Help with PROC CORR

[deleted_user]

I have data that I need to correlate. It is in categories and subcategories. Example:
category - subcategory - set1 - set2
Leadership - A - 4.07 - 2.38
Leadership - B - 4.09 - 2.19
Leadership - C - 3.36 - 2.62
Leadership - D - 3.47 - 2.11
Leadership - Total - 3.75 - 2.33

I want to see which (A, B, C, D) provides the best correlation to the total for that group for each individual set. I have about 15 sets and 10 groups - all with subcategories.

Help!

Thanks so much!

[Paige]

"the best correlation to the total for that group"??

That is not a statistical concept that I am familiar with. Could you define it further?

[deleted_user]

I want to see which subcategory (a, b, c, or d) best correlates to the total.

[deleted_user]

Here is what seemed to work for me:

I transposed the data and then used:

PROC CORR data=name;
var LeadershipT;
with leadershipa leadershipb leadershipc leadershipd;
RUN ;