This, kiddies, is how NOT to do research. (Originally written in October, 2006. Reposted here in response to a post in the LiveJournal Asperger community.)

You may have seen an article flying around the blog- and media-space online lately, about some idiotsresearchers at Cornell University who claim to have found a link between early childhood TV-watching and rates of autism. The media's assertion, based on the Cornell report, is that early TV watching - get this - causes autism.

Lest you think that I'm blowing it out of proportion, here's a link to the article. I'm posting and commenting on some relevant quotes here:

Quote 1: "Watching too much television in childhood could cause autism, experts have warned." Quote 2: "The latest study, in the U.S., looked at whether there was a link between rising rates of autism and the increasing availability of children's TV, videos and DVDs. The research team concluded that the results were "awfully suggestive" of a link between watching TV and autism."

Follow the bouncing media error: the headline says "could cause," but then we get "a link between" in the second quote. We're talking about causation in the first case, and correlation in the second, which the media have now conveniently conflated for your continued incomprehension. *growl*

It gets worse from there. The researchers themselves recommend keeping kids away from the TV as a preventive measure: "[Dr Waldman] recommended those under two did not watch any TV at all, while older youngsters limited their viewing to an hour or two a day." To me that suggests that they, the researchers, believe television-watching is causative. How can they possibly claim that?

And as a researcher, I'm looking at their methodology as reported in the article and it's so shoddy that it's beyond belief. I can't believe that Cornell would have let anyone stamp their university's good name on this junk science:

"As they were unable to obtain any statistics on toddlers' TV habits, they used rainfall levels in different parts of the country to help estimate how much time children spent playing outdoors. (Emphasis mine.) They found that the wettest areas, where, presumably, children spent more time indoors watching TV, had the highest rates of autism. [...] Researcher (sic) also found that areas with the most cable TV customers had the highest rates of autistic children.

So, then, if we follow the bouncing logic ball, the real claim is that children with autism just aren't socialized enough and don't spend enough time playing outdoors, and that if we socialize them better, they won't be autistic. (Sounds very Skinnerian to me. Behaviorism again. Hasn't this been discredited yet?)

Granted, this media report is certainly not telling the entire story. I would need to read the reseach in order to critique it any better than this. But from the media report, it appears that these researchers are claiming, or the media are attempting to use it to claim, that we wouldn't have autistics if we didn't have television, which is frigging ridiculous. There are plenty of historical incidents of autistic traits reported in almost every genius you can name. Did television cause Edison to demonstrate autistic traits? Or Einstein? Come ON, people.

Here's the comment I left on the Daily Mail's website:

This is the problem with the media's reporting of scientific research results. First, all the article states is that there is a correlation between autism in children and television viewing. A correlation simply means that two variables are statistically associated with each other. It does not automatically indicate causation!

Secondly, while it may be true that autistic children watch more television than non-autistic children, it is far more likely that autistic children like to watch TV because it's safe, predictable and familiar to them, and doesn't involve doing things with their bodies which bring up all their balance and coordination deficits, or trying to communicate with non-autistics, which is tiring and uncomfortable. If the media report of the researchers' assertions is indeed correct, then the researchers don't know how to do research, plain and simple. Their methodology as outlined in the article is highly suspect, and their conclusions are laughable.

My partner also points out that for autistic children, who are usually quite highly visually-oriented, TV-watching is quite possibly a visual stim. Even if it's not visible to the naked eye as such, a television is a light that is flashing 25 to 30 times per second. Many autistic children love blinky lights and will stare at them for hours (that's usually what's behind the fan-watching thing, too - the blades of the fan create interesting light patterns).

In any case, these researchers ought to be ashamed of themselves for a number of reasons, and the media, as usual, has taken something that sounded juicy and plastered it all over the net without bothering to do any fact-checking first, upsetting more parents and adult autistics in the process, and giving the anti-autism groups more ammunition that has all the strength of wet tissue paper when examined with a critical eye. That won't stop CAN and other such groups from bandying it about like a prize fish, though, any more than the anti-gay groups have ever had problems using Paul Cameron's idiotic studies to back up their claims.

This is becoming more and more of a hobbyhorse for me, and when I present my CSA paper next month, it's going to be part of the paper. This kind of junk research needs to frigging STOP. So does behaviorism, but that's for another article and another day.

[ETA: Oh, this just gets better and better. According to articles about this in WebMD and Science Daily, the people doing this research were - get this - business professors and economists.

WTF are business professors and economists doing research on autism for?!? Isn't this, you know, a bit OUT OF THEIR FIELD?]

[ETA 2: I found a link to the article. Here it is (opens as a .PDF, just be aware). Going to look at it right now, but in their abstract (the summary at the beginning) they do say this:

"Our precipitation tests indicate that just under forty percent of autism diagnoses in the three states studied is the result of television watching due to precipitation, while our cable tests indicate that approximately seventeen percent of the growth in autism in California and Pennsylvania during the 1970s and 1980s is due to the growth of cable television."

Bolded emphases mine. "Due to"? "Is the result of"? Them's causative words, folks, not correlative. Point: These guys don't know how to do research.]

[ETA 3: And in just skimming the research, I find all kinds of admission that indicate they're not as sure about their results as they claim to be:

"Our empirical methodology assumes that autistic children spent their first three years of life in the same county where they reside when they are recorded in our data set."

That's one hell of a big assumption, isn't it?

"the California data continues to show no evidence of a positive correlation between precipitation and autism."

Yeah. And that's their main data set.

"So, although as indicated we do not believe our tests provide definitive evidence for our hypothesis, we believe the most likely explanation for our findings is that early childhood television watching is indeed a trigger for autism."

So, even though they don't have definitive evidence, they still like their hypothesis? What?

It should be obvious that this study has to do a significant amount of reaching and stretching to attempt to justify what it claims to find. It's far too much of a stretch for me.]  ¶ 4/06/2007 11:32:00 AM