One of the very real reasons that sporadic, non-familial autism is rising is that average paternal age is very high. After a man's early 30s, sperm precursor cells have many more mutations. This is not the only cause of autism, but a very important one. It takes some research to see that for over 50 years it has been known by those who read the research that it is the father's age or the maternal grandfather's age at the conception which matters to future generations.

Link

Not sure about the age of sperm theory. My husband was only 27 when our son with autism was conceived. He was 37 when our neurotypical daughter was conceived. But it may be true in some cases.

In some excellent epidemiological studies conducted by the Israeli military, increasing age of the father correlated to an increased chance for a child with autism.

However, young age did not correlate with a zero chance.

Having one child with autism increases one's chances of having a second autistic child by about 10x, if I recall correctly, but that's still less than a 10% chance for the succeeding child to have autism, and a >90% chance of having a neurotypical child.

It's a statistical chance, not an absolute one. One example can fall well outside the most likely case predicted by the stats.

TK Kenyon

Great read, entertaining and informative article.

Autism is a very tricky condition to diagnose, and sometimes what looks like mere slowness or retardation can be autism. I learnt this when we still lived in the States. A lady I knew turned out to be autistic to a degree. Sure enough, she had severe problems with her digestive system, and she had neurological problems as well.

And she has two daughters who have problems with autism. Her husband was over thirty when all of their children were conceived. The son, is normal, bright, in fact, and one of the few things I really regret about having moved away from the States is that we lost our influence over the boy.

Oh yeah, they watch an awful lot of TV and have cable ;o)).

In fact, Waldman's results are clearly statistically valid. He showed, county by county, that rainfall correlated with autism, and cable TV correlated with autism. Your statistical arguments do not in any way contradict Waldman's results. Waldman and co-workers use statistics skillfully and are fully qualified to do so.

Of course there are genetic predispositions, but that is probably true for most diseases, and does not necessarily help with finding the direct causes of the diseases. Our genetic equipment does not change in decades. There has to be a vitally important environmental trigger to account for the recent dramatic increase in autism.

We need to be searching hard, and with an open mind, for that trigger. Hence Waldman's research deserves very serious attention.

I have an easy-to-read article on my website at www.jungny.com which discusses austism and the statistics of early childcare in more detail.

In the above comment, McDowell fundamentally misunderstands the nature of genetics and evolution. A particular person's genetics never change, though gene expression patterns may, due to many environmental factors such as chemicals or viral infection, among many others. The MET paper correlates a promoter genotype with an increase in the possibility of an autism phenotype. Indeed, the low-MET phenotype may allow damage to another signaling system altogether, or it may worsen damage that a normal genotype may have repaired, or a thousand other explanations, but the results are statistically significant.

Of course there are many factors at work in this very complex disease, but I feel absolutely confident in stating that TV watching is not one of them.

The fact remains that the neurological changes that cause autistic behavior occur before birth, before any television watching is possible.

Television watching might exacerbate already autistic behaviors, possibly by a neurological response to the flickering pixels or by triggering the flight-or-flight response by camera angle changes or volume changes, but it cannot cause damage in utero.

By scientific standards, that autism-TV statistics paper was absolutely terrible. As I stated in the article, even statisticians don't believe most events are monocausal. Waldman, et al, did not attempt to research whether children who were later diagnosed with autism watched more TV than their neurotypical counterparts, or indeed, whether they watched any TV at all. They did not pinpoint which houses subscribed to cable and whether those kids had a higher incidence of autism. They did not endeavor to question whether those TV sets were on.

I can poke holes in the Waldman paper all day long. It was a lazy paper, with haphazard techniques and sweeping pronouncements instead of hard work. If I was a reviewer at a journal and this paper came in, I would have rejected it outright and sent it back with a long list of studies that the authors needed to perform before it should be resubmitted anywhere.

One must also note that it was published by "The Johnson School Research Paper Series 01-07," a compendium of research papers from The Johnson School, which is the business school within Cornell, and not by a respected, peer-reviewed journal (like the MET paper was in PNAS.) I'm not a journal snob, but this paper was not really "published" in the scientific sense of the word. This is self-publishing, the IUniverse of scholarly papers. It was published by press conference, a hallmark of shoddy research, like cold fusion.

TK Kenyon