Monday, January 7, 2013

Exactly!


It’s been a while since my last I.N.K. blog posting — 675 days, or 16,192 hours, to be (too) exact. And that’s what I want to write about: meaningless precision. Also referred to as false precision or fake precision, this is a logical/mathematical fallacy that, once noticed, is encountered frequently. 
Overly precise figures are often used to lend questionable data and unproven (or simply bogus) concepts an aura of credibility or scientific validity. Real scientists, as it turns out, are careful not to use figures that are more precise than their data justifies. 

Examples of meaningless precision are easy to find. Sources of nutritional information are always a good bet. An online recipe (the first I looked at, in this context) called for, among other things, one small eggplant and eight slices of bread. It actually sounds pretty tasty — here’s the link: http://www.eatingwell.com/recipes/eggplant_panini.html. I learned that each serving of this dish contains 659 mg of sodium and 353 mg of potassium, quantities apparently unaffected by my definition of “small” or the thickness of my bread slices. I don’t doubt that the intentions of the recipe’s author were good, even if these figures are impossibly specific. 

I’m less sanguine about a spam email I recently received informing me that I could make $12,587 a month working from home. Or the label of a homeopathic remedy for the flu known as Oscillococcinum. Starting out as a tincture of duck heart and liver (don’t ask me), this preparation has been diluted to one part duck to 10400 parts water. At this concentration, one would have to ingest a quantity of Oscillococcinum many times greater than the mass of the universe to be assured of swallowing a single molecule of that unfortunate fowl.

Pointing out this kind of absurdity isn’t particularly clever or original, but it does have something to do with writing children’s nonfiction, and I’ll get to that in a moment. 

But first: normal human body temperature. It is, as most people know, 98.6 degrees Fahrenheit. Except that it’s not. It’s more like 98.2. Carl Wunderlich, a 19th century physician (http://en.wikipedia.org/wiki/Carl_Reinhold_August_Wunderlich), measured healthy body temperatures of  36.5 to 37.5 degrees Centigrade. This range was rounded, quite sensibly, to 370 C, which converts to 98.60 F. The rest is history.

This example is getting close to home, because it illustrates a problem that I confront when writing about the size of things. A recent book included a pangolin, a kind of armored anteater. Pangolins live in Africa, and according to the website of the African Wildlife Foundation (http://www.awf.org/content/wildlife/detail/pangolin) they range in length from 27 to 42 inches. I always include the metric equivalent of imperial units, since most of the rest of the world measures things metrically, and I'd like to give the U.S. a nudge, however small, in that direction. To include a range in both units makes for text that is discouragingly full of numbers — the pangolin is 27 inches to 42 inches (69 centimeters to 107 centimeters) in length — so I usually choose a figure in the middle of the range, in this case 36 inches — three feet. It’s not quite in the middle, but it does have the advantage of sounding rounded off and not too precise. The trouble arises when the metric conversion — 36 inches to 91 centimeters — is made.  The metric figure is too precise. It implies that pangolins are not variable in length, but exactly 91 centimeters long. I could round off the metric units, but that would yield a too-precise imperial unit figure. Or I could write “36 inches (about 90 centimeters),” except that those “abouts” get annoying when there are dozens of creatures described. Or I could use a rounded off metric figure that would not be an accurate conversion. But, to quote Richard Nixon, that would be wrong.

I recognize that this issue doesn’t rank highly on the list of problems facing humanity. Still, it bugs me. I share it with I.N.K. readers in hopes that someone will suggest a workaround. Or maybe just to get it off my chest.

3 comments:

Jim Murphy said...

Getting things off one's chest is always a good thing. Okay, not always, but it can have certain health benefits. And while I don't usually have to deal with the issues you describe here, it does come up every so often. I try to explain the variables in the text, but if it gets clumsy I put a brief explanation in my backmatter. By the way, like just about everyone else, I grew up understanding that the healthy body temperature was 98.6. When I was eight or so, I mentioned this to my doctor and he said the notion of a fixed body temp number was nonesense. Different people, different bodies, different normal temperatures, he explained. Then he added a really valuable piece of advice, "Don't believe everything you hear even if everybody says it." Maybe more kids need to hear this.

Vicki Cobb said...

Welcome back to I.N.K., Steve! Along with meaningless precision another pet peeve is most people's lack of understanding of acceptable risk. Over the weekend I heard a story of a household product that most people have that causes about 800 [note I rounded off] deaths a year. "Horrors!" some people say. "Get it off the market, no death from a household product is acceptable." What's the product? Natural gas.

rglaser said...

As a nonfiction editor I run into this conversion problem ALL the time. It is nice to hear someone else has thought through this exact same logical argument of when to round or be precise. I find the best ways around it depend on the age level. For the very young I skip the measurements, for lower elementary I like making comparisons--is a pangolin at three feet long about the size of a raccoon? Or a little bigger--maybe a black lab? For upper elementary you can add something visual--scale diagrams or rulers showing the measurements next to a silhouette.