Every day we're bombarded by new studies making outrageous health claims. Occasionally they claim to have discovered that something is wonderfully good for us, and we should change our lifestyle and/or immediately buy something so we can run fast, jump high, and live forever. More often they claim something is horribly horribly bad for us, so we should change our lifestyle, and/or immediately buy the cure, so we can run fast, jump high, and live forever.
Sturgeon's law, which states "Ninety percent of everything is crud" is about 5% too low when applied to statistical studies. If we assume they're all lies, we'll be right most of the time. But it's more fun, and much more satisfying, to debunk them. Extensive debunking usually requires time, patience, expertise, and knowledge of statistics, but fortunately there's a very handy shortcut. We simply ask if the conclusion is reasonable.
As an example, take James Repace's claim that 300 mile per hour winds are necessary to clear second hand smoke from a room. If we had an extensive background in physics and toxicology, we could determine the accuracy of that statement with pages of calculations involving air flow, room size, and the toxicity of second hand smoke. The result would be time consuming and impressive to other people with the same expertise, but no one else would understand it or give a damn. Applying a bit of common sense to the claim is a far simpler approach.
Imagine sitting in a living room with Aunt Irene, an antique collector who has just finished smoking a cigarette. You're in the middle of hurricane Ivan, and the windows are open. According to James The Expert, those 150 MPH winds, while capable of blowing Irene's roll top desk through the wall, (followed, no doubt, by Aunt Irene, and you), they'd be far too weak to clear the air of the wisps of smoke lingering from her cigarette. You'd have to double the wind speed to do that.
Is that reasonable?
In his most recently published study, Repace claims that the air in a smoky bar is more dangerous than the air filled with vehicle exhaust. Ignoring trivial stuff like carbon monoxide, he instead concentrates on soot, calling it PPH to make it seem less trivial. His measurements are in nanograms billionths of a gram. He claims to have measured the PPH in bars at 134 nanograms, compared to 18 nanograms at the side of the road during rush hour.
First we need to ask if his measurements can be trusted. It's difficult to measure such small amounts accurately in an uncontroled environment. Add to this the fact that crusaders are usually untrustworthy. Add to this the fact that Repace personally gave me fraudulent evidence on a different issue just a few months ago, and we can pretty easily dismiss anything this clown ever spews. But let's ignore all that for a moment. Let's pretend we're dealing with an honest researcher and his measurements are accurate.
A nanogram is one billionth of a gram. A grain of salt weighs about 120,000 nanograms. Is it reasonable that 134 billionths of a gram of soot would be dangerous?
We can also apply the reasonableness test to the researcher of any given study. Who is paying them? What is their history? Google makes such research fairly easy.
In the case of Repace, we're dealing with someone who helped author the 1993 EPA study on second hand smoke, the most thoroughly debunked study in the history of junk science. Much of the funding for his projects comes from the Robert Wood Johnson Foundation, an uber nanny that has generated even more junk science than the EPA in their holy quest to stamp out tobacco and alcohol. If we follow the money, we learn that RWJF is funded primarily by the dividends on six billion (yes, billion) dollars worth of Johnson and Johnson stock. Every time someone quits smoking and buys one of Johnson and Johnson's outrageously profitable stop smoking aids (nicotine patches, nicotine gum, nicotine inhalers, nicotine breakfast cereal, etc.) it literally puts money in their pocket, which they then use to hire "experts" like Repace. That makes Repace a shill for the pharmaceutical industry. Shouldn't we consider that whenever he makes a proclamation?
Let's try the reasonableness test on another recent "study." Economist Frank Sloan claims smoking costs society nearly $40 a pack. Now, we could point lots of other studies that found the cost is either even, or a net savings. This is because smokers die sooner, collecting fewer retirement benefits and generating a lower lifetime health bill. Also, over their lifetime they pay tens of thousands of dollars more in taxes than anyone else in their income bracket. Frank claims to have included those facts, but managed to inflate his numbers by declaring that each year of lost life was worth $100,000. Where did he get that number? To quote my great grandmother: "He pulled that one out of his ass." He then conjures up a $5.44 for the cost of second hand smoke, another nanny ass number, then for good measure pulls out another nebulous $1.44 for society as a whole.
In his original press release he never metioned that a significant portion of his funding came from wait for it The Robert Wood Johnson Foundation. Are you starting to notice a pattern here?
I haven't seen the original study, but I suspect it would take a degree in economics to debunk it directly. So let's take the quicker approach asking Are The Numbers Reasonable? For this one we have to do a little math.
First, we need to know how many packs of cigarettes are sold in the US each year. That wasn't as easy as I thought it would be. Googeling the question didn't work too well, but I was able to find several sites that talked about per capita consumption of individual cigarettes, and most offered similar numbers. This one from the United Nations claims the US consumes of 2,092 cigarettes person per year. First, we'll test the reasonableness of that number.
Only a quarter of the population smokes, so we multiply that number by four to get the annual consumption of a single smoker. It comes out to 8,368, which breaks down to 418 packs, which in turn breaks down to 1.15 packs a day. That seems a little low, but it's not unreasonable, so we'll go with it. (One advantage of this approach is that we're don't need precise numbers. Reasonable approximations are enough to draw general conclusions, especially if our calculations err on the side of caution.)
Multiplying 418 packs times 70,000,000 smokers (a little less than a quarter of the US population, again erring on the side of caution) gives us 29,288,000,000 packs of cigarettes sold every year. Multiply that by $40 a pack, and we come up with a cost of $1,171,520,000,000. That's right folks, according to this study, cigarette smoking costs the US 1.17 trillion dollars every year!
The Gross National Product last year was about $11 trillion dollars. That means that smoking costs us more than 10% of everything. Wow, what an impressive number! What a shocking conclusion!
What an amazing pile of bullshit.
The next time you read a study claiming that something is amazingly good or amazingly bad for you, you can take the time to find the original abstract, examine the sample size and methodology and check the backgrounds of the "experts" to determine if the study is valid or not. Or, you can take the simpler approach, and simply ask: Does this sound reasonable?
Learn the basics of staticts in two (fairly) easy lessons, Epidemiology 101, and Epidemiology 102.
© 2005 Dave Hitt