Calculating the sample size needed for population proportion estimation

[Marco Manfrini]

Dear users,
I am new to Stata, and my question is: what is the correct procedure for calculating the sample size needed for population proportion estimation, given that p0 is the attended proportion and e the margin of error (alpha and beta may vary)? I did not find the procedure in the graphical interface of power and sample size analysis neither in the manual for executing commands from the prompt.
Thank you in advance.

[Al Bothwell]

Hi Marco,

I use a user-written package called svysampsi. To install, run this code in Stata

 ssc install svysampsi

After installation, type

 help svysampsi

[Rich Goldstein]

for the built-in procedures, a general starting place is

help power

[Andrew Lover]

For measuring a single simple proportion (no survey design effects etc.), you can just decide what precision is "sufficient" to answer your research/policy question, and iterate using cii. Then add say 5-10% for non-response, or missing data, and you should be all set.

cii proportion 100 20

might be sufficient for some studies, but not for others.

[Marco Manfrini]

Done. Thank you all!