Hacker News Reader

Comment by dschiptsov | Hacker News Reader

Yet another example of meme-science. Only idiots would try to explain a complex, subtle phenomena to a single environmental factor.

View on HN

"There is a large inherited component to autism, but lead researcher Dr Marc Weisskopf said there was mounting evidence that air pollution may play a role too."

The way of thinking about Autism should be like the way we think about human language. Yes, the ability to speak, or rather our abilities to acquire and learn a language are, of course, inherited, but when there is no one to speak to a child, or a child refused or unable to participate, or there are other obstacles in the environment, after passing a certain age he will never speak at all.

It seems that autism has the same pattern. It has genetic aspect and environmental aspect and one of personal experience and amount of practice, so no one could identify all the factors, leave alone find a "single cause". Of course, it is genes, for, say 80% (which is rather a flawed mental construct), like it is in everything else - height or IQ, for example (why should it be different?), but the rest 20% are, of course, not mere "air pollution".

It is the same meme-nonsense (with which we're bombarded daily by mainstream media) as to say, "look, children in a toxic environments have worse grades and misbehave more". OK, perhaps they are, but this is an "empty message", a "meme" based on a naive common-sense. Correlation does not imply causation, and even if this factor is a "valid" one, its actual implication is infinitesimally small - it, perhaps, contributes to "stress" or states related to "depression" (which is another meme). Most of these meme-studies are just misuse of statistical tools together with a confirmation and self-serving biases. People construct statistical "evidences" for their own flawed constructs, instead of methodically removing everything "mental" and see what remains.

Let's apply the Principle of Charity here. Perhaps the researchers are idiots, far below the average education and intelligence of, say, Hacker News users. But perhaps they're as familiar with such notions as "complex phenomenon" and "environmental factor" as we are.

We should choose the more charitable interpretation, not because there aren't idiots out there, but because doing so leads to more interesting discussion in the long run—and usually in the short run too.

http://en.wikipedia.org/wiki/Principle_of_charity

Why not? Are researches not hungry for grants, quick fame and cheap sensations? Do they never neglect subtleties of statistical models, like necessity of a third "control group"? Are they all flawless and brilliant, or there might be some halo effect, overconfidence and jumping to conclusions? Are they never trying find a support for their hypothesis form correlations barely distinguishable statistically? Kahneman suggest that this is rather common.

The first question is "How does they manage to properly test this "link" with a distinct control group". Could you explain, please.

So, I prefer to read Simon Baron-Cohen.

Oh, I agree with you on all of that and more, but if we're to have substantive discussion, we need to guard against generic dismissals. This is the weed that grows most abundantly around here. (Generic endorsement is probably just as bad, but we don't have that problem on HN.)

If you have specific information about this or relevant research, by all means contribute it. But railing against idiocy in a generic way lowers the signal/noise ratio of the threads: if generic it must be predictable, and if predictable it can add no information. The Principle of Charity is so good for counteracting this that we may add it to the HN guidelines.

Intellectual charity doesn't require accepting anything false. It means that when there are several reasonable interpretations, you should pick the strongest. This goes against the default tendency we all have, to pick the weakest to then have the pleasure of overturning it. That pleasure is strong—so strong that one may begin seeing idiots for it—and it overpowers the pleasures of curiosity, which is what we're going for.

The word "link" is used throughout the article, and its message seems to essentially be Yet, the research is unable to conclusively say that pollution causes autism as there could be other factors that were not accounted for in the study.

Reply on news.ycombinator.com