I have a dataset of 3500 cases, with two categorical variables of 3 levels each.  I create a 3 X 3 matrix of distributions.  Then I select all the points in one of the 9 squares.  Then I choose “show measures for selection.”  I expect to see 393 in that one square and 0 in all the rest.  But I see something else – see attached.  Is this a bug or do I not understand what “show measures for selection” is doing?  (I do know that this move doesn’t tell me anything new – but I’m still curious why it tells me what it does…  It look to me like it’s just looking at one of the two categorical variables in considering the selection.)