A New Year’s BOGUS has been born! (is it just me, or did that happen awfully fast?). And let me tell you, this
- Image by rochelle, et. al. via Flickr
one is a ten-pound whopper with a stinky diaper chock-full of…well, you get the idea. An “evolutionary biologist” published an actual paper asserting that the gender of babies is determined by what the mother eats. Apparently, good old King Henry was justified in chopping off all those wives’ heads – it wasn’t his sperm at fault. It was her breakfast.
Unfortunately, I am not joking here. To understand the destructive impact of this kind of BOGUS reporting, imagine you live in a country repressive to women (don’t feel constrained, there are lots to choose from). If you can get stoned for adultery after you were gang-raped, how damaging do you think this media-spread reversal-of-fact will be – that women are now responsible for the births (or not) of sons?
So how does someone get a paper this BOGUS (full of this much mom-blaming misinformation) published, and then widely covered in the (NPR!) media? First, here’s the way gender-selection actually works: The simple matter of an X or Y chromosome (from, exclusively, the sperm – all eggs are Xs) determines the gender of a baby. XY is a boy. XX is a girl. That’s been known, since, oh, about the time my grandmother could have peered into a microscope (if, that is, she hadn’t been just a eensy bit busy raising cotton, as well as seven kids).
An “evolutionary biologist” (which, BTW, seems to be code for: someone-who-says-whatever-s#*t-pops-into-his/her-head) decided to “analyze” (obviously without ever taking an advanced statistics course) factors about mothers’ lives to see what popped out as marginally significant in gender selection. Then, oh goody! let’s publish it!
Ah the curse of modern software…garbage in/garbage out.
See, the other, well-known (by those who care to know) issue here is that if you take ANY group of people, and analyze factors about them – something will pop up as marginally significant. The more factors you throw in (and there were over 130 in this study!), the more false positives you will get. This is a mathematical fact.Hence the need for both something fancy, called a “logistical regression model” which tries to dampen down this effect, and the need for something simple, called basic biologic facts (or, as may grandmother would say, “horse sense”).
Say, for example, we take all U.S. Senators and Representatives (this being Inaugural week) and we decide to analyze factors that might determine which seat they choose when Congress is in session. We make them keep food diaries, and list all the grooming products they use, their birth order, income, and number of stars in their daily horoscope. Our analysis shows that high-fiber cereal, Aqua-Net hairspray, and a horoscope stating “you will meet the man of your dreams!” are all significant factors in where a Congressional butt lands. This is great stuff, isn’t it?
So do we rush to publication, thereby boosting our chances of promotion, padding our resumes and feeling, as we give interviews to NPR, like we’re the Oprah of the pocket-protector set? Or do we, instead, pay attention to the person flapping their arms and shouting in the back of the room “You fool- those seats are ASSIGNED.”
Unfortunately, even when an offended, thoughtful scientist -way to go, Stanley Young! – tried to nip this in the bud by writing to the shameful journal that published this study – he even detailed in writing exactly how wrong this is – he only got included in the media coverage (yes, even on NPR!) as a “skeptic.” As opposed to being, say, “right.”
For this, we award to the authors and to NPR our first of 2009, special (Before Opening Gargantuan mouth, Understand Statistics – or, your choice – Babies Obviously Grow Under Stones) BOGUS Award.
What do you think? Will a woman eating an all-cereal diet undo fetal XX chromosomes? What would you over-analyze to find a BOGUS result – maybe the factors that influence the reaction time of gravity? Weigh in with the comments section below!
Related posts:
You obviously get it and I’m glad that you had fun writing this wonderful article. Actually, the journal did help immensely, in getting the authors to provide the data set. The authors resisted our request. It was a long slog to get the data, etc. so it was almost one year to get our comment published. Even now, the original authors are sticking to their story. Others should ask for data sets from studies with implausible claims and check them out. It is a target-rich environment. We estimate that 80-90% of claims coming from medical observational studies are wrong!
Actually, the X and Y chromosome part is anything BUT a simple matter. While the researcher may or may not have found something, this critique is somewhat bogus. If it were simply due to the X and Y chromosomes and randomly receiving such, then probability would tell us that a large population would produce a sex ratio of 50/50. That is, however, not the case. The human population produces more males than females. This cannot be readily explained by simply looking at the chromosomes. SOMETHING else is going on there; what that is is yet to be determined.
Actually there is already a lot of great research about gender proportions in society – starting with the differences between X and Y sperm motility, then the differences in fetal viability (including increased disease rates from Y-linked traits, the most obvious being as hemophilia), then the increased neonatal mortality rate between males (higher) and females, and continuing to the dramatically higher juvenile death rates from injury, accident and trauma. All that prior to the legal age of adulthood, where increased injury/accident/homicide rates continue, as well as increased rates of age-specific diseases such as heart attacks and emphysema (many of whom are related to behaviors such as smoking). The remaining unaccounted-for differences are, in reality, quite tiny, even across the age spectrum, much less at birth.
You seem to be missing the point. Science has YET to account for the ratio of boys to girls being greater than 1 AT BIRTH. Accidents, mortality, heart attacks, smoking, etc after birth have nothing to do with it. The probability of male vs female at conception MATHEMATICALLY is 50/50. Flip a coin a hundred times. You will get approximately 50 heads and 50 tails. Now flip it 1000 times. You will get even closer to 50% heads and 50% tails. Amazing! Now flip it 6 billion times. Your outcome with be almost exactly 50% heads and 50% tails. How come, then, does it not work out that way with conception? There is a significant difference between the mathematical probability and the actual observed outcome that has yet to be explained scientifically. To further complicate things, it has been observed that heterochromatic genders in species such as birds (in which they are female, instead of male like with mammals) are the opposite as far as the sex ratio goes.
Having been a woman all my life (I was born that way) I think the most important point made in Doc Gurley’s original post is the damage this (Bogus) information can to do women all over the world, not just in certain countries. As an OB nurse in the US for many years, I have seen fathers react with dismay (or sometimes anger) when the learn they have another daughter. How can anyone in good conscience publish ANY information without adequate research and verification? Do they not stop to think that this is not just esoteric trivia, but information that can change people’s lives?