|
Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
Sampling an exponantially distributed population - how many samples? I have an industrial process which emits a chemical to the air occasionally. There is already a background level of this in the air of say 4ppb. My normal emissions might raise this by 1ppb. Occasionally - a handful of times per year the process might give off a higher than usual emission, raising levels in the surrounding air to say 4000 ppb or more for a period of a few seconds or a few minutes. Ambient measurement is very costly but for various reasons is the only option. To get a statistically valid sample to show a 98% compliance with emission standards how much monitoring do I need to do?
READERS RESPOND: Re: Sampling an exponantially distributed population - how many samples? This appears to be the sort of problem where some sort of cumulative approach is called for. I assume that you are assuming some sort of exponential decay over time of the stuff that is in the air ? What is the rate of decay ? Can you tell when an emission has occured and the size of that emission or do you calculate this from the ambient amount in the air and a certain assumed decay rate ? I you calculate backwards what about weather conditions, wind, inversions, etc, as well as the fact that emissions of similar substances from other sources may influence your measurements.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
|
|||||||||
|