Karen wants to do two things:
1) test if the proportions are the same or different, and then
2) if (1) shows significant evidence of a difference in proportions
between the groups, she will want to test each pair of groups to see
where the difference lies.
Thanks for the detailed and informative answer May. Given that this is
fairly complex, is there a command somewhere, e.g.
TestEverythingYouCouldEverThinkOf, that would automate this process? If
not, would such a command be difficult to create? Besides the current
question on proportions, this would seem to be useful for any sort of
problem involving group comparisons, such as in Anova.
