If the mean of a variable is referenced in a tabulate, and then once again the variable is referenced as a denominator, will the mean be used?
Suppose I have a dataset which contains 12 records. One for every month. On that record is a variable called call_dropped, and another variable called total calls.
For a monthly report, the percent of calls dropped would be expressed as
proc tabulate data=mydata
calls_dropped = ''*sum='Calls Dropped'*f=comma6.
total_calls = ''*sum='Total Calls'*f=comma9.
calls_dropped =pctsum='Percent Dropped'*f=pctfmt8.2,
month=' '
/box='Call Data';
BUT, If I wanted a yearly average, I would do the following:
calls_dropped = ''*mean='Avg. Calls Dropped per Month'*f=comma6.
total_calls = ''*mean='Avg. Total Calls per Month'*f=comma9.
calls_dropped =pctsum='Avg. Percent Dropped per Month'*f=pctfmt8.2,
month=' '
/box='Call Data';
Since there is no pctmean, what would the denominator for this expression be? The mean of all 12 months (as I would like), or the sum of all 12 months (not good).
