1. The quote says "diagnoses of autism" not "diagnoses of severe autism" so if the definition of autism has expanded over time due to improvements in our understanding of autism (which it has) the diagnosis argument is far more effective.
2. It is easier for children of poorer people to get an autism diagnosis now than it was 50 years ago, so even severe cases might not have shown up in the stats if the child was born to a poor family.
I meant it could explain some of it, but not the full 16x times.
I am talking under the presumption that researchers today are aware are correcting for that fact that there is some selection bias (if that's what I can called) in stats.
