Steve Erbach
erbachs at gmail.com
Tue Jan 18 19:56:48 CST 2005
Stuart,
This is just too much. Using the Sample StDev for a sample size of two.
So what's the margin of error, eh?
Steve Erbach
Neenah, WI
On Wed, 19 Jan 2005 08:14:04 +1000, Stuart McLachlan
<stuart at lexacorp.com.pg> wrote:
> On 18 Jan 2005 at 16:03, pedro at plex.nl wrote:
> > But now. You gave me an example of the calculation {So if the points are
> > 8.4 and 9.3, the average is 8.85. The "standard deviation" would be 9.3
> > - 8.85 = 0.45.}. Steve your calculation is the stdevp. This calculation
> > is for a whole population. But, i understand, that it doesn't matter.
> > The calculation can't be done in access because it needs a field name.
> > It it possible to write the two values each time to a temp field, do the
> > calculations and write them then to the result table.
> >
>
> SD of a population uses Sqr(..../N), SD of a sample uses Sqr(...../(N-1))
>
> Since there are only two data points (N=2), the difference between StdDevP
> and StdDevS is SQR(..../2) v SQR(......) Since Sqr(2) = 1.414, you end
> up with StDevS = StdDevP *1.414 or
>
> StDevS = Abs(A1-B1)/2*1.414 = Abs(A1-B1)*0.707
>
> --
> Stuart
>