Revolution <[EMAIL PROTECTED]> wrote in sci.stat.edu:
>Do you think there exist a series of questions, when answered, that
>can determine (with a high degree of probability)
Do you mean with a high degree of _confidence_?
> the gender of the
>respondent? (Assume "Are you male?" is not one of the questions)
What about "Do you have a Y chromosome?" or "Do you have a
[particular anatomical feature]?"
>1) Suppose all questions are answered truthfully.
>
>2) Suppose questions are answered deceptively for the purpose of
>altering the results.
If all you're allowed to do is ask arm's-length questions, and the
respondent may be lying, and you can't verify the answers
independently, then obviously you can't have a high degree of
confidence in any conclusions you draw from the answers.
I sense that the questions I'm answering are not really the ones on
your mind. Can you explain a bit more clearly what you're actually
trying to find out?
--
Stan Brown, Oak Road Systems, Cortland County, New York, USA
http://OakRoadSystems.com/
"People that read tabloids deserve to be lied to."
-- Jerry Seinfeld, on /60 minutes/
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
