Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: RE: Question about Hausman test results: V_b - V_B not positive definite
From
Nick Cox <[email protected]>
To
[email protected]
Subject
Re: st: RE: Question about Hausman test results: V_b - V_B not positive definite
Date
Sun, 15 May 2011 17:53:03 +0100
You use it within Stata. That is, open Stata directly rather than
trying to open Stata by e.g. clicking on that .ado. The second route
clearly won't work because your machine associates the extension .ado
with Adobe Photoshop.
In Stata you use -help xtoverid- to find out how to use it, as Eric
already suggested.
Nick .
On Sun, May 15, 2011 at 4:19 PM, Christina SAKALI
<[email protected]> wrote:
> Thank you for the information. I downloaded the file you suggested
> (xtoverid) and it was installed on my PC, however when I tried to open
> it (with adobe photoshop) it wouldn't open. Can you please provide
> some more info on what this file is and how I can use it?
>
> Thanx a lot!
>
> On 14 May 2011 14:10, DE SOUZA Eric <[email protected]> wrote:
>> The standard Hausman test for fixed vs random effects is valid under very strict conditions (see Wooldridge's textbook, for instance). Clearly these conditions are not satisfied in your case.
>>
>> Download the user written routine -xtoverid- and use that instead:
>> -ssc install xtoverid-
>> Read the help file first.
Christina SAKALI
>
>> I am running a panel regression with 121 observations. My question regards the choice of fixed effects versus random effects specification.
>>
>> I carried out a hausman test and the results suggest that I cannot reject the Ho (Prob > 0.05) which I believe it means that the random effects model is preferred for my data.
>>
>> However I also get the message that the variance of the coefficient difference is not positive definite.
>>
>> Can someone explain to me what this means and whether I can trust the Hausman test results to be valid. Should I choose the random effects specification as more appropriate for my data?
>>
>> (Results from hausman test are provided below).
>>
>> Kind regards,
>>
>> Christina
>>
>> . hausman
>> You used the old syntax of hausman. Click here to learn about the new syntax.
>>
>>
>> ---- Coefficients ----
>> | (b) (B) (b-B) sqrt(diag(V_b-V_B))
>> | Consistent Efficient Difference S.E.
>> -------------+----------------
>> ------------------------------------------------
>> gg | 1.130961 1.075676 .0552858 .
>> trade | 1.544293 .4932033 1.05109 .4425236
>> sec | 3.217286 3.053046 .1642406 .
>> tert | 4.319199 3.989446 .329753 .
>> trans | 5.480038 5.574542 -.0945033 .
>> ------------------------------------------------------------------------------
>> b = consistent under Ho and Ha; obtained from xtreg
>> B = inconsistent under Ha, efficient under Ho; obtained from xtreg
>>
>> Test: Ho: difference in coefficients not systematic
>>
>> chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B)
>> = 5.64
>> Prob>chi2 = 0.3427
>> (V_b-V_B is not positive definite)
*
* 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/