## July 31, 2008

### Today's Chart

Update: more visual aids here.

Arnold Kling suggests another way to present education data to determine if funding matters in education:

On the X-axis, plot the percentage of students in a county who are above the FARMS line (that is the "free and reduced meals" indicator of poverty). On the Y-axis, plot the percentage of students that pass the math exam. For each county in Maryland, put a data point on the chart. Next to each data point, put the County's ranking in terms of expenditure per pupil.

Next, draw the line of best fit through the data points. Counties that fall above the line are adding relatively more value than counties below the line. If education spending matters, then Montgomery County and other high-spending schools should be above the line. It would be interesting to see whether this is in fact the case.

Here is the chart for 497 school districts in Pennsylvania for 2005. On the X axis I have the percentage of students in the district that do not receive free and reduced meals. On the Y-axis I have the pass rate for PA's PSSA 11th grade exam (math and reading). There were too many school districts to add the funding data but I did do the calculations.

For school districts falling below the regression line the average total expenditures was \$11,417.

For school districts falling above the regression line the average total expenditures was \$11,214.

The overperforming schools actually spent less on average. Go figure.

Back in March I ran few different regressions on expenditures and FARM perfromance, household incomes, teacher salaries and parental education. The results are not always what you'd expect.

Update: Brett from DeHavilland blog has the numbers from Tennessee. Brett writes, "After looking at the correlation between TCAP and poverty rates, we looked at correlation between free/reduced lunch rates and value-added performance of the schools. Virtually no correlation to be found: in other words, some schools with 100% free/reduced lunch rates are contributing tremendously to student learning, and some with almost no free/reduced lunch participants are dropping the ball." Notice the variance (R2) is virtually the same as what I calculated for PA. (Note: Brett 's graph shows the percentage of FARM students not non-FARM students.)

Update II: Unbroken window runs the data for New York and finds the same relationship. Although, it appears that the some schools are maxing out the test and distorting the data.