March 04, 2010

Cancer and statistics

If you recall the rekindled controversy about cancer screening a few months back, here is a detailed explanation about why the issue is a lot more complex and counterintuitive than the sound-bite reports or staged debates between lobbyists make it out to be.

As Dr. Kramer vividly illustrates, the statistical tools used to evaluated the "success" of screening methodologies are extremely susceptible to unintended manipulation. The greater ability to detect health problems often does not correlate with a greater ability to decrease mortality.

Thus the results of screening can be as harmful as they are well-intentioned. This Wednesday (3/3/2010), the American Cancer Society followed exactly the advice Dr. Kramer gives (11/25/2008), saying that "Routine [prostrate] screening isn't recommended for most men."

This again affirms my belief that in high school math classes, teaching statistics should take precedence over teaching anything beyond basic arithmetic. Related thoughts here by Car Talk guy (and MIT grad) Tom Magliozzi.

Labels: , ,

# posted by Anonymous Dan
3/08/2010 6:33 AM   

A course on statistical theory would be more valuable than one on statistical methods. You teach the bell curve to students and suddenly everything in the world is a Guassian distribution, even when it is not.

If I had a vote I would require every high school student to have a year long course on Opportunity Cost. Let's call it the "No free lunch" seminar. Better yet have a one semester course on this topic and another semester course on how to rob Peter to pay Paul.

If we fail to educate the rising generation on the merits of Austrian economics let's at least make sure they are well prepared to function as DC lobbyists.