nos-trum. pronunciation: \nos'-trum\. noun. Etymology: Latin, neuter of noster our, ours.
1. a medicine of secret composition recommended by its preparer but usually without scientific proof of its effectiveness.
2. a usually questionable remedy or scheme.
See here for more discussion.

Thursday, February 11, 2010

Random Thoughts on Randomness, or the Fallacy of Uniform Randomness.

There's another situation--in California, of course--where people are getting very confused about detecting clusters of disease and what a random pattern of disease distribution looks like.

Here's the story.  The residents of a small town of 1400 (Kellerman, CA) noted that there had been several birth defects and three infant deaths over a two year period.  The town is close to the "largest hazardous waste landfill in the West." 

Ipso facto, right?  Gotta be the reason...

So, the Terminator ordered a study by the EPA.  They came in, took a look at the cases, and compared what was occurring in other similar towns that didn't have a hazardous landfill nearby.  The investigators found that "the rate of birth defects in Kettleman City from 1987 to 2008 was only slightly higher than that of neighboring towns and in surrounding counties."  Due to the low numbers, I suspect that this means there was no statistical significance.  Further, there was no pattern to the hazards that could explain the cases that were alleged to be due to the landfill.   Let me say this differently:  specific toxic chemicals cause specific defects.  You don't see a toxic compound that can cause any kind of problem.  I once had a patient who alleged that the anthrax vaccine made him throw-up, and his regurg glowed in the dark (I'm not kidding).

No.  Sorry.  There's no biological plausability for flourescent puke.  The truth goes like this:  vinyl chloride causes liver cancer--it doesn't cause fibromyalgia, arthritis, pneumonia, or the heartbreak of psoriasis.  Whenever you see a cluster of cases that show a broad range of diseases and symptoms, you have to match them up to one or more specific hazards that can cause those symptoms.   They couldn't do that in Kellerman.

The reaction to the EPA's report was typical:  "that can't be right."  There's more to the story here, and I refer you to the link.  They did find arsenic in the water supply, but...I hate to tell you this...if your water comes from well water, that's not unusual.  See the article if you need more.

The fallacy of reasoning that's involved here turns upon how we tend to think about random events.  Many of us think that if bad stuff is sort of evenly spread over the landscape, then it's random.  But that's not random, that's uniform.  A truly random landscape will show gaps, and spots of increased bad stuff, which actually confirms that there is a random process at work.

This is called the clustering illusion, or the Texas Sharpshooter Fallacy (click the link; this article is one of my favorites).

Now, nothing here means that there isn't a public health problem at work.  It just aint what the residents think it is...so far.

Doc D

Opinions are entirely my own.

No comments:

Post a Comment

Followers

What I'm Reading - Updated 3 May

Blog Archive