On 1 Mar 2004 13:27:17 -0800, Euh <[EMAIL PROTECTED]> wrote:

> I took 3 differents samples from the flask and, for each sample, I did 20 
> counts under the microscope (= at 20 different

I've considered all counts from any samples as equal cases.
Maybe we need consider only means of counts as cases, and
then calculate a mean of the cases (samples).

The formula for required number of cases is
n = (t^2*s^2)/(d^2)

n - number of cases
t - Student factor
s^2 - dispersion
d - maximum allowable error

You need to define these parametrs:
1) confidence probability of 0.95 is quite enough
for biological investigations; so, choose p=0.95;
t(0,95)=1,96;

2) maximum error; assume +-10 particles;

3) calculate dispersion using your data
(according the data you gave bellow your posting
s=34 for all 60 cases).

So, n=(3,84*1156)/100=44

I offer to calculate 45 counts per sample;
and use 5-10 samples
.
.
=================================================================
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