Adventures in Autism has read an article I also came across recently in which SurveyUSA conducted a phone survey in California to see if there were any reported differences in prevalence of autism among vaccinated vs. non-vaccinated kids. In a sample of 17,674, 991 (about 5.6%) were reported as being completely unvaccinated. Among this group (the control group) autism was significantly less common.
As the article itself points out, this is not a definitive conclusion; but it does raise questions that warrant further investigation. And as adventures points out, it also proves that it should be possible to find a statistically meaningfully-sized control group for purposes of comparison.