The study
Mandatory Regionalization and Its Limits examines how California districts experience and navigate special education governance through regional SELPAs. It was authored by a postdoctoral scholar at Stanford and a doctoral researcher at UC Davis and published in May 2026 as a Getting Down to Facts III technical report, the research initiative hosted by Stanford’s SCALE Initiative.
The mixed-methods design paired newly standardised financial data from 120 of California’s 136 SELPAs with interviews with nine senior special-education administrators, recruited from a stratified random sample and spanning eight SELPAs across six counties.
How Evidano fit the workflow
A human-led process, with AI used as a supplement and checked at every step.
- 1Researchers completed manual coding and thematic analysis first.
419 minutes of audio were transcribed; first-cycle coding produced 150 in-vivo codes, consolidated through pattern coding and analytic memoing.
- 2Deidentified transcripts were uploaded into Evidano.
- 3Evidano generated AI themes and sub-themes used to assess alignment with the researcher’s themes and to enhance analytic reflexivity.
- 4All AI-generated codes were manually cross-checked and revised by the first author.
“an AI-powered qualitative data analysis platform that supports content, thematic, and cross-group analyses”
“The AI-generated analysis was used to assess alignment with the first author's themes and sub-themes and to enhance analytic reflexivity.”
“All AI-generated codes were manually cross-checked and revised by the first author.”
Why AI-assisted over manual coding
Slow cross-checking of themes, with limited time for a second analytic pass.
An AI pass to assess alignment and support reflexivity, with the researcher reviewing every code.
AI use was reported in the study’s methods section.
Every AI-generated code was reviewed and revised by the researcher.
Transcripts were deidentified before being uploaded.