06-07-2016 07:49 PM
Hello I'm having trouble obtaining 95% CI for a proportion. This proportion is a result of a binary variable crossed with another binary variable. I want to know the 95% CI for percentage of mothers that breastfeed exclusively (yes or no, binary) for women of diferent nationalities (also binary). I've tried variations of proc freq but its giving me the confidence interval of the odds ratio instead of for the actual values of % breastfeeding exclusvely. Any help?
proc freq data=meley.burum2; table momsnationality*exclusive/chisq; run;
*proc freq data=meley.burum2 order=freq;
06-07-2016 10:51 PM
It sounds like you want GROUP analysis.
proc freq data=meley.burum2 order=freq;
06-08-2016 12:11 AM
Sort by nationality and get the CI for each.
proc sort data=meley.burum2; by momsnationality; run; proc freq data=meley.burum2; by momsnationality; tables exclusive / binomial; weight count; run;