Two months ago, I eagerly took a look at Nate Silver's new, independent news site, FiveThirtyEight.com, to see what his data-driven approach would do to improve science coverage.
Sadly, the lead story in his science vertical was an unfortunate piece that pretended to provide a statistical analysis of health coverage, but fell far short of anything approaching the rigor that Silver applies to his political polling. (For example, one term in the equation was "gut feeling.")
Two months later, I'm sorry to report that things are no better.
The lead science story today is another misleading and poorly reported piece about a very important subject: Whether the children of older fathers are at increased risk of ADHD, autism, and other "genetic problems and psychiatric and behavior disorders." (I'll disclose at the outset that I've written stories in which researchers say that the risks of older fatherhood are substantial, such as this piece from Scientific American Mind.)
The FiveThirtyEight.com story–by Emily Oster, an economist at the University of Chicago–argues that the risks associated with older fatherhood are not nearly as severe as they've been made out to be.
She begins with some silliness about fathers' discovering they have biological clocks: "I challenge any woman…not to feel just a teensy bit of schadenfreude at the increasing discussion of aging sperm and its effects on childbearing," she writes. "I’m sure we can all look forward to the day when television features 25-year-old women on dates with desperate 37-year-old men discussing their sperm-freezing plans."
This is funny, but only until you begin to think about what's at stake here. The other genetic problems and psychiatric disorders that Oster refers to sound a lot worse when you spell them out. They include schizophrenia, hydrocephalus, decreased intellectual capacity, cleft lip or palate, and others–along with the ADHD and autism that Oster mentions.
Oster then goes through a number of studies, showing their weaknesses. Aging men tend to have children with aging women, she points out, which makes it difficult to determine whose advanced age is responsible for the increased risks faced by their children. She also notes that men who have children later in life might be different from men who have children younger, and that difference could be responsible for the increased risk.
Those are fair points. But her criticisms fall apart in the face of a study she didn't mention–the 2012 deCODE Genetics study from Iceland that found that men are more likely to pass along mutations as they age. "A 36-year-old will pass on twice as many mutations to his child as a man of 20, and a 70-year-old eight times as many," the study concluded, according to a news article in Nature.
This finding sidesteps Oster's criticisms. In this research, the age of a male's partner or differences among men are not important factors. As men get older, they pass along more mutations to their children. That's disturbing. And Oster was wrong to ignore it.
Further, the deCODE study got a huge amount of attention in the press. How could Oster have missed it?
She concludes that "the jury is out on the relationship between paternal age and child psychiatric problems. Yes, there is some possibility that it matters."
Researchers have much to learn about the increased risks faced by the children of older fathers, but the risks shouldn't be dismissed. Oster's piece will be reassuring to older men, and to older parents. But parents who read more broadly will get a different impression. The evidence for risks associated with older fatherhood is substantial, and it's growing.
Readers who see only Oster's piece will not get the straight story. FiveThirtyEight is giving us bad information.
And it isn't based on data.
-Paul Raeburn
Dan Lack says
You are right on the money…and other than a spectacular record of the election predictions, other than maybe horse racing, who cares about anything else…this “data driven” business, is so overrated, but I’m sure Nate turned it into a bonanza for himself..