NUMBERS
Long ago, a group of concerned scientists released a statistical study of a particular neighborhood. There in Love Canal there were health issues galore - and most probably related to the toxic waste stored in the landfill atop of which Love Canal had been built.
The study was interesting. The researchers had compared rates of various health issues to a generally similar community not too far away, but not part of Love Canal. The best way to do such a study would be to locate each health issue within the whole of each community: study and control. the researchers did not do so, they sampled each neighborhood.
Sampling is okay too. Run the comps to determine how many samples will be required to arrive at a satisfactory level of accuracy with the population quantities of each community. Calculate how many circles of whatever arbitrary size will be needed to include that number of sampled persons. Plot randomly chosen points on a map of each area, the number of circles needed to attain accuracy. Generate the circles around these points, then go out and interview (having to show medical records would be a good thing).
this is not how it was done. the researchers did an intital and quick plot of known addresses with helth issues in Love Canal. then they drew circles around the resulting clumps (randomness is not perfectly 101010101010) and sampled.
Not to say that i would live in Love Canal. Simply to say that the numerical analysis was flawed.
How many of the numbers given to us by the experts are bogus, subject to poor interpretation, or skewed?
_________________
Who is John Galt?
Still Moofy after all these years
It is by will alone that I set my mind in motion
cynicism occurs immediately upon pressing your brain's start button
11 Feb 2007, 9:06 am
