Tom Mort wrote...
...
What I't like is to have something that the businesses could
enter their past results for a period of time and the number of
samples they proposed to take per month and it would tell them
what the odds (assuming current conditions and variability are
within the range of past) that the average concentraion from
this many samples would be under the average limit.

I was afraid of this. How much stats background do you have?

When, exactly, would businesses sample themselves? When would the wast
water facility or agency sample their output? Does the latter sampl
each business on site?

Your individual businesses are going to need to establish baselin
samples, which means a few days of sampling at specific, relativel
frequent intervals. From that, they may be able to establish relativel
homogeneous periods within the day when they can take one of tw
samples going forward.

Anyway, your sampling procedure should *NOT* involve combinatorics. Yo
can use standard statisitcal techniques to set sample sizes based o
normal distribution assumptions (unless the sample concentration value
show pronounced asymmetry) along with regional overall and specifi
industry descriptive statistics to supplement individual busines
sample values.

This isn't my line of work, so I'm not comfortable enough to go int
the statistical details. However, I'd be very surprised if there wasn'
something available through SIAM that didn't address your needs closel
if not exactly (there are one heck of a lot of engineers
mathematicians and statisticians working in the chemical an
electronics industries, and post-Super Fund, they devote A LOT of tim
and attention to matters of managing discharges from their factories)

--
Message posted from http://www.ExcelForum.com