Slashdot Mirror
Scientific Brain Linked to Autism
squoozer writes "The BBC is reporting that a leading scientist in area of Developmental Psychopathology, Professor Simon Baron-Cohen, is indicating that there is good chance that there is a scientific basis to the observed phenomenon that children with highly analytical parents are more likely to be autistic. He believes the genes which make someone analytical may also impair their social and communication skills. A weakness in these areas is the key characteristic of autism."
This certainly doesn't explain Boomer Esiason's kid.
Baron-Cohen's evidence is called "ecological correlation". As another poster has pointed out already, ecological correlations are subject to the "ecological fallacy". However, ecological correlations are a good way to do preliminary investigations into unknown sources of new pathologies. As anyone who has had to dignose a complex system knows, you start by gathering gross phenomenology about the system which is rather inexpensive, and then start teasing apart the various potential confounding variables as you find promising (but possibly deceptive) lines of research validated by the preliminary data.
We're at such a primitive state of understanding of the phenomenon of autism that there is intense disagreement as to whether there as actually been an explosive epidemic in autism over the last 20 years or whether there is simply an explosion in the number of social workers who are prone to make the diagnosis given the same kinds of problems they've always faced.
Without getting into the nuances of this debate we can simply say this:
If we behave as though there is a real explosion in the number of cases, we are acting wisely since the cost of being wrong is far less than is the cost of being wrong about there being no explosive epidemic.
Having said that, we have the first reason to discount Baron-Cohen's research:
While Baron-Cohen provides no data to back up the plausible sounding argument for why there might have been increased assortive mating among "nerds". Byrna Siegel makes the same argument. There are just as plausible arguments that assortive mating among "nerds" has decreased over the same time period -- not the least of which is the simple fact that nerds are working in male saturated environments where availability of mates is low, the cost of living is high and job stability is low. In other words, nerds are reproducing at a much lower aggregate rate than they used to, when they were living in more scattered, more gender balanced, more affordable and more inbred rural towns.
The second reason to discount Baron-Cohen's research is that he doesn't use the technique of "strong inference" which you really have to do when you're dealing with such a tenuously supported preliminary investigation. Strong inference means taking at least 2, preferably more, hypotheses and subjecting them to similar tests to see which of them wins in a rational comparison. There are lots of suspected ecological correlations out there -- mercury to autism, vaccination to autism, etc. and he doesn't compare the degree of his ecological correlation to the degree of these other ecological correlations. Note that what I'm not saying here that the ecological fallacy isn't in play here, nor am I saying that there might be better data supporting or refuting a given hypothesis (say, statistical case studies of individuals). What I am saying is that if you're going to use strong inference you need to apply similar tests to the different hypotheses and see which of them comes out on top so you can prioritize your subsequent research rationally.
The third reason to discount Baron-Cohen's research is that he doesn't even provide hard numbers for the nerd-autism ecological correlation (this is giving him the benefit of the doubt that nerds aren't having their effective fertility destroyed by other ecological factors).
So what would happen if you tried to do a real, strong inference ecological study of autism comparing the various hypotheses against each to see which has the strongest ecological correlation?
You'd come to the conclusion that the place to look for the cause of autism's explosive increase over the last 20 years is in areas of high Finnish ancestry where something is impo
Seastead this.
It's really too bad that Dilbert was taken off the air. That show was the shit.