I can only explain this question through an example and I cannot go for a logit or logistic.

This is the prevalence of tobacco consumption in rural and urban areas
Variable Urban Urban Rural Rural Total Total P value
Age % 95% CI % 95% CI % 95% CI for everything 0.001
31-40 38.2 35.3-41.1 38.2 20.7-,25 23.9 27-31 x
41-50 31.3 29.1-33.5 31.3 23.8-28.5 28.2 26-30 x
51-60 18.8% 16.5-21.4 18.8 23.8-28.5 23 21-24 x
60 & Above 11.7 10.1-13.6 11.7 23.2-27.3 19 17-21 x
This table will have gender, marital status, wealth quintile, and religion, and 9 more things. I am trying to see if there is any difference between rural and urban tobacco consumption and the prevalence from a survey from the selected group. The survey weight should not be applied because it cannot be generalised to the whole population

The codes I had to use were

Code:
proportion iv_age if dv_tobaccouser==1, over (Region)
Note tobaccouser==0 means "non user"

and a chi-squared using the tab command to the combined value but I need to find the p-value of x in the table.

Code:
foreach v of varlist iv_*{
    tab `v' (region) if dv_tobaccouser==1, col chi2
}
This is a faulty code because the indepvar and depvar are in the wrong places. I'm in the initial phase of learning.

I am unable to figure out how to get the confidence interval and the p-Values without a survey setup.


Q. 1. Is there any other way or the last way is to use the analytical weighing method?
Q. 2. Do you think I should break the dataset into tobacco user urban and tobacco user rural and generate variables for the factor variables?
I tried the

Code:
by region: tab tobaccouser i.indepvar, col chi2
the error code was r(101)
factor-variable and time-series operators not allowed

I hope I was able to ask the right question rightly.

Thanks
-----