Carver By the Numbers

The estimable Jim Bacon noticed that the disabled students at Carver outscored their abled peers on the reading tests in 2017.  He said:

While Carver students as a whole out-performed their peers in Richmond schools and state schools, those classified as disabled out-performed their peers by mind-blowing margins. Either Carver has cracked the code on teaching disabled students or… it has been aggressively manipulating test results.

A deeper dive into the numbers suggests that Bacon is being too kind by suggesting that “crack[ing] the code” might be an alternative.

First a quick bit of history:

And a short timeline:

image

Turning to the data, here are the 3d grade reading pass rates by year for Carver, Richmond, and the state.  “Y” indicates disabled; “N” indicates not disabled.

image

The state data, blue, show the statewide drop with the new tests in 2013.  Throughout, the disabled students, the red points, generally underperformed their more abled peers, the yellow points, by about thirty points on the new tests.

The yellow lines are Richmond.  The disabled performance before 2013 reflects their cheating on the VGLA tests.  Since then, Richmond’s disabled students have generally scored below the state average by up to ten points while their more abled peers have been low by about fifteen. 

(In light of Richmond’s overall lousy performance, we can wonder whether those disabled numbers are artificially boosted.  But that is a question for another day.)

Then we have Carver, the green lines.  There is some year-to-year variation, as can be expected from a smaller population.  The non-disabled scores were low before the new principal and have been stratospheric since. 

The scores of the disabled Carver students have been spectacular.  Before 2013, it looks like Carver was abusing the VGLA, along with too many other Richmond schools.  The disabled scores plummeted in ‘13, with the abolition of the VGLA.

After the arrival of the new principal in 2012, Carver’s disabled students often outscored the non-disabled state averages and, in 2015, outscored even the non-disabled Carver students.

Friar Occam would tell us to select the simple explanation: Those Carver scores were “aggressively manipulated.”  That is, they were cheating at Carver, wholesale, both before and after 2012.

Next, fourth grade reading:

image

The missing Carver disabled datum for 2013 probably represents a population small enough to trigger the VDOE suppression rules.  Otherwise, these data tell the same story as the third grade numbers (with a notable higher score by the disabled population in 2017).

Fifth grade reading:

image

These data are a variation on the same theme, but with the stratospheric Carver scores persisting into 2017.

Turning to the math tests:

image

image

image

Note the missing Carver data for 2012.

There are some interesting details here, notably the lower 2017 pass rates in some cases.  Jim Bacon posits “that something changed in the way the SOL tests were administered to make manipulation more difficult.”

The Big Picture is clear, however, imho: The people running Carver have been cheating, prodigiously.  Whatever their technique (I’m hearing tales they posted the answers on the blackboard), they have obtained spectacular pass rates for the non-disabled students and have achieved even better than equal opportunity score boosts among the disabled population. 

I don’t think we need to wait for the retesting for confirmation.  I think that jail would be too good for the staff at Carver.  Stocks for the Carver staff and a bushel of tomatoes for each Carver parent would be a good start.