[EMAIL PROTECTED] wrote:
>
> Robert J. MacG. Dawson <[EMAIL PROTECTED]> wrote:
>
> : If I attempt to survey 1000 people and 950 answer, of whom 600 give a
> : positive response, I can consider the extremes of 650 in 1000 and 600
> : in 1000, create confidence intervals, and say (eg) that _in_any_case_
>
> Don't follow this. Why 650 in 1000, maybe 950 in 1000?
If 50 members of a true random sample don't answer, and 600 answer in
the affirmative, the extreme possibilities are:
(0) all nonrespondents _would_ have answered in the affirmative if they
had answered, so that a complete response would have been 650 out of
1000;
(1) all nonrespondents _would_ have answered in the negative if they
had answered, so that a complete response would have been 600 out of
1000.
Any other possible outcome would be betwen these extremes, so that if
analyzing both these extremes gives the same result [subject to cewtain
obvious monotonicity assumptions] all outcomes would do so.
-Robert Dawson
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
