A couple of years ago, Baylor University released a report on home school students at Baylor and how well they performed (Profile of First-Time Freshmen from Home Schools, Fall 2005 to Fall 2009). Some home-schoolers point to this report as evidence for how good home schooling is, because of statements in it like
Home school students took a slightly higher credit hour load during their first semester compared to the entire first-time freshmen population. The home school group also had a higher cumulative GPA at the end of their first year.
Of course, the report is useless for saying much about home schooling, since it only looks at how well the students admitted to Baylor did. But to get into Baylor, home schoolers have to have very high test scores (according to the report, 61.3% of the home school students who enrolled at Baylor had SAT reading+math ≥1300, while only 25.3% of the overall freshman classes did). Because the report does not match the home school admittees to a population with a similar distribution of SAT scores, any conclusions about how well the home school students perform are meaningless. It would have been interesting to see whether there was a difference in performance, retention, levels of depression, … between home schoolers and regularly schooled students, but only after first controlling for differences in admissions policies and practices (for example, by matching SAT scores, gender, and income levels).
What one can reasonably conclude is that to get into Baylor as a home school student took exceptionally high test scores (only 6.4% of home school students who enrolled had SAT scores less than 1100, while 21.7% of the overall class did). The median SAT score for home schoolers who enrolled was 1325, while for the class as a whole it was 1200. The entire distribution for home schoolers is shifted up about 120 points from the distribution for the whole class. This difference may reflect the greater reliance on test scores rather than GPA for home school applicants, as calibration for home school GPA is difficult.
The difference is not proof of discrimination against home school students, of course, as we have no idea what the pool of home school students looked like (for all we know, Baylor may have accepted all that applied), but it is suggestive of a higher threshold for home school students.
What is most disturbing to me is that the Office of Institutional Research and Test at Baylor thought that this profile supported their conclusions about differences in performance related to home schooling—I would expect better controls from a collegiate study, even one done by staff rather than faculty.