Thank you so much to "PaigeMiller" and Bllardw" for their responses. My total sample size is fixed. I'm estimating the proportions of people for 12 categorical indicators disaggregated by sex, age, education . Here is the example of data: variable name: indicate 1 Answer categories code great 1 moderate 2 small 3 not at all 4 variable name: indicate 2 Answer categories code great 1 moderate 2 small 3 not at all 4 . . . variable name: indicate 12 Answer categories code great 1 moderate 2 small 3 not at all 4 an example of my syntax is (I'm using SUDAAN in SAS) proc crosstab data=mydata design=BBR; weight myweight; repwgt mybootstrapweight; tables gender * indicate1; class gender indicate1; setenv colwidth = 1; output nsum serow rowper lowrow uprow /filename= out_ filetype=SAS; data out; set out_; row= round (rowper, .1); lowR= round (lowrow, .1); upR= round (uprow, .1); cv= round ((serow/rowper *100), .1); keep gender indicate 1 nsum row lowR upR cv; run; I need to marked the output of "nsum" based on the thresholds, for example nsum>=220 be marked as A (good quality), 110>=nsum>=220 be marked as E (marginal quality) and nsum<110 be marked as F for poor quality. Hope I could explain it well now. Thank you so much.
... View more
