Home Educated and Now Adults: Their Community and Civic Involvement, Views about Homeschooling, and Other Traits

7306 respondents, of whom 5254 were analysed after excluding anyone home-educated fewer than seven years

US adults who had been home educated during their K-12 years — the first cohort of the modern homeschooling movement; predominantly white, Protestant, and skewed young (a large share aged 18-24)

The canonical developer-led adult-outcome survey, and the origin of most "adults who were homeschooled are civically engaged and functioning competently in every way measured" claims in circulation. It is a descriptive survey with no comparison group of its own: respondents are compared against published national averages for the general adult population, which is not a matched comparison and does not adjust for age, education, income, race or religiosity. Recruitment ran through homeschooling organisations' newsletters, word of mouth and personal networks — a frame that selects on continued affiliation with, and therefore favourable attitude toward, the homeschooling movement, and that systematically cannot reach alumni who left it. No response rate is computable because the sampling frame has no defined size. 2,052 respondents (28%) were dropped for having been home educated fewer than seven years, without a stated rationale, which removes precisely the crisis-driven and mixed-sector cases. No parental income or education data were collected, so the confound that would explain the results is not even measurable. The author founded and directs NHERI, the movement's research arm, and published the report through it; no peer review. Grade D, and the grade is generous.

Ray reports that home-educated adults are heavily involved in local community organisations, more active in politics and civic affairs than the general population, and report high levels of life, job and financial satisfaction; elsewhere he reports lower rates of criminal conviction. The archive's reading is that these numbers describe the population of enthusiastic homeschool alumni reachable through movement newsletters, and that essentially nothing follows about home education as a cause. Two things make that reading concrete rather than reflexive. First, when the same civic-participation question is asked of probability samples, the answer changes sign: Hill & den Dulk 2013 (NSYR panel) find homeschool volunteering odds around 0.42 of public-school peers, and Sikkink & Skiles 2015 (random-sample Cardus) find less volunteering and less charitable giving. Second, Cheng & Watson 2025 — also a probability sample — find MORE volunteering among long-term homeschoolers only. Ray's result is unfalsifiable by its own design and should not be counted as evidence in either direction; it is retained as the source of a widely repeated claim rather than as support for it.

High, untreated, and unmeasurable in this dataset — the report collected no parental income or education data at all. Selection into the sample is additionally correlated with the outcome (movement-affiliated adults are by definition civically organised around homeschooling), which is a selection problem on top of the heredity problem.

Never replicated by anyone outside NHERI, and contradicted on its headline civic outcome by two independent probability-sample analyses (Hill & den Dulk 2013; Sikkink & Skiles 2015) while being supported on that outcome by a third (Cheng & Watson 2025). The pattern across those four is the finding: whenever the sampling frame stops selecting on movement affiliation, the civic advantage becomes unstable or reverses.