4
$\begingroup$

I am analyzing a multiply-imputed complex sample survey data using Stata. For normally distributed numerical variables I want to report the mean and standard deviation. However, the Stata command for estimating mean of a multiply-imputed survey data mi estimate: svy: mean [varlist] does give the standard error of the mean, not the standard deviation. I tried to search for valuable help using Google but in vain. My question is this: Under such circumstances, is it possible to obtain an unbiased estimate of the standard deviation using the formula $\sigma$ $=$ $SE$ $\Huge.$ $\sqrt{n}$?

$\endgroup$
2
  • $\begingroup$ i know. the answer to your question depends on the microdata. is it one of the ones covered on asdfree.com? $\endgroup$ – Anthony Damico Oct 15 '14 at 13:24
  • 1
    $\begingroup$ Not covered on asdfree.com. I collected my own data from a population of university students using a two-stage stratified cluster sampling technique. $\endgroup$ – Ayalew A. Oct 15 '14 at 13:35
4
$\begingroup$

If you really want to get to the standard deviation of the population distribution, you should mi xeq : generate y2 = y*y the squares of the variables, and then

mi estimate (sd : sqrt( _b[y2] - _b[y]*_b[y] ) ) : svy : mean y y2
mi testtransform sd

Note that the interface of multiple imputation and inference with complex survey data is extremely poorly researched into given the ubiquity of the issue. I outlined the literature and the steps elsewhere on statalist.

$\endgroup$
2
  • $\begingroup$ Should I do the mi xeq: generate y2 = y*y on M=0 and then follow it by mi updateor just do it on one of the imputations M=1 or M=2 or ... or M=20? $\endgroup$ – Ayalew A. Oct 16 '14 at 5:37
  • 1
    $\begingroup$ mi xeq should do it on all of them at once. $\endgroup$ – StasK Oct 17 '14 at 3:46
4
$\begingroup$

Short answer: NO, you can't do that, because the relationship between the standard error and standard deviation is only valid for simple random samples. For a longer answer look at Stata's FAQ. It does not include the issue with multiple imputation, but it shows how the standard deviation could be computed from survey data.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.