Calculation of the SE for a proportion in Stata 15

[Brendan Girdler-Brown]

Can anyone explain, please, why Stata calculates the normal approximation SE for a proportion [following the command .proportion variable, citype(normal)] using the formula for the SE squared as p*q/(n-1) when, in fact, the derivation of the formula from first principles is given by p*q/n?

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata Code in code delimiters, readable Stata output, and sample data using dataex.

I don't see where you got that formula from - the only equations I can see in the documentation in Stata 15 that look like your equation use n. If you really can show that Stata is not using the formula it says it is using, then contact the help folks at Stata.

[Joseph Coveney]

I second that: take it up with Stata Technical Services. I can confirm that the implementation in proportion seems to be incorrect.

[Brendan Girdler-Brown]

Thank you for your replies. This question has been referred to Stata Technical Services and I will report back.