A recent "report" was published by physician Janette Sherman and epidemiologist Joseph Mangano that showed some scary numbers relating to the meltdown of the Fukushima Daiichi nuclear power plant in Japan. According to the essay, which was picked up by Al Jazeera English as fact, there's been a 35 percent spike in baby deaths in the Pacific Northwest since the nuclear meltdown.
Dead babies!? Terrifying news!
If it were true.
Today Scientific American writer Michael Moyer points out that the data behind the doom-saying duo's pronouncement is flawed, to say the least.
The crux of Sherman and Mangano's argument is data published by the Centers for Disease Control that measure infant mortality.
The authors took data from eight West Coast cities--Seattle, Portland, Boise, San Francisco, Sacramento, San Jose, Santa Cruz, and Berkeley--and compared the average number of baby deaths per week in the four weeks prior to the meltdown to the number of baby deaths per week in the 10 weeks after the meltdown.
From this the authors get their 35 percent increase.
This method of data mining, however, is highly flawed, as SA reports:
Let's first consider the data that the authors left out of their analysis. It's hard to understand why the authors stopped at these eight cities. Why include Boise but not Tacoma? Or Spokane? Both have about the same size population as Boise, they're closer to Japan, and the CDC includes data from Tacoma and Spokane in the weekly reports.
More important, why did the authors choose to use only the four weeks preceding the Fukushima disaster? Here is where we begin to pick up a whiff of data fixing.
This "whiff of data fixing" that Moyer smells is laid out when he compares numbers stretching back further than four weeks before the meltdown (as any normal scientific process would demand).
Upon doing this, Moyer actually finds that infant deaths have gone down very slightly.
While it certainly is true that there were fewer deaths in the four weeks leading up to Fukushima (in green) than there have been in the 10 weeks following (in red), the entire year has seen no overall trend. When I plotted a best-fit line to the data (in blue), Excel calculated a very slight decrease in the infant mortality rate. Only by explicitly excluding data from January and February were Sherman and Mangano able to froth up their specious statistical scaremongering.
Seattle Weekly reached out to Dr. Janette Sherman (a woman whose own website describes her as "Physician. Author. Activist") for comment on whether she feels she fudged her results.
She said she wouldn't comment until tomorrow because it's 4:30 p.m. in Virginia and that's apparently too late for comments.
We'll update when she does.