3 ms·

In fact, due to sheer randomness there will be small towns with extraordinary (hight or low) cancer rates. Of course there will be great temptation to explain t

In fact, due to sheer randomness there will be small towns with extraordinary (hight or low) cancer rates. Of course there will be great temptation to explain this with nearest plant (or healing springs) etc, where the truth may or may not be there at all.

bsder 7y ago
While something may very well be random chance, that does not mean we shouldn't investigate. The only way to differentiate whether something is causal or random is to examine it.
Dylan16807 7y ago
The more standard deviations from average, the fewer towns you expect. By the time you hit a 50x rate for a reasonably large population, the expected number of towns this bad is zero. And look at the map, the whole area around it is spiked to a lesser extent. It's not random.
- juskrey 7y ago
  Standard deviation does not "work" here. What I mean is that for rare 1:10000 diseases 1 patient per population of 100 people is x100 higher rate comparing to 1 patient per city of 10000 people. Obviously sometimes you have sick people in small villages. The obvious mistake is to calculate the rate per village inhabitant, which is made exactly on the poster.
  - Dylan16807 7y ago
    That's what I was getting at with "reasonably large population". You're right that for a rare disease 10000 is not enough to make standard deviation work. But when you're talking about "cancer", all of it, it's much more than you need for very strong statistical significance.
Spooky23 7y ago
Not really. Almost all cancer clusters can be traced to a source. Sometimes natural like radon gas, but usually industrial. If you’re going to live near a military base, semiconductor manufacturer (especially older ones), or defense/aviation contractor, don’t drink the water. Industrial sites just dumped chemicals in site or in poorly engineered dumps nearby. The military likes to setup burn pits with all sorts of nasty things or just dump stuff as well. Look at maps of brownfield/superfund sites in places like Burlington, MA, around Stanford in California, Bethpage, NY, Endicott, NY and similar sites. Chemical plumes in groundwater and cancer clusters are common themes.
- mauvehaus 7y ago
  Adding to your list: tanneries, machine shops, and dry cleaning outfits. All of which use solvents that are, succinctly, not good for human health. Old rail right-of-ways are also to be looked at with suspicion. Once upon a time trains sometimes left point A full of waste for disposal at a facility at point B and arrived with a lot less waste than they were loaded with (thus saving the folks at point A a lot of money).
  - Spooky23 7y ago
    I forgot about dry cleaners. A friend of a friend worked construction during summer vacation, and when excavating a site that was corner gas station and adjacent laundry, a few workers were overcome by fumes from the gasoline/diesel/dry cleaning muck that was a few feet underground. He quit that day!

bsder 7y ago

While something may very well be random chance, that does not mean we shouldn't investigate. The only way to differentiate whether something is causal or random is to examine it.

Dylan16807 7y ago

The more standard deviations from average, the fewer towns you expect. By the time you hit a 50x rate for a reasonably large population, the expected number of towns this bad is zero. And look at the map, the whole area around it is spiked to a lesser extent. It's not random.

juskrey 7y ago
Standard deviation does not "work" here. What I mean is that for rare 1:10000 diseases 1 patient per population of 100 people is x100 higher rate comparing to 1 patient per city of 10000 people. Obviously sometimes you have sick people in small villages. The obvious mistake is to calculate the rate per village inhabitant, which is made exactly on the poster.
- Dylan16807 7y ago
  That's what I was getting at with "reasonably large population". You're right that for a rare disease 10000 is not enough to make standard deviation work. But when you're talking about "cancer", all of it, it's much more than you need for very strong statistical significance.

Spooky23 7y ago

Not really. Almost all cancer clusters can be traced to a source. Sometimes natural like radon gas, but usually industrial. If you’re going to live near a military base, semiconductor manufacturer (especially older ones), or defense/aviation contractor, don’t drink the water. Industrial sites just dumped chemicals in site or in poorly engineered dumps nearby. The military likes to setup burn pits with all sorts of nasty things or just dump stuff as well. Look at maps of brownfield/superfund sites in places like Burlington, MA, around Stanford in California, Bethpage, NY, Endicott, NY and similar sites. Chemical plumes in groundwater and cancer clusters are common themes.

mauvehaus 7y ago
Adding to your list: tanneries, machine shops, and dry cleaning outfits. All of which use solvents that are, succinctly, not good for human health. Old rail right-of-ways are also to be looked at with suspicion. Once upon a time trains sometimes left point A full of waste for disposal at a facility at point B and arrived with a lot less waste than they were loaded with (thus saving the folks at point A a lot of money).
- Spooky23 7y ago
  I forgot about dry cleaners. A friend of a friend worked construction during summer vacation, and when excavating a site that was corner gas station and adjacent laundry, a few workers were overcome by fumes from the gasoline/diesel/dry cleaning muck that was a few feet underground. He quit that day!