At 04:07 PM 1/13/2003, Rich Ulrich wrote:
- a not-so-quick comment on the document and on power-estimation.there are two issues involved:
On 8 Jan 2003 18:14:53 -0800, [EMAIL PROTECTED] (Dennis Roberts) wrote:
> I worked on a small handout for my stat class about the margin of error
> (like the +/- 3% used in polls) . While I fully acknowledge that I got the
> idea from Moore and McCabe's book "The Practice of Statistics", I put it in
> the context of examples used in my own classes.
1. what level of confidence do you want to work with ... ie, in the long run, what % of the CIs will encapsulate the parameter (forget precision)
2. what level of accuracy do you want to work with ... ie, how CLOSE do you want your point estimate to be away from the parameter (forget level of confidence)
these are not unrelated issues ... that is, one should think about BOTH of these matters but, the handout ONLY dealt with #2 ... not #1
the purpose of the handout was to try to counteract what we tend to do, that is, build 95% CIs blindly ... without much/any concern for precision ... we just live with the level of precision that happens to turn up ... which, according to the anti martha stewart forces ... is NOT a good thing
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
