Re: The Promise of Forth

Jonah Thomas wrote:
John Doty <jpd@xxxxxxxxxxxxxxxxxxxxxxx> wrote:
Jonah Thomas wrote:

So you didn't look it up, and you didn't ask your wife or anybody to
explain it to you. You just assumed that you were right.

And as I thought, you didn't fade away because you saw you'd made an
obvious mistake.
What obvious mistake?
You could ask your wife how it works, or consult an
intro-to-epidemiology textbook.

This has an example.
http://www.cdc.gov/EXCITE/classroom/outbreak/steps.htm
Search for "Attack Rates by Items Served at a Church Supper" to find the
relevant part.

Often fewer than half the people who eat a tainted food product get
symptoms. If you look primarily at which items the sick people ate then
typically it will be a statistical blur. But here's a logical approach
that usually works: Since epidemic food poisoning is rare, usually it
will be only one item that is tainted. In an ideal world, one food will
stand out -- nobody got sick unless they ate it. It will turn out that
all of the people who did get sick did eat it, though there may be
several other foods that all of the sick people ate.

Why do you think I don't understand that?


In the real world, there will be a small fraction of people who got sick
and who reported they did not eat the food that made people sick. This
is because people aren't particularly good at remembering what they had
for dinner 3 days ago. In their example 5% of the people who ate the ice
cream had forgotten it by the time they were questioned. To deal with
this they use an extremely simple statistical method to compute relative
risk.

If you don't get one or two items that have a relative risk that stands
out far beyond the others, then the epidemic was probably spread some
other way entirely.

Or has multiple causal factors. That's my wife's sort of epidemiology.

The problem with this kind of reasoning is that you haven't gotten beyond descriptive statistics. That's sloppy thinking, and my wife would tell you so (she complains about this kind of sloppiness in her colleagues all of the time). Your toy example is large enough and clean enough that logic works, but how small a sample can you take before it breaks? What do you do with data contaminated by errors? That's reality. She'd *love* to see data this clean, but it very rarely happens in real life. Real epidemiological practice can't be based on such fables.

But that seems to be all you and Guy have, fables, while I have mostly been giving you real world examples back. Fables are where logic actually works: it finds necessary properties of imaginary objects. Reality is more difficult.

--
John Doty, Noqsi Aerospace, Ltd.
http://www.noqsi.com/
--
History teaches that logical consistency is neither sufficient nor necessary to establish practical, real world truth. Those who attempt to use logic for that purpose are abusing it.
.

Relevant Pages

  • Re: Sickness & Elevated Readings
    ... i have done nothing different except get sick. ... I've had two bouts of what I can reasonably label food poisoning in the ... The first one set in about 8 hours after I ate the ... Gatorade-alike in the fridge and some apple juice ...
    (alt.support.diabetes)
  • Re: The Promise of Forth
    ... Often fewer than half the people who eat a tainted food product get ... If you look primarily at which items the sick people ate then ...
    (comp.lang.forth)
  • Re: Older MREs
    ... contents in a pan with some water, warmed it up and ate it. ... It smelled okay. ... or old canned food it would definately make you sick. ...
    (misc.survivalism)
  • Re: Restaurants that have made you sick
    ... what you ate for lunch didnt cause you to take a ... > sick in 4 hours, you would have been able to taste the rancidness. ... > poisoning isnt coffee, it aint instant. ... now thats fast food... ...
    (alt.vacation.las-vegas)
  • Re: Julies RADP 13 Trip Report - Day 3
    ... I enjoy breakfast in the food court, ... the very tiny trunk of the Yaris, then drove back into the main Epcot ... there is no way that I would have wanted to eat at ... ate sushi for the first time. ...
    (rec.arts.disney.parks)