Twisted Statistics
 https://plus.google.com/u/0/communities/113621697602518150080
There are problems with statistics like these that it may be showing a trend that unrelated to the variable being tested. Often food studies do this. People who tend to eat more than X are healthier than those who do not. The problem is that those who eat X are often living very different lifestyles than those who do not. Poverty, healthcare access, jobs that expose people to different toxic environments are variables that are not mentioned or tested. Assessments into autism vary because in other cultures due to misdiagnose, or the lack of desire to diagnose at all children who are different are not subjected to medical intervention. Most often other disabilities (such as learning difficulties/deafness/anxiety disorders) are diagnosed. When comparing vastly different cultures.
In America a child is likely to go to kindergarden at 3 years old. If he is different from his peers in anyway it is noted and sent for tests (often children may be in childcare WAY before 3 years) medical testing is recommend. The parameters for "normal" in this setting are vastly different than those say in Cuba. Children go to school later and tend to be part of large families, with many family members looking after children. Only children vastly different from siblings would be worthy of note and then tested. This would reduce people seeking tests for autism.
My husband is on the spectrum and it is our belief it is genetic. His mother was "deaf" except her inability to deal with people, basic social situations and vivid temper, in our conclusion put her squarely on the spectrum.
While extremes on the spectrum can make life very difficult for parents, their wiring is a blessing as well as a curse. I think in years to come we will look back at Newton and see someone with aspergers, the same way dyslexics do with Einstein.