Can anybody give me any insight as to how to write a program that forms confidence intervals for one component from a stratified table? I have a list of stratified dollar amounts in a table, and want a confidence interval for each of those amounts.
Your question isn't very specific about the end goal.
The basic formula for a normal approximation confidence interval for a proportion, p, is
p +- 1.96*SQRT(p(-1p)/n)
If you are looking for some sort of pooled confidence limit, you might be able to get it using PROC LOGISTIC with the STRATA statement. You might also be able to get something useful using PROC FREQ and the CMH option.