The np chart was used in the Red Bead Experiment as there was close
to a binomial process going on (though technically closest to a
hypergeometric).
Hello, all,
I've heard this from a number of people, and I've often wondered: why a
hypergeometric? I have always understood that the hypergeometric
distribution characterizes data when you are sampling without replacement.
In the Red Bead, we dump the beads back into the bowl after each pull. Is
the idea that if we loaded each paddle one bead at a time, we would be
pulling each of the fifty in that sample without replacement, so each
bernoulli experiment (each bead on the paddle) constitutes "a sample?"
Probably of more practical concern is this: can we tell the difference? I'd
be willing to bet that, in this situation, there's not a test sensitive
enough to discern the difference between the hypergeometric and the
binomial. The binomial may be--at least theoretically--a flawed model for
this situation, but it is certainly a useful one.