Does television viewing trigger autism?

Gregg Easterbrook says yes, citing this new study.  Here is part of the abstract:

…we empirically investigate the hypothesis that early childhood television viewing serves as such a trigger [for autism].  Using the Bureau of Labor Statistics’ American Time Use Survey, we first establish that the amount of television a young child watches is positively related to the amount of precipitation in the child’s community.  This suggests that, if television is a trigger for autism, then autism should be more prevalent in communities that receive substantial precipitation.  We then look at county-level autism data for three states – California, Oregon, and Washington – characterized by high precipitation variability.  Employing a variety of tests, we show that in each of the three states (and across all three states when pooled) there is substantial evidence that county autism rates are indeed positively related to county-wide levels of precipitation.  In our final set of tests we use California and Pennsylvania data on children born between 1972 and 1989 to show, again consistent with the television as trigger hypothesis, that county autism rates are also positively related to the percentage of households that subscribe to cable television.  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.  These findings are consistent with early childhood television viewing being an important trigger for autism.

I am unconvinced.  Precipitation, in these states, is a coastal phenomenon and is proxying for heterogeneity in the gene pool.  Perhaps the coastal areas attract a more "autism-ready" group of individuals.  In fairness to the authors, they do try to control for income and education and population density and diagnosis capacity, among other variables.  Note two worrying features in the results: in California precipitation is not correlated with autism rates at all (there is a north vs. south split for rain, rather than the coast vs. inland), and precipitation is a better predictor of autism than cable viewing is directly. 

Here is the latest autism news on the genetic front.

Addendum: Steve Levitt is also skeptical.


One of the common theories about rising levels of autism is increased pollution. The precipitation levels and cable adoption indicators both support the pollution theory. Pollutants are more common in urban areas where cable was adopted sooner, and precipitation can increase pollution by washing pollutants into groundwater, which then becomes drinking water (also by moving people indoors where pollutants tend to concentrate).

I know we're meant to be more persuaded of results arrived at using exotic instrumental variables that have some utterly meaningless and coincidental positive correlation with the variable of interest... and yet, I never am. Rainfall just can't be a good proxy for so many different things...

And I thought this was a joke. Then I RTFA. Sigh.

Whatever the validity of the study's conclusions, it's interesting that there's a correlation between precipitation and autism. Also interesting that this correlation does not hold north/south in California. All in all, seems progress has been made.

Such vacuous analysis makes us skeptics wonder if there is really any such thing as autism.

When does autism express itself? My children's pediatrician recommended abstaining from TV until my kids are two. They're 7 mo. old now, and they seem to behave generally in a way totally inconsistent with autism (or at least my limited understanding of it). Does/can this change?

Just based on Tyler's entry, I think the authors mean to use the precipitation thing as a marker for higher than average TV watching. The cable thing is interesting too. Do people with cable watch more TV? Probably. What about the programming, could that be a factor too? We don't measure the amount of TV watched in every single community, but we measure rainfall just about every where, since the two are positively correlated I think the authors are using it as a proxy for increased TV viewing.

Maybe I'm simplifying it way too much, but I hope someone will point out where I'm wrong here, here's what makes sense to me:

- Babies brains are still developing after they are born, basically the synapses get wired together after birth as they learn. Right?

- Constant simulation, or over-stimulation could cause those connections to get wired improperly.

- Could those mis-wired connections in the brain stunt the kid's ability to communicate?

Is this possible? Think how much TV programming has changed, and how much we stimulate our infants/toddlers. Any comments?

Donald wrote:

First, we don't really know whether children in high-precipitation areas actually do watch more TV--that's an assumption for which no evidence is presented. Second, even if we validata that assumption, we're still left with only a correlation between TV-watching and autism, with n causal mechanism.

They actually do show that children in high-precipitaton areas watch more television. From the abstract:

" Using the Bureau of Labor Statistics’ American Time Use Survey, we first establish that the amount of television a young child watches is positively related to the amount of precipitation in the child’s community. This suggests that, if television is a trigger for autism, then autism should be more prevalent in communities that receive substantial precipitation. "

The results from that are in Table 2 if you are interested.

Donald wrote:

"Finally, there has to be a large number of things, from activities of children to
physical conditions in the areas with more precipitation (more vegetation, leading to more pollen, for example), so the assumed correlation of precipitation with TV watching can't be the only possibility."

I think you are right that this test is unable to differentiate between the multiple pathways that precipitation could be influencing autism diagnosis rates. I am nonetheless intrigued that there could be any causal relationship between rain and autism, as that does narrow the theories down somewhat.

