Date: |
Mon, 16 Sep 1996 19:54:10 -0700 |
Subject: |
|
From: |
|
Parts/Attachments: |
|
|
<<Disclaimer: Verify this information before applying it to your situation.>>
Hi All,
In discussion with a netfriend, we were considering the analysis posted by a
new member to the list, Melissa. I'm afraid I don't remember anything
but her first name. Her analysis assumed, if I remember correctly, that
the error factor was calculated for the entire group tested. I suspect
that it is calculated for the group that demonstrates a positive test
result. In a hypothetical construct, if one tests 10,000 healthy people for
Ema antibodies, one should get about 40 positives. If the test is 98%
specific, then 39.2 should be true positives, and .8 should be false
positive. That would mean that in testing 10,000 people, only one should
show a false positive.
I would welcome further instruction in this area, as I have assiduously
avoided statistics in recent years, and am now paying the price. I am not
very sure of what I am saying.
Best Wishes,
Ron Hoggan
|
|
|