View Single Post
Old 01-10-08, 04:35 PM
  #19  
WalterMitty
Old Enough to Know Better
 
WalterMitty's Avatar
 
Join Date: Mar 2006
Location: OK
Posts: 415
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Likes: 0
Liked 0 Times in 0 Posts
Some of the worst statisticians I have ever met are professionals in the medical, and other "organic" sciences.

The first problem *everybody* has is the understanding that correlation *is not* causation. For example, many more people die at hospitals than at race tracks. This is a provable, verifiable fact.

No thinking person should say that this proves that hospitals are inherently more dangerous than racetracks either for the customers or employees of either enterprise.
However, I know, and have known people that would make just that argument. Perhaps you know people that are deathly afraid of hospitals but would be happy to attend a race with you?

There are problems of adverse selection (more sick and dying people go to hospitals than race tracks) and problems of biased sampling (the fat guy killed in the crash at the race track may have his vital statistics added to an overall morbidity dataset even if weight was clearly not a factor).

If the headlines are to be believed, a significant percentage of us are over weight based on the BMI chart. I have been all my life; even the day after I graduated from ARMY BCT and could do a marathon in full combat gear in less than 12 hours.

So…if the entire population is already predefined as having an “undesirable” characteristic, you can easily prove that the majority of the people dying from heart attacks have this “undesirable” characteristic. This is also true if you die after a fall from a tall building, are killed in a car crash, or are struck by a meteor from outer space.

What gives me the giggles is that the same holds true for everybody that lives past 100. Most are “overweight” and many had “other bad habits”.

The problem this presents to us as a society is that serious study dwindles as soon as a popular consensus is reached based on correlation, which, rather than being a “cause” of the medical/social/physical problems is simply the statistical artifact of a dominant characteristic found in the greater population of data points.

Read about the “End of Medicine” (I apologize in advance for the commercial in the middle of it).

Until our medical professionals learn how to step up, our medical system will remain only one step above witchcraft and continue to get too expensive to buy.

Two Doctors having a conversation: “Well, it looks like we’ve got another stiff on our hands.” says the first. “Yep” replies the second, “but while everybody ever born for the past several thousand years has died (at least once) I’ve got data that indicates about 80% of them die in hospitals”. So, concludes the first, “natural causes complicated by sleeping in hospitals?” “That’s how it looks to me.” they both agreed, and headed off to the golf course for a quick round before dark.
WalterMitty is offline