With, or without, replacement? And do you insist on using Microsoft's poor
prng, or be willing to settle for a much better one? (I'm thinking along the
lines of PopTools, a free add-in).

Finally, if you are looking for the distribution of the means, why not just
calculate it directly instead of using simulation? Your population size is
tiny; or are you just throwing out some numbers (5 of 30, 50 samples) for
the sake of concreteness?

DaveB


"stats" wrote in message
...
I have a data set of 30 numbers. I need to use excel to draw 50 random
samples of size 5 and then compute the mean for each sample. How do I do
this?