Thursday, July 09, 2009

Bad Science, Part II

If you are a lottery player, this is your lucky day!!!

For a *very* limited time, I am prepared to share with you the chance to become INCREDIBLY RICH. I am prepared to share with you the mathematically proven way to DOUBLE, TRIPLE, even QUADRUPLE your odds of HITTING IT BIG in the lottery over those poor souls who buy just one ticket.


Sound like a pitch you've heard, in various forms, on late-night TV, or in various pop-up ads at your favorite websites? If you have, you're being sold a very crafty pitch via the use of what's known as a change in relative risk.

If you were to send me that $49.95, you know what I'd send you in return? A very fancy, foil-printed, three-fold brochure instructing you as follows:

* To DOUBLE your odds of winning over buying one ticket, buy two.
* To TRIPLE your odds of winning over buying one ticket, buy three.
* To QUADRUPLE your odds of winning over buying one ticket, buy four.

You'd probably be be pretty hacked off, but you'd also be lighter in your wallet that $49.95, because I had sent you *exactly* what I promised. Nothing in those instructions is false; if you buy two lottery tickets, your chances of winning are doubled over buying just one, and so one. Statistically and mathematically rock solid.

But if you know that your absolute odds with one ticket are somewhere in the vicinity of 1 in 100 million (depending upon the particular lottery), doubling your odds probably doesn't make you feel much better, does it? That's the unsold dark side of relative risk; what sounds like a great deal expressed in relative terms isn't nearly so appealing in absolute terms.

The art of selling changes in relative risk over absolute risk has risen to high form, practiced primarily by pharmaceutical companies, agenda-based political organizations, and your garden variety late-night hucksters.

The worst offenders in my view are pharmaceutical companies, who are glad to tell you that "prescription-only Miraculex" will cut your risk of catching Throbbing Globulitis by 40%!!!, but just not bother to tell you your risk of catching it in the first place is only 1 in 1,000,000, meaning your risk will drop to 1 in 600,000. Yet millions of people fall prey to precisely this kind of advertising in the hopes that throngs of uneducated consumers will rush to their doctors, demanding prescriptions for this new wonder drug.

Journalists are the next worst offenders, because relative risk stories are fundamentally more sensational and often help "sell" a story when tied to some particularly trendy cause. An example of relative-risk that doesn't, on the surface, appear to be selling something, appears here, with a story about the discovery of a 26% reduction in risk for breast cancer among women who experience migraines headaches.

One of the worst pharmaceutical offenders is undoubtedly the Sanofi-Pasteur, manufacturers of the anti-meningococal vaccine Menactra. Ads for the vaccine paint this terifying picture of happy older children stricken dead with meningococcal disease (which is not, but can lead to, meningitis). What they tell you amidst the fear in small print at the bottom of the screen is that only 1,000 to 2,600 cases of menigococcal disease occur in the US annually, meaning your absolute risk of contracting meningococcal disease is no better than 1:115,000 (assuming a US population of just over 300 million people). Yet Sanofi-Pasteur is delighted to frighten you into believing you or your kids are "one sip" away from death without Menactra.

The point isn't to shoot down pharmaceutical companies or journalists, but to demand action. When news outlets post stories involving changes in relative risk, such changes should also includes changes in absolute risk as well, in the spirit of complete information. Complete information benefits all of us, so we can make informed decisions, not ones based on hysteria, propaganda, and bad science.

Especially bad science that stretches truth to the point of unrecognizability.

No comments: