This story digs into Indiana’s statewide voucher program, the largest of its kind in the country. We uncovered an important trend in the Indiana education system – how the state created a system of choice in which schools, not parents, where often the ones choosing. By analyzing state data, we compared special ed rates in voucher schools to their public school neighbors in every district. Indeed, we found that special education rates were considerably – and consistently – higher in public schools. We also found evidence of voucher schools denying admission to LGBTQ students.
Vouchers can be a complex system to grasp, and we sought to explain our findings through clear, focused visuals. The story incorporates several graphics that show the patterns from our analysis – how the voucher program developed to admit more white, suburban students and largely cover students who would not have attended public school in the first place. In a graphic examining special education rates across Indiana’s counties, we overlay each data point to show the unmistakable pattern that schools across the state admit far fewer special education students.
What makes this project innovative?
Although the story focuses on Indiana, school vouchers are a politicized issue up for national debate. By examining the "model" state for private school choice, we provide a focused explanation of the system while making it accessible for a national audience.
What was the impact of your project? How did you measure it?
Source and methodology
To perform this analysis, we collated data from two different systems: annual reports from the [Indiana Choice Scholarships Program](https://www.doe.in.gov/choice) (whose appendices included data about voucher enrollment by year per private school, as well as data about which private schools were drawing voucher students from which public schools) and school enrollment data from the the Indiana Department of Education's [COMPASS system](https://compass.doe.in.gov/dashboard/enrollment.aspx).
First, we pulled the voucher program data tables out of the report PDFs and tidied them for analysis. Then, using the COMPASS system, we pulled special education and overall enrollment data for these private schools (bulk download for most, but several we had to look up one-by-one manually). We also used the COMPASS system to download special education and overall enrollment data for the public school districts.
Using pandas and [Jupyter Notebook](https://github.com/nprapps/school-choice/blob/master/special_education.ipynb)), we merged the available data about each voucher-accepting private school, then aggregated the private school data by public school district. This allowed us to calculate special education enrollment as a share of all enrollment in private schools that draw from a given public school district. Then we merged the data about special education and total enrollment in each public school district, allowing us to make private vs. public school comparisons. In almost all cases, private schools were underserving special education students, compared to the public school districts they drew from.
Then we did a few design explorations to find the best way to depict this disparity. We ultimately landed on a kind of interactive slopegraph, with public school special enrollment percentages across the top and private school percentages across the bottom, with a line connecting the two values for each public school district.
Reporting and writing: Cory Turner, NPR; Eric Weddle, WFYI; Peter Balonon-Rosen, Indiana Public Broadcasting