Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Calculating RSE for point estimates


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Calculating RSE for point estimates
Date   Tue, 6 May 2008 22:54:18 +0100

I am not aware of an .ado to calculate relative standard error. Perhaps
people on this list are more used to looking at t or z statistics, which
carry the same information. I doubt that there could be a single .ado to
do what you want, but the general recipe is surely to get hold of the
ingredients and do the work after estimation. For example, after a
-regress- you can do this: 

mata : sqrt(diagonal(st_matrix("e(V)")))' :/ st_matrix("e(b)")

For many people, perhaps you, that too counts as "by hand". Perhaps that
is what you are doing already. 

Nick 
n.j.cox@durham.ac.uk 

Dayna N KIRK

Hello! I am using STATA 9.2 for Windows
 
Data supression guidelines for statistical reliability vary, however
Healthy People 2010 and many U.S. health departments use a relative
standard error (RSE) > 30% as a central criteria for data supression,
primarily for BRFSS and YRBS.  While I can easily get STATA to present
the standard error for point estimates in -svyset-, I have to calculate
the RSE by hand (RSE = 100 x (SE(r)/r)).  This is time consuming given
multiple counties, regions, race/ethnicities, etc. I have searched the
archives, read the manual three times and asked some experts to no
avail. Is there a simple command like -se-, -ci-, etc. for relative
standard error?
 
Has anyone tried to create an ado file to address this simple, yet daily
need?

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index