I'm working in SAS 9.4, and the following code is exactly what I'm trying to do, with the exception of the rate calculation because proc tabulate doesn't support mathematical operations (i.e., it runs without the inclusion of rate=num_deaths/pop*100000=''*f=8.1 which is not kosher code). My input dataset contains all combinations of demographic categories and their associated value for num_deaths and population. I need columns for years, under which would each have num_deaths, population, colpctn, and rate (num_deaths/pop*100000). The rows would contain each level of the demographic variables age, sex, and race_eth. proc tabulate data=death_summary;
class year sex race_eth age;
var num_deaths pop rate;
table sex race_eth='Race/Ethnicity' age,
year=''*(num_deaths='Count'*sum=''*f=8. num_deaths='%'*colpctsum=''*f=8.1 pop='Population'*sum=''*f=8.
rate=num_deaths/pop*100000=''*f=8.1) / misstext='0' nocellmerge;
run; I have tried proc report, but I'm failing at nesting columns, and depending on how I've structured it, it seems to prefer to summarize observations rather than num_deaths and population. I cannot calculate rate in death_summary, because I need the calculation to be dynamic after the proc tabulate groupings. I also intend to, once this runs, to add a "where" statement to filter by state, so that I can have a macro that runs it for each of the 50 US states. Does anyone have any suggestions on how I could calculate rates while maintaining the groupings both by column years and row demographics?
... View more
