Bang per Buck SGP Analysis

An earlier analysis showed that Division SOL scores are essentially uncorrelated with division expenditure per student and that Richmond spends much more per student and obtains much lower SOLs that most other divisions.

The recent release of the Student Growth Percentile data opens a new path for analysis.  Unlike the SOL, the SGP is essentially uncorrelated with economic status.

We have the 2014 SGP data but VDOE is yet to post the expenditure data for the 2014 Superintendent’s Annual Report.  So we turn to the 2013 data (here with the facility costs, debt service, and contingency reserves removed from the calculation of total cost per student).

First, the division average reading SGP scores vs. the division average expenditure per student.

image

The fitted curve might suggest that the reading score increases with increasing expenditure but the R2 of just under 1% tells us that the two variables are essentially uncorrelated. 

Richmond is the gold square, spending above average sums per student and obtaining, as with the SOL, below average scores.  The red diamond to the left is Hampton, which Brian Davison already has identified as a division with a large population of economically disadvantaged students but a robust SGP, suggesting superior teaching there.

On the subject of robust, the two divisions with SGP averages >55 are Poquoson on the left and Charles City on the right.

The other red diamonds are, from the left, Newport News and Norfolk.  All three of the red diamond jurisdictions demonstrate that it is possible for an old, urban jurisdiction in Virginia to deliver average or superior education at near-average costs.

The math data show a similar picture as to Richmond albeit a less flattering one for Hampton:

image

Again, the SGP results are barely correlated with expenditures (R2 = 2.7%).

The outperforming divisions, all with SGP averages >60, are, from the left, Bland, Botetourt, Buckingham, and Bristol.

I’ll venture a conclusion: Virginia schools, especially the expensive ones such as Richmond, don’t need more money; they need more of whatever Hampton and Poquoson and Charles City and Bland and Botetourt and Buckingham and Bristol are doing.