Mime-Version: |
1.0 |
Sender: |
|
Subject: |
|
From: |
|
Date: |
Fri, 10 Apr 2009 08:17:52 -0400 |
Content-Type: |
text/plain |
Content-Transfer-Encoding: |
quoted-printable |
Reply-To: |
|
Parts/Attachments: |
|
|
Dear all:
I'm going to caution once again from making leaps from associations to causality. You
canNOT claim that because two events occurred at the same time, or one event preceded
the other, that the one event caused the other. This is "ecologic fallacy". As I've said
before, this is what Nora Ephron used to claim that breastfeeding caused allergies
because both are increasing.
Furthermore, what happens in chemical test tube does not always predict what happens in
the petri dish, what happens in a petri dish doesn't always predict what happens in a
multicellular organism, what happens in avian species doesn't always predict what
happens in a mammal, what happens in a mouse doesn't always predict what happens in
a monkey, what happens in a monkey doesn't always predict what happens in a human.
There are any number of failed interventions that worked just fine in monkeys that failed
in humans. Some notable disasters occurred in a Lord Voldemort topic because what
happened in monkeys was not what happened in humans.
There are REAL problems with cow's milk for which there is good evidence of an effect
through careful science using methodologies that eliminate "ecologic fallacy".
Going down the path of speculation based on a few tenuous leaps of biochemistry on the
one hand and a very MULTiCAUSAL behavior does NOT prove causality. More
importantly, it does not inform our interventions.
A MULTICAUSAL behavior that is often "DIAGNOSED" as oversupply (when sometimes it
isn't oversupply) deserves far more in depth examination than leaps of speculation.
Speculation IS needed to:
a) define a hypothesis
b) develop a potential intervention
c) design a study
However, any study is useless and sometimes even harmful if you don't have:
1) Careful and tight definitions of the problem (not just sloppy labels like oversupply)
2) Control groups who either do not have the condition or do not receive the treatment
And for multicausal conditions, I have to say that one cannot look at the situation in a
simple linear fashion. You also need to look at what is known as "effect modification"
which often involves using statistical techniques to find interactions. This was done
beautifully in a study called "Mothers Milk and Sewage" which debunked the formula
industries claims that the much higher infant mortality rates in developing countries from
formula were merely due to sloppy science. It turned out that the much higher mortality
rates were due to a strong interaction between lack of latrines (hence greater
contamination) and formula. Piped water really did not have much of an effect.
Much more is involved in designing tight studies --- but I would say that a mental
exercise of potential connections is never enough to provide proof.
Even in qualitative research there are many many steps to constantly check and recheck
assumptions, develop alternate explanations and come to validity.
Best, Susan Burger
***********************************************
Archives: http://community.lsoft.com/archives/LACTNET.html
To reach list owners: [log in to unmask]
Mail all list management commands to: [log in to unmask]
COMMANDS:
1. To temporarily stop your subscription write in the body of an email: set lactnet nomail
2. To start it again: set lactnet mail
3. To unsubscribe: unsubscribe lactnet
4. To get a comprehensive list of rules and directions: get lactnet welcome
|
|
|