from your article:
That?s impressive, but we have to keep in mind: This wasn?t a random cross-section of homeschoolers.
Participants were recruited from a special subset of the homeschooling population--families who subscribed to a fee-based testing service.
Compared to their peers in the public schools, these kids were more likely to have affluent, well-educated parents. Were the parents also more committed to educating their children? Perhaps.
Oh, wow. So affluent, well-educated people who homeschool their kids, using fee-based testing do better than the average kid.
But most home schoolers aren't those things...