I wrote a paragraph on that in the post. I predicted a publication bias in favor of positive results, assuming the community is not biased on the particular issue of vaccines & autism. This prediction is probably wrong, but that hypothesis (lack of bias) is what I was testing.