In a study to be released today in November's Archives of Pediatric and Adolescent Medicine, Cornell University economist Michael Waldman found that in areas of California, Oregon and Washington that experienced high levels of rain and snowfall during the years 1987-2001, autism rates among school-aged children rose when measured in 2005. Those children diagnosed with autism would have been under 3 during the periods of high precipitation, the period during which autism is generally diagnosed.Waldman doesn't actually suggest that rain causes autism. Instead, his theory is that, by keeping kids indoors too much, we subject them to a million and one possible environmental toxins (including too much TV). One of these, surely, could be the cause of autism.
Indeed, it is possible to find a statistical link between rain and autism. It's also possible to link ANYthing new or newly prevalent to autism. Make a list of everything that has changed or increased in prevalence since 1992, and you will find a statistical link between that thing and the rise of autism diagnoses. More office cubicles. More items manufactured in China. More cell phones, computers, headsets, digital cameras, MRIs, ultrasounds, anti-depressants, video games, pokemon characters, TV channels, midi files, organic fibers, genetically engineered corn, bioterror attacks, older adults... you name it.
The point here is simple. Statistical associations can be useful tools - but they are also a terrific way to create FUD (fear, uncertainty, and doubt), particularly among a group of people who are already on edge. Be careful, as you read headlines like those in the LA Times (Link found between autism and rainfall) that you separate out actual causal evidence from mathematical boondoggles.