Just in time for Christmas, the National Center for Education Statistics (NCES) has finally released its latest figures on homeschooling – the gift at the top of any homeschool researcher’s list! This is a big deal. If you’ve read much homeschooling literature you’ve seen NCES’ 2003 data used over and over, because it’s the best we’ve got for estimating the national population of homeschoolers. In 2003 NCES estimated that 1.1 million children were being homeschooled (2.2% of school-age children). The new estimate for 2007 is 1.5 million (2.9%). This is a 36% increase in four years.
The report is so brief everyone should just click on the link above and read it through rather than read a summary from me. But some of its findings are really surprising to me and challenge some of the assumptions I’ve been making of late, so let me mention them. I had begun to believe, based largely on the statistics reported by some states and on declining attendance at some of the biggest state homeschooling conventions, that homeschooling was not growing as fast in the last few years as it had in the past. If the NCES sample is an accurate representation of national trends, I was wrong. It must be noted that by the survey’s definition of homeschooling students enrolled in virtual schools are included in the tally, but nevertheless, this is big news.
A couple of other findings surprised me as well. The 2007 figures show a significant increase of parents who say they homeschool for religious reasons (from 72% in 2003 to 83% in 2007), and a decrease of parents who say they do so because of a child’s special needs (from 29% to 21%). But at the same time, the survey found a significant increase of parents who chose homeschooling for “other reasons,” (from 20% to 32%) confirming me in my belief that homeschooling is appealing to an increasingly heterogeneous group of people for a wide range of reasons. (Parents surveyed could give more than one reason for homeschooling)
Finally, it should be noted that there is a methodological margin of error that makes it possible that the figures NCES reports here are either too high or too low. We are dealing with extrapolations from a limited sample. But it is the best data we’ve got, and we’ll certainly be seeing it over and over for the next several years.