From: Harvey Newstrom (newstrom@newstaffinc.com)
Date: Thu Dec 02 1999 - 13:01:32 MST
Robert J. Bradbury <bradbury@www.aeiveos.com> wrote on Thursday, December
02, 1999 4:41 am,
> So, the real problem is not with the term "natural" or "atypical" but
> with the term "homosexual" -- it decribes a gross behavioral
characteristic
> that might have multiple underlying, but distinctly different, physical
> (or environmental) causes.
The problem with this term has plagued biological statistics for a long
time. Kinsey claimed that 10% of the population was "gay". But what was
actually defined was a whole range of sexualities: Totally gay, mostly gay,
tending toward gay, bisexual, tending toward straight, mostly straight, and
totally straight. Which of these count as "gay" and which do not?
At the risk of opening another can of worms, let me give this classic story
of statistical error:
1. (Assume these numbers for the sake of the example. I know that the
exact numbers are very much disputed.)
2. Gays represent 10% of the general population
3. Gays represent 75% of the HIV-positive population.
4. Faulty conclusion: HIV is a "gay" disease. Gays are found more often
in the HIV population than would be expected by random chance.
5. Error: The term "gay" used in statement 2 is not the same definition as
the term "gay" used in statement 3. We are comparing apples and oranges.
6. Explanation: The belief that gays represent 10% of the population comes
from the Kinsey study. They only counted mostly-gay people as "gay". They
*excluded* people tending toward gay, bisexual, tending toward straight, and
mostly straight. The belief that gays represent 75% of the HIV-positive
population comes from the CDC in Atlanta. They counted anyone who had ever
had a homosexual experience in their entire life as "gay". They *included*
people tending toward gay, bisexual, tending toward straight, and mostly
straight.
The intermediate group of people considering themselves straight but who had
at least one homosexual experience were *excluded* in one definition while
being *included* in the other definition.
7. Corrected interpretation:
10% of the general population is totally gay or mostly gay. 10% of the
HIV-positive population is totally gay or mostly gay.
65% of the general population is tending toward gay, bisexual, tending
toward straight or mostly straight. 65% of the HIV-positive population is
tending toward gay, bisexual, tending toward straight or mostly straight.
The two groups above can be combined to show that 75% of the general
population has had a homosexual experience, and 75% of the HIV-positive
population has had a homosexual experience.
That leaves 25% of the general population who have never had a single
homosexual experience, and 25% of the HIV-positive population who have
never had a single homosexual experience.
8. Corrected conclusion: The numbers found in the HIV-positive population
are found in the same ratios as the general population. This is what would
be expected by random chance. There is no statistical link between sexual
orientation and HIV.
I know somebody will dispute the exact numbers given here. I don't care
about the exact quibblings. The point is that the method of counting gays
in the general population undercounted compared with the method of counting
gays in the HIV-positive population which overcounted. The statistics were
flawed by comparing apples to oranges and comparing results.
(The CDC has since revamped its definitions and keeps better records for
people in different groups, and for people in multiple groups, such as gay,
drug user, prostitute, blood transfusion recipients.)
-- Harvey Newstrom <mailto://newstrom@newstaffinc.com> <http://harveynewstrom.com> Author, Consultant, Engineer, Legal Hacker, Researcher, Scientist. ----- Original Message -----
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:05:55 MST