Need to expand subsample of species lengths

**rshanaghan** · 12-19-2012, 12:13 PM

Not sure if this issue is tough, or if I'm missing something obvious. I work on a scientific fishing vessel, and its my job to obtain lengths for all species we catch. For the particularly dense species, I only measure subsamples of 30 individuals. The problem is, to analyze this data, I need to expand this subsample, to represent the whole catch. I know the size of the whole catch, but not all the lengths. My question, is whether there is a way to use my data set of 30 entries to create the entire data set, which in some cases exceeds 500 individuals. Thanks for any help you can provide!

Ryan

[SOLVED]

**Melvinrobb** · 12-19-2012, 12:33 PM

Can you provide some numbers or examples. Not sure what you mean by "create the entire data set".

**FDibbins** · 12-19-2012, 12:42 PM

this is a stats question, and my stats knowledge has rusted away. google extend sample size to population size

**rshanaghan** · 12-19-2012, 03:49 PM

Melvinrobb - For an example, the lengths (in cm) for a subsample of cancer crabs are
14
9
10
8
8
7
8
6
9
8
8
10
8
9
8
8
9
7
12
10
6
6
7
8
5
8
9
9
9
6
7
That's 30 individuals, but we caught 175 that day. I need to have a data set which represents all 175 individuals.

**Melvinrobb** · 12-19-2012, 03:57 PM

I believe what you want is the NORMINV function. Calculate the MEAN and the STDEV of the actual lengths you have (easily done using the functions). then in a new column enter this formula:

Formula:

Please Login or Register  to view this content.

Copy the formula down 175 lines. This will give you realistic numbers based on the subset you have.
the RAND function returns a random number between 0 and 1. By using this function, if the random number is 0.5, the formula will return the mean. 0.68 will one standard deviation above the mean, 0.95 equal two (i believe...), etc...

**rshanaghan** · 12-20-2012, 11:54 AM

I tried that formula, and that definitely is what I need, but it won't calculate values for it. It keeps giving me a "Volatile" response for the RAND() function inside the NORM.INV function. This causes it to give a #NUM! error.