I think that the fact that the survey was conducted by a government organization and that the targets of the survey were families who had decided not to participate in government-run education, for one reason or another, makes a certain selection bias seem likely.

I'd expect a measure of homeschooling families to want nothing to do with the federal department of education.