[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: standard errors after ivtobit

From   "Schaffer, Mark E" <>
To   <>
Subject   st: RE: standard errors after ivtobit
Date   Tue, 11 Nov 2008 00:12:51 -0000


> -----Original Message-----
> From: 
> [] On Behalf Of Rike
> Sent: 10 November 2008 23:35
> To: statalist
> Subject: st: standard errors after ivtobit
> Hi,
> I've been trying to find an answer to this but I couldn't so 
> here I go.

Did you check the manuals?  There is (almost?) always a "Methods and
Formulas" section in the entry for an estimator.

> Using the twostep ivtobit regression, the standard errors are not
> correct because in the second stage you are regressing y on the
> predicted error from the second stage regression (among other things).

No, the SEs are fine.  If you do two-step "by hand" you have to adjust
the SEs, but Stata's built-in -ivtobit- (like just about all Stata
commands) does this for you.

> My questions:
> 1) Are the reported standard errors already adjusted? If not, is there
> a simple option/command that I can use? I need the correct standard
> errors for t-statistics.
> 2) what about the mfx command? Can this be taken at face value or does
> it also need some kind of adjustment?

Same answers as above (but see the discussion of the -nonlinear- option
in the online -help mfx-).


> Thanks.
> *
> *   For searches and help try:
> *
> *
> *

Heriot-Watt University is a Scottish charity
registered under charity number SC000278.

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index