A reader should be able to trace an education statistic to the record behind it. This page explains how ChalkLens (chalklens.org) counts students, compares surveys, reports test scores, and records changes.
1. Source order
We assess sources in this order:
- State education registries: Annual filings from the 23 U.S. states that require homeschool registration, published by state education departments.
- Federal surveys: National Center for Education Statistics (NCES), National Household Education Surveys Program (NHES), and U.S. Census Bureau Household Pulse data.
- Assessment records: Norm-referenced scores from Iowa Assessments, Stanford 10, TerraNova, ACT, and SAT, with sample and income details where available.
- Audited surveys: Academic studies and peer-reviewed surveys that report sample size, confidence intervals, and question wording.
2. Comparing national estimates
National counts don’t always agree. The 2022–23 NHES model puts homeschooling at 3.4%; Census Pulse samples and state files sit closer to 6.2%–6.5%.
We publish a blended range instead of hiding that gap. One report, for example, gives 3.408 million students, or 6.262%, with a sensitivity interval of 3.067 million to 3.749 million. It also separates independent home education from full-time virtual charter schools.
3. Reading test scores
Voluntary test samples often score higher because college-bound families take part at higher rates. We label that limitation and distinguish those results from mandatory state cohorts. Multi-variable analyses account for parent education and household income.
4. Updating state policy data
Attendance rules, Education Savings Accounts (ESAs), microgrants, and notice requirements change during legislative sessions. We review state dockets and check the statutes again each session.