Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: sample selection bias: rho and wald test


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: sample selection bias: rho and wald test
Date   Mon, 7 Jun 2010 08:43:26 -0700 (PDT)

--- On Mon, 7/6/10, sara.martinezdemorentin@unavarra.es wrote:
> I have a doubt related to the heckprob command. The estimated
> coefficient for rho is not statistically significant from 0,
> which I think suggests that there is not a selection bias.

I guess that this statement refers to the Wald test (the test
that is reported after each coefficient).

> However, the wald test reports that I should reject
> the null hypothesis that the equations are independent. 

I guess here you mean Likelihood ratio test, The test reported
in the line labeled: "LR test of indep. eqns. (rho = 0):".

Asymptotically the two should be the same, but in finite
samples the two are different. Often people prefer the
likelihood ratio test over the Wald test.

> The obtained rho lies outside the -1, +1 range. 

I guess that you mean the obtained /athrho, and not the
rho. Rho will always be within the -1 to 1 range, /athrho,
the Fisher's z transformation of rho, can be with the 
-infinity to +infinity range.

-- Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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   |   Site index