[EMAIL PROTECTED] (dennis roberts) wrote in 
[EMAIL PROTECTED]:">news:[EMAIL PROTECTED]:

> we hardly EVER talk about is the concept of 
> "margin of error" ... that is, trying to PRE plan our accuracy (that is, 
> have our sample statistic be within m of the parameter ... ) through the 
> control of sample size n
> 
> why is this?
> 
Perhaps because people generally do hate the answer when they do the power 
analysis? 
<whine-mode-on> Aw, do I really need 300 subjects? Can't I really do it 
with 60? That's too much work. My friend says I only need to specify the 
width of the confidence interval. <end-whine>

David Winsemius.
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to