, 4th and 8th grade math and 8th grade reading - it’s not that the black-white achievement gap is declining, but not by enough to be statistically significant. These gaps are actually growing. Sigh.
Lucky for Duncan, Joel Klein doesn’t have a leg up in this area - neither superintendent has been successful in narrowing racial achievement gaps. (See NYC achievement gap analyses in painful detail here.)
And while we’re on the topic of Joel Klein and NAEP scores, a word on his latest statistical gaffe about sampling. From the outback, here’s the exchange between a reporter and Klein at his Australian National Press Club talk:
Reporter: Are you concerned that the only independent national assessment of American children shows that there is actually no improvement in New York, with the slight exception of year four maths? Klein: [The NAEP tests] are sampled, whereas our state tests are mandatory. They're the accountability and, if you look at the data that I provided to you, across the board those state tests show that we're outperforming the rest of the state and everyone else.
Given that a sample is appropriately drawn - and there is no reason to believe that the NYC NAEP sample is not - it should produce unbiased estimates of the achievement of the underlying population, NYC kids. This guy is being discussed as a candidate for Secretary of Education, and in the same breath telling Australians that our national test is inaccurate because it’s based on a sample. Oy.