M
MACRE0
I have one solution, but it's rather cumbersome. I could create a lis
10,000 long with each unique item maintaining its relative percentag
(to 4 decimal places) then putting =RAND() and selecting the highes
number for each selection (60 in one of my cases) then run thi
several hundred more times for an adequate sample size on which t
_begin_ my analysis.
Initially I figured I could simply multiply percentage of occurrence b
a rand() number and take the maximum there 60 times repeatedly. Howeve
it quickly became evident that I was dually weighted by the percentag
(i.e. the ratio of the taking the max between say one item at 6
*RAND() vs. another at 2%*RAND() is more like 30% : 0.2% ) A possibl
solution I figured might be taking the Max of the Square Roots of the
Outcome by 60, but I'm not certain that isn't an arbitrary way o
augmenting the disproportions. I haven't tried it yet so I don't kno
for sure what would happen.
I require a more compact way of creating possible lists of 60 items o
which to perform confidence testing. Any input, be it conceptual o
for practical application, would be welcomed.
Thanks,
MACR
10,000 long with each unique item maintaining its relative percentag
(to 4 decimal places) then putting =RAND() and selecting the highes
number for each selection (60 in one of my cases) then run thi
several hundred more times for an adequate sample size on which t
_begin_ my analysis.
Initially I figured I could simply multiply percentage of occurrence b
a rand() number and take the maximum there 60 times repeatedly. Howeve
it quickly became evident that I was dually weighted by the percentag
(i.e. the ratio of the taking the max between say one item at 6
*RAND() vs. another at 2%*RAND() is more like 30% : 0.2% ) A possibl
solution I figured might be taking the Max of the Square Roots of the
Outcome by 60, but I'm not certain that isn't an arbitrary way o
augmenting the disproportions. I haven't tried it yet so I don't kno
for sure what would happen.
I require a more compact way of creating possible lists of 60 items o
which to perform confidence testing. Any input, be it conceptual o
for practical application, would be welcomed.
Thanks,
MACR