PURPOSE
The purpose of this product is to demonstrate variation in per-pupil revenues across the country, and to highlight differences in revenues even among districts with similar characteristics.
DATA
Sources:
School district-level, 2013 finance data from the Census f33 data
County-level cost-of-living index data from the Council for Community and Economic Research
School district-level student poverty estimates from the Census, Small Area Income & Poverty Estimates (SAIPE).
School district-level median household income from the Census, American Community Survey and
School district-level, enrollment characteristics [number enrolled, FRL, IEP, LEP] from NCES, Elementary/Secondary Information System (ELSi)
Data exclusion criteria:
District-level school finance data was used as the basis of the district-level data set; other data sources were merged onto this dataset with unmatched entries from the additional data sets being dropped such that every row in the data has school finance data.
Starting with the full sample of 14,462 school districts in the school finance data set, the following exclusions were made:
1. Finance data missing an NCESID
2. Charter districts or districts where charter classification is missing [most of the districts with missing data seem to be ones we would want to exclude anyway, e.g. “computer association” districts, “interlocal cooperatives” etc.]
3. Special education, technical and vocational districts [excluded by finding districts that some form of special, technical or vocational in their name, with exceptions made for districts that appear to be more general districts]
4. Districts with ten or fewer students
5. Districts that enroll a proportion of IEP students that is more than two standard deviations (sd=.05) above the mean of 0.14
6. Districts missing cost-adjusting values
7. Districts with state+local revenues below $500
8. Districts with less than one student per 10 square miles, on average, were excluded because of cost-considerations that differentially impact sparse districts
In total, 1,711 districts (12%) were excluded leaving 12,751 districts in the Power In Numbers sample
For the regressiveness analyses, two additional exclusions were applied to the sample: 1) DC and Hawaii were removed, since they each have only one school district, and 2) school districts with less than one student per 10 square miles, on average, were excluded because of cost-considerations that differentially impact sparse districts.
METHODS
Cost-adjusting
All of the revenue figures presented are cost-adjusted to convert per-pupil revenues into figures that account for variation in the purchasing power of a dollar across different regions. See our first Power in Numbers piece for a discussion of why cost-adjusting is so important in studies of school finance.
We applied a cost-adjusting conversion from the Council for Community and Economic Research (C2ER) to each district's revenues. District revenues were adjusted using the C2ER county-level, cost-adjusting rate for the county identified by NCES as being associated with a district.
Difference between district and national average per-pupil revenues
In order to explore variation in per-pupil revenues across the country, we compare each district’s cost-adjusted revenues to the average cost-adjusted per-pupil revenues in the country. The national average of $11,866 is the average per-pupil revenue for all districts in the Power in Numbers sample.
Each district’s percent difference from the national average revenues is computed by taking the difference between district and national-average revenues, divided by the national-average revenue for each district.
Largest 200 districts peer groups
In order to demonstrate the arbitrary nature of school funding, even among relatively similar districts, we created peer groups based on cost-adjusted median household income (MHI), free/reduced lunch (FRL) rates and enrollment size. The 200 districts with the highest enrollments (DC and Hawaii were excluded because there is only one district within the state/territory, so only 198 of the top 200 are included) are divided into the following fourteen peer groups:
1. Low-income, high-FRL, large: districts with a MHI below $35,000, FRL rate of 75% or higher and enrollments of 50,000 or more.
2. Mid-income, high-FRL, large: districts with a MHI between $35,000 and $52,250*, FRL rate of 75% or higher and enrollments of 50,000 or more.
3. Mid-income, mid-FRL, large: districts with a MHI between $35,000 and $52,250*, FRL rate between 46.5%** and 75%, and enrollments of 50,000 or more.
4. Mid-income, low-FRL, large: districts with a MHI between $35,000 and $52,250*, FRL rate less than 46.5%** and enrollments of 50,000 or more.
5. Low-income, high-FRL, small: districts with a MHI below $35,000, FRL rate of 75% or higher and enrollments of less than 50,000.
6. Low-income, mid-FRL: districts with a MHI below $35,000, FRL rate between 46.5%** and 75%, and enrollments of any size.
7. Mid-income, high-FRL, small: districts with a MHI between $35,000 and $52,250*, FRL rate of 75% or higher and enrollments less than 50,000.
8. Mid-income, mid-FRL, small: districts with a MHI between 35k and $52,250, FRL rate between 46.5% and 75%, and an enrollment less than 50k.
9. Mid-income, low-FRL, small: districts with a MHI between $35,000 and $52,250*, FRL rate less than 46.5%** and enrollments less than 50,000.
10. High-income, mid-FRL: districts with a MHI of $52,250* or higher, FRL rate between 46.5%** and 75%, and enrollments of any size.
11. High-income, low-FRL, large: districts with a MHI of $52,250* or higher, FRL rate less than 46.5%** and enrollments of 50,000 or more.
12. High-income, low-FRL, small: districts with a MHI of $52,250* or higher, FRL rate less than 46.5%** and enrollments less than 50,000.
13. Very high-income: districts with a MHI of $75,000 or higher and any FRL rate or enrollment size.
14. Very large enrollment: New York City, Chicago and Los Angeles.
* $52,250 was the national MHI in 2013 according to the Census.
** 46.5% was the average district FRL rate in our Power in Numbers sample.