Hi,
I see that I didn't explain myself very well leading to this blossoming of
messages on the topic!
My final goal is to define (randomly) not equally sized groups, consisting of at
least c observations.
I thought that the best way to achive this objective was sampling from a
uniform (without replacement), the sampled values representing cut points
between contiguos groups.
In order to get groups of at lest c observations, I simply checked the samples
using the condition
min(diff(sort(x)))) < c,
skipping those that did not meet the condition.
So, I guess that the fact that the distribution from which I'm sampling is no
more uniform does not affect my final goal.
Is this correct?
thanks you,
melina
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
|