Stats Help!!!! Need it asap!

[CanaNG]

I'm sure this is in the wrong forum but I need help asap in determining the minimum sample size required to have about an 80 or 90% confidence that the mean value is representative. Unfortunately we don't have anyone versed in statistics on the team.

Here's a little back ground. I'm taking the spinner tests of all our commingled sand&coal as well as our CBM P&F control wells and determining the relative contributions of each zone on a well by well basis. Then I'm comparing the two type of wells to show that the flow data from the comm. sand & coal wells are interchangable with that from the CBM P&Fs. Anyways, I need to figure how to determine what the minimum number of wells needed to ensure within a 80-90% confidence that the mean is representative of flow from that zone. There are 335 CBM P&F wells and 600 commingled wells.

Note that not all wells have every zone perforated, therefore I assume I would have a different population for each zone. Some zones only have 30 of 335 wells perforated. Also, if I use all the data I have, I should have no margin of error, correct? This would be because my sample is my population?

If anyone has some formulas or calculators (excel, interntet) that would be great. I found this one site:

http://www.raosoft.com/samplesize.html

but I don't know if that was designed for queries that have a range of values as opposed to a yes/no answer.

If you need more info please let me know. Need this sorted out by the end of the day. Got to love time crunches.

Thanks

 

[TEXWS6]

Are you going to compare the mean of the data set to a standard or given data point based off of historical data?

 

[TEXWS6]

I'm up all night watching a frac, so I'll go ahead and chime in.

1) Determine if your data is normal or not.

2) Once you determine normality, use appropriate hypothesis testing of data.

Hypothesis testing normal data will tell you if all the means are/not equal. This is just my two cents and will probably be the simplest. You could run a 1-2 sample tTest, but you must have a standard to compare against for the 1 sample tTest or compare two samples against each other on the 2 sample tTest.