PROC STDRATE and Standard Error (Output)

[joe_perez]

Hi guys, 

I am using proc stdrate/direct method to make a simple adjustment age/sexe of my study population.

I wish to have a better understanding of either the formula or the conceptual procedure used by SAS to calculate the Standard Error displayed by-default in the Output; Is it possible to either change/add/or replace this information for example by the standard deviation or the variance ? 

Where could I obtain more details about these statistics (applied by SAS) ... the details are scarce in my opinion, or I am not looking at the right place. 

Thanks you so much for your help.

[StatDave]

The formulas are clearly presented in the STDRATE documentation online. See the Details: Direct Standardization section. The standard error in the results is just the square root of the variance, and the formula for the directly standardized rate or risk is shown in that section.

[joe_perez]

Thank you so much.

I'll verify all that again. 

Have a great day !

[joe_perez]

Sorry for coming back. 

Using a PROC UNIVARIATE,, the Standard deviation in the Output is indeed the square root of the variance,

but using a PROC STDRATE (to calculate an adjusted rate/adjusted proportion), the Standard Error is the equivalent of the Standard deviation (but in this case applied to a proportion) ?

... thanks in advance for your support. 

[StatDave · Accepted solution]

The directly standardized rate or risk is a statistic computed from the observations in the provided data. The standard error of that statistic is the standard deviation of the distribution of that statistic. It is the appropriate measure of variability on the values of the statistic. The standard deviation, as computed in PROC UNIVARIATE or PROC MEANS, is a measure of the variability of a set of observations themselves, not of a statistic computed on those observations.

[joe_perez]

That was very very helpful ...