So... I looked up some of the papers on homeschooling research, and what I see, frankly, troubles me. The statistics in favor of this are produced by a fairly biased group (homeschool associations) and papers I see tended to have some methodology flaws.
Like this one, from a homeschooling group:
http://www.academicleadership.org/article/download/392/ALJ_ISSN1533-7812_8_1_392.pdf
Where the methodology was:
1. Parents contracted with the testing services to have tests administered to their children/students.
2. The testing services certified test administrators, some of whom were the students’ parents.
3. The testing services sent tests, answer forms, and a letter explaining how parents could access and complete the questionnaire to the test administrators.
4. Tests were returned to the testing services who then scored them or sent them to the test publishers for scoring. Unlike in most preceding studies, the large majority
of parents (i.e., the parents of 69.4% of the 11,739 students included in the study) did not know their students’ scores ahead of time; that is, before completing the
questionnaire and thus participating in the study.
5. Electronic copy of the test results and survey questionnaire results were sent from the testing services and the online survey administrator to the researcher. These
data sets were merged to provide 11,739 cases with matching identification numbers
Now, in public school, I know darn well who sat in what chair and took what test and how long it took and whether there was any interaction (asking for clarification, attempting to peek at books.) For this round, they just sent out tests and accepted them -- there's no evidence of who actually took the test and how they took it.
As a scientist, I find this result unsupported.
Furthermore, I find that some of the major studies on homeschooling come from the Bob Jones University and I believe they are not a neutral and objective source since they produce materials (Creationist and Biblically oriented) for homeschooling (
http://en.wikipedia.org/wiki/Bob_Jones_University).
From (I hope this pastes correctly):
http://ericae.net/digests/tm9905.pdf
In Spring 1998, 39,607 home school students contracted to
take the Iowa Tests of Basic Skills (ITBS; grades K-8) or the
Tests of Achievement and Proficiency (TAP; grades 9-12)
through Bob Jones University Press Testing and Evaluation
Service. Students were given an achievement test and their
parents were asked to complete a questionnaire entitled
“Voluntary Home School Demographic Survey.” A total of
20,760 students in 11,930 families provided useable
questionnaires with corresponding achievement tests. The
achievement test and questionnaire results were combined to
form the dataset used in the study.
For the record, I've seen the Bob Jones school material recently. Its content is appalling (the gist of the history book that I reviewed was that the Native Americans ran around and threw sticks at each other until the Christians came in to build this great nation and all the true progress in this nation was made through missionaries and the kind guidance of White Christian Males.)
As an Uppity Woman Anthropologist, I found their version of history to be blinkered with the same paternalistic assumptions as my own education (in 1950, when Blacks were still called "boy" and our shameful family secret was that my great-grandmother was Native American.)
From NCES, I find stats on home schooled children, but not age vs expected grade level:
http://nces.ed.gov/programs/digest/d10/tables/dt10_040.asp
And although I see a highly touted report that homeschooled students did better than the national average on the ACT, there's no actual set of statistics from the company on how many home schoolers took the test and what their score was. And what percentage of homeschooled kids actually apply to college.
http://www.act.org/newsroom/data/2006/charts/index.html
As a scientist, I'd like to know where that number came from and how it was reported (if self-reported, I imagine that parents whose kids did NOT score well are not going to run out with "We got a total score of 150 on both sections of the ACT!"
Lots of noise, lots of flawed reports, nothing concrete -- which makes me suspicious.