Stupid Stats

C’mon, now… really?

Uterine size in non-pregnant women varies in relation to age and gravidity [number of pregnancies]. The mean length-to-width ratio conformed to the golden ratio at the age of 21, coinciding with peak fertility.

Claiming that a uterine golden ratio coincides with peak fertility is highly suspect. The good folks at Ava Women claim that, “Most women reach their peak fertility rates between the ages of 23 and 31.” Information at Later Baby states, “Female fertility and egg quality peak around the age of 27.” And WebMD says, “A woman’s peak fertility is in her early 20s.” So, there seems to be some debate about when peak fertility actually occurs. Consequently, this strikes me as retro-fitting, and it seems that Dr. Verguts and his colleagues may have played loose with the age of peak fertility in order to make a connection to the golden ratio.

In their defense, though, it’s not the first time that folks have gone uptown trying to find a connection to the golden ratio. A claim by The Golden Number states, “[The DNA molecule] measures 34 angstroms long by 21 angstroms wide for each full cycle of its double helix spiral,” and 34/21 ≈ 1.6190476, which is approximately equal to φ, 1.6180339.

Though this guy — an honest-to-goodness biologist — seems to disagree:

I’ve also heard folks say that people are perceived as more beautiful if certain bodily proportions are in the golden ratio. The most extreme example of this that I’ve found involves the teeth:

…the most “beautiful” smiles are those in which central incisors are 1.618 wider than the lateral incisors, which are 1.618 wider than canines, and so on.

In a study of 4,572 extracted adult teeth, Dr. Julian Woelfel found the average width of the central incisor to be 8.6 mm. If the teeth in a beautiful smile follow the geometric progression described above, well, that would imply that the first molar would be just 8.6 × 0.618⁵ ≈ 0.8 mm wide, which isn’t reasonable and, moreover, is not even remotely close to the average width that Dr. Woelfel found for the first molar: approximately 10.4 mm.

But all of these claims involving the golden ratio are not even close to being the stupidest statistics I’ve heard in my life. Mary Anne Tebedo made a remark on the floor of the Colorado State Senate in 1995 that may hold that distinction:

Statistics show that teen pregnancy drops off significantly after age 25.

Of course, it’s hard to call that a statistic, since it’s completely nonsensical. Maybe it’s only the stupidest statement I’ve ever heard.

Then there’s this one, from the New York Times on August 8, 2016, which couldn’t be more useless:

No presidential candidate has secured a major party nomination after an FBI investigation into her use of a private email server.

Well, duh. Email didn’t even exist before the 1970’s. Moreover, besides Hillary Clinton, has any presidential candidate ever had their use of a private server investigated by the FBI? This is like saying, “No one has ever been named People‘s Sexiest Man Alive after writing a math joke book.” (Not yet, anyway.)

Randall Munroe made fun of these types of “no politician has ever…” claims in 2012 with his cartoon Election Precedents:

And it’s true:

But perhaps my all-time favorite is one that Frank Deford — may he rest in peace — included in his piece “The Stupidest Statistics in the Modern Era” on NPR’s Morning Edition:

He’s [Brandon Phillips] the first National League player to account for as many as 30 steals and 25 double plays in one season.

About this stat, Deford commented, “Steals and double plays together? This is like saying, ‘He’s the first archaeologist to find 23 dinosaur bones and 12 Spanish doubloons on the same hunt.'” (I sure am going to miss him.)

The preponderance of dumb stats shouldn’t come as a surprise, though. A recent study found that people deemed real news headlines to be accurate 83% of the time and fake news headlines to be accurate 75% of the time. So, if we can’t tell truth from fiction, how can we possibly distinguish useful statistics from inane?

If you’d like to test your ability to detect fake news, check out Factitious from American University.