Also, note they mention the Amish community - which I thought was interesting. Since Amish don't watch television (hard to do when you don't have electricity) then they should have several hundred cases of autism based on the rest of the population. But the Amish have unusually low rates of autism - fewer than ten cases, if one informal study the authors note is correct. Another piece of evidence to consider.

TV as a contributor to autism sounds like a long shot, but autism is such a severe problem that it deserves to be investigated. So, here's a more promising natural experiment, with a much sharper increase in television watching:

The South African experience. The regime allowed no television until experimental broadcasts in 1975, followed by full-time broadcasting in 1976. South Africa is unique among countries with a large modern health system in its late, sudden adoption of television during an era when autism's defintion was more clear than during the 1940s and 1950s in America. I have no idea what the autism trends were in South Africa, but this might provide a good test.

The problem with using proxies in a situation where it is already difficult to find properly controlled cases is that it introduces even more potential lurking variables. For example, Parents with children who have special education needs, such as autisic children, tend to migrate to areas with good special education programs, which are also likely to have better screening. Since autism has a heritable component, migration induced by parent or sibling cases can create a correlation between initial diagnosis and location. While incomes were controlled for on an individual level, average incomes will have a causitive relationship with school funding, which will have a causitive relationship with special ed funding. If cable and precipitation correlate with regional income beyond personal income effects (which they probably will in areas where major commercial centers are ports and thus costal), then it is still possible to pick up an income effect, which could only be disentangled if location was controlled for, which would control out the two proxies.

How does Autism correlate with droughts?

Year on year variation in precipitation would eliminate genetic and socioeconomic variation. It still wouldn't affect other precipitaion causal mechanisms (dust, pollen etc.)

Did anyone check to see what the correlation between rainy counties in California (um, San Jose?) and fathers who are engineers might be? Y'know, kinda geeky guys who tend to be shy and, uh, maybe get married later in life and HAVE CHILDREN LATER THAN THEIR PEERS?

And TV viewing causes autism? Don't the initial signs show up at about 18 months (shortly after a series of vaccinations and therefore the source of the post hoc ergo propter hoc link to thimerosal)? Okay, I'll give the authors the benefit of the doubt and conjecture that maybe between the ages of 0 and 18 months these kids start making the choice to stay inside to watch TV.

Ok, kids, let's open up our "Be a Published Social Scientist Kit" and start doing science! It rains in Germany, there are lots of geeky German engineers, engineer graduation rates are a proxy for autism, rain causes autism, QED. Ever notice that there are relatively fewer engineers in the Sahara? Eh? Especially the former French colonies? Eh? Eh? A nod's as good as a wink to a blind bat, knowhatahmean, squire?

I wonder if smaller families have something to do with autism? Perhaps children without siblings to constantly annoy them are more likely to become autistic? I also think that a generation or two ago a mildly autistic person who was quite and good at performing repetitive tasks would be considered a good worker rather than suffering from a condition. When I was a child the adults in my life all seem obsessed with getting me to sit down and shut-up. Nowadays adults want children to socialize instead of just sitting down and shutting-up so it doesn't seem surprising that more children are being diagnosed as autisic now.

I have a son with autism and have done much research. First, California's state laws and universities make it an extremely popular place for families with autism to relocate to. Services for individuals with autism are extremely unevenly distributed, and many families move in order to get help. California universities are at the forefront of autism treatment research. It is extremely difficult to sort out the true rate of increase in California because of the effects of mobile population. Just about all solid research indicates that autistic individuals are born, not made. My son had odd behaviors from the moment of his birth, although he has made significant progress.

One of the reasons TV watching is thought to be implicated, is that autistic individuals tend to repeat statements that they hear in their environment, especially funny or exciting statements. Autistic children frequently repeat phrases they heard on TV, over and over and over. This is a symptom of autistic behavior, not a cause of autism. We have never had a television in our house, so my son repeats phrases that my wife and I say, over and over. I can see how an uninformed person might assign causation to that effect. But a researcher has no excuse for not being familiar with the condition and the literature surrounding it. This is extremely shoddy work.


"Classic" autism diagnoses may also increase because of a more (or less) accurate classification of conditions with similar symptoms even when the subject is clearly symptomatic. For example, a parent who has heard about autism may stress problems with verbal communication when describing symptoms, which could affect the outcome of differential diagnosis of autism vs mental retardation, or developmental delays, which would affect the outcome of differential diagnosis of Asperger's syndrome vs "classic" autism.

Comments for this post are closed