Date Posted:12:17:25 03/28/05 Mon Author: Mike K. Subject: Statistical evidence In reply to:
Joe
's message, "Re: Cult Apologist Lobby" on 22:15:23 03/25/05 Fri
"If you have out of a hundred or two hundred thousand Jehovah’s Witnesses, 10 or 15 who say something like that
Fact is, any statistician would treat this very, VERY differently.
One in a million is an odd occurrence.
Two may be variance, but 10 already... hey, that's not ten in a million, but one in hundred thousand.
And since we're talking like 1000 in 500000, we're already talking in the area of percentage.
And it's not like their testimony is "white noise", but forms a consistent pattern.
During my investigation of medical data, it was for those "one in ten thousand" things I was to look for, NOT for the 9999 that were not special.
Medicine actually is a good example:
if you have a body, and 9999 times, your heart frequency and blood pressure show no anomaly, it may indicate that you're healthy. BUT: no doctor would be such a fool to ignore the one instance when heart frequency and blood pressure exceeded normal bounds.
To diagnose sickness, this one "abnormal" instance is MUCH more important than all other instances.
In the investigations I made, I even went to "filter out" all those 99.99% of normal occasions, and my job was to locate patterns in the remaining 0.001% - yes, you have VERY little raw data, but it's your duty to find the patterns there.
And it's not like you can't find patterns in the report of dropouts.
So, be the good doctor and filter the "normal cases" out.
