Re: st: Bug in estat teffects?

[vwiggins@stata.com (Vince Wiggins, StataCorp)]

Dave Garson <garson@n csu.edu> asks about finding the standardized
estimates using -estat teffects- after -sem-.

> [...]
> Whether I run the sem standardized or unstandardized, if in
> postestimation I ask for "estat teffects,standardized" the direct
> effect in question is always the unstandardized value (.73) and the
> partition is based on unstandardized values.
> [...]

It is easy to miss in the output, but the standardized estimates are
there.  The unstandardized estimates appear where they always do --
under the "Coef" column.  The standardized estimates appear in the far
right column. 


-- Vince 
   vwiggins@stata.com

 
 
------------ Original message ---------------------------
 
From: Dave Garson <garson@ncsu.edu>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:24.0) Gecko/20100101 Thunderbird/24.4.0
Subject: st: Bug in estat teffects?
Sender: owner-statalist@hsphsun2.harvard.edu
Lines: 29

I am running a model with the sem command. One direct path in the model 
has a value of .73 unstandardized or .34 standardized.

When total effects are partitioned into direct and indirect paths, " 
Direct effects are the path coefficients in the model.", as per "help 
sem teffects", which also states the "standardized (option) reports 
effects in standardized form".

Whether I run the sem standardized or unstandardized, if in 
postestimation I ask for "estat teffects,standardized" the direct effect 
in question is always the unstandardized value (.73) and the partition 
is based on unstandardized values.

To me this appears to be a bug in Stata, causing the "standardized" 
option to be ignored. However, is any reader able to shed light on this? 
I wish to do a standardized partition as I can in SPSS or SAS, and get 
the same results. Before I tell my students that this is impossible in 
Stata (except manually), I would like to check with the knowledge in 
this group.

Thanks,
Dave
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/