What happens when you
build a table to show data for a variable that has three different
possible values (for example, 'A', 'B', and 'C'), but in your data
set, you end up with records for only two ('A' and 'B')? This situation
often comes up when you are dealing with survey data. A question had
three valid responses, but the respondents only chose two of the three
answers. It would be nice if the PROC TABULATE table would show all
three choices, even though one was never chosen.
You cannot do this with the MISSING option. That option shows only that you have records in the data set where there is missing data. In this case we are not missing individual values, we are missing records. There ...