Here is a most interesting article describing research linking autism rates in children with television watching.
The Cornell team searched for some independent measure of increased television viewing. In recent years, leading behavioral economists such as Caroline Hoxby and Steven Levitt* have used weather or geography to test assumptions about behavior. Bureau of Labor Statistics studies have found that when it rains or snows, television viewing by young children rises. So Waldman studied precipitation records for California, Oregon, and Washington state, which, because of climate and geography, experience big swings in precipitation levels both year-by-year and county-by-county. He found what appears to be a dramatic relationship between television viewing and autism onset. In counties or years when rain and snow were unusually high, and hence it is assumed children spent a lot of time watching television, autism rates shot up; in places or years of low precipitation, autism rates were low.
It’s always impossible to judge the quality of research from a mainstream media report, and there’s a lot here that’s suspect. But four hours of tv a day for kids on average? That’s damn scary.