Question/confirmation on randomization

[Dave R.]

Hi,

at work we are selecting random numbers. In the past, the process used has
been to use the TOOLS>Data Analysis, to randomly pick x number of numbers
from a list of patient ID codes.

The problem with this, is in our random list we want a maximum of one
occurence of any patient ID. So the problem in the past with using the
Tools>Data analysis route was that there would frequently be duplicates,
requiring re-pulls that exclude numbers we've already picked.

So I have suggested using the method of randomizing whereby we'd use RAND()
in a helper column next to our patient IDs, then sort by the random numbers,
and copy/paste the top x patient IDs to a separate location, and thereby
have our random sample of unique patient IDs.

The issue that comes up about this is that it is not the SAME method as we
have used in the past, but in my eyes we are still creating 2 random lists
of numbers, with the same program. My question for everyone is if they think
this is a comparable method (i.e. taking into account randomization methods
within XL 2000) that is simply more efficient.

Thanks for any ideas

 

[Guest]

Dave
Random is random. Your sample will still be just as random with the new technique as with the old

I have used your new technique on simple dice and card game sheets

Good Luck
Mark Graesse
(e-mail address removed)

----- Dave R. wrote: ----

Hi

at work we are selecting random numbers. In the past, the process used ha
been to use the TOOLS>Data Analysis, to randomly pick x number of number
from a list of patient ID codes

The problem with this, is in our random list we want a maximum of on
occurence of any patient ID. So the problem in the past with using th
Tools>Data analysis route was that there would frequently be duplicates
requiring re-pulls that exclude numbers we've already picked

So I have suggested using the method of randomizing whereby we'd use RAND(
in a helper column next to our patient IDs, then sort by the random numbers
and copy/paste the top x patient IDs to a separate location, and thereb
have our random sample of unique patient IDs

The issue that comes up about this is that it is not the SAME method as w
have used in the past, but in my eyes we are still creating 2 random list
of numbers, with the same program. My question for everyone is if they thin
this is a comparable method (i.e. taking into account randomization method
within XL 2000) that is simply more efficient

Thanks for any idea