topic How can you standardize/z-transform variables after imputing missing values using proc mi? in Statistical Procedures

How can you standardize/z-transform variables after imputing missing values using proc mi?

ejamro — Tue, 16 Nov 2021 18:42:52 GMT

Re: How can you standardize/z-transform variables after imputing missing values using proc mi?

PaigeMiller — Tue, 16 Nov 2021 19:06:53 GMT

PROC STDIZE

Re: How can you standardize/z-transform variables after imputing missing values using proc mi?

ejamro — Tue, 16 Nov 2021 19:12:04 GMT

Thanks for your reply. Is there some trick to using proc stdize on imputed data? The resulting mean is very close to zero when I use proc mi and then proc stdize (for example, mean=-8.19047E-16), but it isn't exactly zero as I'd expect.

Re: How can you standardize/z-transform variables after imputing missing values using proc mi?

PaigeMiller — Tue, 16 Nov 2021 20:15:25 GMT

@ejamro wrote:

Thanks for your reply. Is there some trick to using proc stdize on imputed data? The resulting mean is very close to zero when I use proc mi and then proc stdize (for example, mean=-8.19047E-16), but it isn't exactly zero as I'd expect.

There is no trick. Your mean is zero except for roundoff error. Nothing to worry about.