How Do I Calculate Weighted Unadjusted and Adjusted Percentage?

[Roman Johnson]

Hi, all:

I have already set my design weight specifications (e.g. strata, psu) for my survey data, but I do not yet how to calculate the unadjusted and adjusted percentages for my variables of interest (e.g. mammogram use by citizenship status). Could someone please give me the correct syntax to calculate an unadjusted and adjusted percentage using the svy command? I couldn't find it using the help svy command.

[Marcos Almeida]

In the "Quick Start" of the Stata Manual about - svy: tabulate twoway - you get the information.

[Steve Samuels]

And if by "adjusted" you mean "standardized", then see the stdize and stdweight options and the SVY manual entries on direct standardization and poststratification.

[Steve Samuels]

In a previous post, you said you were working with the National Health Interview Survey. The weights for that survey already include a post-stratification adjustment to national totals. Therefore no further adjustment is needed.

ftp://ftp.cdc.gov/pub/Health_Statist...6/srvydesc.pdf

[Hassen Ali]

Dear all, Thank you very much!! I have learned a lot from your daily posts and replies.
Respectfully,Hassen

[Roman Johnson]

Hi,

By adjusted percentage, I meant doing a predicted marginal statement with the svy command.

[Roman Johnson]

So, I'm assuming that if I calculate adjusted percentages using the svy predicted marginal command, I calculate unadjusted percentages using a svy two way cross tab?

[Steve Samuels]

Yes. However the output of svy tabulate (a table) differs from that of margins (estimates, standard errors, and confidence intervals in one row). Moreover with tabulate, you specify the statistics you want to see.

But you can get output similar to margin's with svy proportion and the over() option. (proportions displays proportions of "No's" and Yes's" in separate sections.) Try out the possibilities and see for yourself.