04-02-2013 05:32 PM
I want to use proc tabluate to make a cross tab of bad rates using a variable bad that is 1 if the observation is bad and 0 otherwise. My sample code is attached. As I see the main issue, I want to treat bad as a class and analysis variable in order to get the sum of the bads and divide by the number of observations. I know I can't do that. I'm sure there is a very elegant way to do this. Many thanks for people's time!
04-02-2013 08:45 PM
You are not the first person trying to do this the hard way. With a 0/1 variable, make it an analysis variable and calculate the mean. That's the same as the sum of the "bads" divided by the number of observations.
04-04-2013 11:34 AM
Adding to Astounding note, and N gives number of records considered and Sum yields number of bad. I use this approach a lot as my customers will almost always ask something like "How many bad?" or "How many records?" when seeing 0 or 100 percent, or some value above/below a theshold of interest, for some group. It heads off follow-up request to provide all 3 up front.