In the past, I have used JMP to compute the correlation coefficient between two variables. I have this new data set where logically, there is a 1 to many relationship. A record can have as many as 20 unique values for Column z. Rather than stringing all the values together in Column z, it was decided that a NEW record would be created for each separate value to be loaded in column z. So,if the record must have the values A and B, two records are created, with column z loaded with A and B.
Does this have any bearing on the computed correlation coefficient, as you now have additional rows in the data set? would it be better to create a separate column that identifies the existence of a specific value? So, if I am tracking occurrences of "A", i would create a new column called "A exists" and populate the fiield wih Y or N???