I got "800 or so" too. Note that the answer does not depend on the
population size of 126,635 (except to the extent that it is much larger
than 800).
On Mon, 22 Nov 1999, J. Williams wrote:
> Usually, determining the sample size requires an error rate you can accept and
> a selected confidence level. In many types of studies the knowledge of the
> variance is helpful. There are tables in many standard texts which can give
> you the necessary n for your poll. With a 70-30 break, I would estimate you
> will need about 800 or so. You can also use the little formula below to get a
> better picture of the size you might need:
>
> n= Z^2 pq/E^2 where Z = square of the C.I. in standard error units; p=est.
> proportion "for" something; q=est. proportion "against" something; E = sq. of
> maximum allowance for error between the actual prop. and the sample prop.
> Generally, the greater the n, the lower the standard error of the proportion.
>
>
> In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] (KazakOR) wrote:
> >The city in which I live has a population of 126,635. I wish to sample this
> >city and generate poll results with a margin of error of 3.5%. How many people
> >must I contact to arrive at this confidence level?
___________________________________________________________________________
Michael P. Cohen phone 202-219-1917
National Center for Education Statistics fax 202-219-1736
555 New Jersey Avenue NW #402 Internet [EMAIL PROTECTED]
Washington DC 20208-5654 USA
---------------------------------------------------------------------------
