Contextualizing Homeschooling Data: A Response to Rudner

not applicable (commentary on Rudner 1999)

US home-schooling population, as represented in the Bob Jones University testing data

Published critique in the same journal issue-year as the study it addresses.

Rudner's 20,760 students are 1.28-2.08% of the US home-schooling population by the authors' own conservative arithmetic, selected twice over: families willing to pay for testing (states requiring tests generally provide a free route, so paying implies income), and families willing to route that testing through Bob Jones University. The authors note that HSLDA funded the study, that Rudner had testified on behalf of home-schoolers represented by HSLDA, and that HSLDA distributed the report to members of Congress. Their conclusion is not that Rudner is wrong about his sample but that the sample cannot bear the demographic or achievement generalisations made from it — including the widely repeated claims about home-schooler income, marital structure and racial composition.

Not applicable; the critique is about sample selection, which is the mechanism through which heritable parental traits enter the estimate.

The substantive point — that testing-service samples cannot support population claims — has been restated by every subsequent independent review of this literature.