Home vs. public schoolers: Differing social opportunities

41 (21 home-schooled, 20 public-schooled adolescents)

Adolescents in central Oklahoma, mean age 14.48 (home) and 14.85 (public); home-schooled group had been home-schooled a mean of 4.10 years and came from religiously/ideologically motivated families

Small, geographically confined, self-selected — grade D on every axis — but methodologically the most informative study in the socialisation corpus, because it measures BEHAVIOUR rather than opinion. Participants kept a diary of every interaction lasting two minutes or more for one month, then rated closeness on a 3-point scale. No parent-report anywhere in it. That design converts "are homeschooled children isolated?" from a matter of whose rating you trust into a countable quantity, and the answer turns out to be non-obvious: total social contact is equivalent, but its composition and its intimacy are not. Limitations the author states: n = 41, one region, one season (spring), exclusively religious/ideological homeschoolers, no controls beyond rough age matching.

Home-schooled adolescents were NOT socially isolated in volume — total contacts 49.29 vs 56.25 for public schoolers, not significantly different — and had significantly MORE contact with younger children (8 vs 4, p < .05) and no deficit in contact with older people. What differed was peers and depth. Peer contacts were roughly a third of the public-school figure (11.0 vs 30.0, p < .01); public schoolers reported significantly more daily interaction (p < .01); and public schoolers rated their relationships as closer overall (2.12 vs 1.90 on a 3-point scale, p < .035). The author's conclusion is the honest formulation of the socialisation question and worth preserving verbatim in spirit: home schoolers are not "at risk" in the number of people they interact with, but are "at risk" in feeling less closeness toward, and receiving less support from, their peer friendships. Note that Medlin's reviews cite the peer-contact contrast as 11 vs 20 rather than 11 vs 30 — the direction is agreed, the magnitude is reported inconsistently across secondary sources and has not been checked against the original by this archive.

High and untreated. Families self-selected into home education and the samples are not matched on parental education, income, religiosity or child temperament. A child with fewer peer contacts and lower rated closeness may have had those traits before, or be from a family that values them less. Nothing in the design separates the schooling arrangement from the family that chose it.

Never replicated with a larger or better-matched sample, and thirty years old — the homeschooling population has since diversified far beyond the religious/ideological families sampled here, and co-ops and hybrid programmes have grown, so the peer-contact figure should not be assumed to hold today. Hamlin 2019 (NHES, nationally representative) finds homeschoolers heavily involved in organised out-of-home activities, which is consistent with the "volume is fine" half of this result but does not test the "closeness is lower" half. Nothing in the literature has retested the closeness finding at all.