3:52 a.m.
Hi all:
I am using gretl to teach the intro. to econometrics class.
Gujarati's text along with other popular texts such as Hill,
Griffiths,Lim mentions the Jarque-Bera normality test. Iknow that Lee
Adkins' ebook offers a gretl script for this test, however, I was
wondering if this test could be implemented.
If yes, I also wonder if it could be made available as an option for
the normality of the residuals under model tests. Is it possible to
also have an option to set the number of bins for the chi^2 test here?
Cheers,
Talha
--
"Remember not only to say the right thing in the right place, but far
more difficult still, to leave unsaid the wrong thing at the tempting
moment." - Benjamin Franklin (1706-1790)
--
11:12 a.m.
On Fri, 9 Nov 2007, Talha Yalta wrote:
I am using gretl to teach the intro. to econometrics class.
Gujarati's text along with other popular texts such as Hill,
Griffiths,Lim mentions the Jarque-Bera normality test. Iknow that Lee
Adkins' ebook offers a gretl script for this test, however, I was
wondering if this test could be implemented.
I suppose so, though the literature suggests that Doornik-Hansen,
which is what we have, is better.
If yes, I also wonder if it could be made available as an option
for the normality of the residuals under model tests. Is it
possible to also have an option to set the number of bins for
the chi^2 test here?
On the second question the answer is No, since it's not a "number
of bins" thing; see Doornik and Hansen (1994), referenced in the
bibliography to the Gretl Users' Guide.
http://www.doornik.com/research/normal2.pdf
Allin.
9:28 p.m.
New subject: Normality Tests (once again)
On Fri, 9 Nov 2007, I wrote:
On Fri, 9 Nov 2007, Talha Yalta wrote:
> Gujarati's text along with other popular texts such as Hill,
> Griffiths,Lim mentions the Jarque-Bera normality test. Iknow that Lee
> Adkins' ebook offers a gretl script for this test, however, I was
> wondering if this test could be implemented.
I suppose so, though the literature suggests that Doornik-Hansen,
which is what we have, is better.
Let me expand on this a little, since it's an issue that comes up
from time to time.
Both the Jarque-Bera test and the Doornik-Hansen test (which is
derived from an earlier variant by Bowman-Shenton, and is
implemented by gretl) are based on the skewness and kurtosis of
the sample data.
The fact that the third and fourth sample central moments are
involved is what accounts for the 2 degrees of freedom in the
chi-square distribution to which these tests are referred (and not
the placement of the sample data into 2 bins).
As I understand it, the Doornik-Hansen test is better sized in
relation to the chi-square(2) distribution than Jarque-Bera.
That is, if you compute the Jarque-Bera statistic and refer it to
chi-square(2) -- which is the usual approximation -- the resulting
p-values may be quite misleading, even for fairly substantial
samples (e.g. n = 70). See for example
http://www.alglib.net/statistics/hypothesistesting/jarqueberatest.php
To get accurate p-values for Jarque-Bera you need specific
quantiles obtained via Monte Carlo methods. My understanding is
that you can refer the Doornik-Hansen statistic to chi-square(2)
with relative impunity, even for relatively modest sample sizes.
If anyone has information otherwise, please let us know.
Allin.
2:38 a.m.
New subject: Normality Tests (once again)
On Fri, 9 Nov 2007, Allin Cottrell wrote:
On Fri, 9 Nov 2007, I wrote:
> On Fri, 9 Nov 2007, Talha Yalta wrote:
>
>> Gujarati's text along with other popular texts such as Hill,
>> Griffiths,Lim mentions the Jarque-Bera normality test. Iknow that Lee
>> Adkins' ebook offers a gretl script for this test, however, I was
>> wondering if this test could be implemented.
>
> I suppose so, though the literature suggests that Doornik-Hansen,
> which is what we have, is better.
Let me expand on this a little, since it's an issue that comes up
from time to time.
Let me expand on the expansion a little :-)
First, you can compare by yourselves the two tests via the attached simple
script, which does 16384 replications with a sample size of 32; with the
help of a few seconds of CPU time, you'll see what Allin means.
Second, the typical use of a normality test is in the context of a
regression model. What happens if you have to reject the null? Well, if
your sample is big enough, the answer is: nothing to be worried about.
Normality is not needed in large samples for OLS to have its nice
distributional properties. (Of course, when the null is rejected, you
ought to ask yourself why; more often than not, it is a symptom of
heteroskedasticity.) Rejection of normality, by itself, may only matter in
small samples: but then, the last thing you want is a normality test that
has poor finite-sample properties.
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti
1:23 p.m.
New subject: Normality Tests (once again)
On Sat, 10 Nov 2007, Riccardo (Jack) Lucchetti wrote:
First, you can compare by yourselves the two tests via the
attached simple script, which does 16384 replications with a
sample size of 32; with the help of a few seconds of CPU time,
you'll see what Allin means.
Very nice! To make this sort of thing a little easier I've
enabled in CVS the "--silent" flag for the freq command: the test
result is stored but nothing is printed.
Allin.
3:42 a.m.
Sorry I couldn't reply over the weekend. I appreciate all the comments
and Allin and Jacks's explanations. I now see how it is indeed a
better idea to use this as an example for showing gretl's scripting
functionality as Peter suggested.
Cheers
Talha
On Nov 9, 2007 7:12 PM, Allin Cottrell <cottrell(a)wfu.edu> wrote:
On Fri, 9 Nov 2007, Talha Yalta wrote:
> I am using gretl to teach the intro. to econometrics class.
>
> Gujarati's text along with other popular texts such as Hill,
> Griffiths,Lim mentions the Jarque-Bera normality test. Iknow that Lee
> Adkins' ebook offers a gretl script for this test, however, I was
> wondering if this test could be implemented.
I suppose so, though the literature suggests that Doornik-Hansen,
which is what we have, is better.
> If yes, I also wonder if it could be made available as an option
> for the normality of the residuals under model tests. Is it
> possible to also have an option to set the number of bins for
> the chi^2 test here?
On the second question the answer is No, since it's not a "number
of bins" thing; see Doornik and Hansen (1994), referenced in the
bibliography to the Gretl Users' Guide.
http://www.doornik.com/research/normal2.pdf
Allin.
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
--
"Remember not only to say the right thing in the right place, but far
more difficult still, to leave unsaid the wrong thing at the tempting
moment." - Benjamin Franklin (1706-1790)
--
6201
days inactive
6204
days old
5 comments
3 participants
participants (3)
-
Allin Cottrell -
Riccardo (Jack) Lucchetti -
Talha Yalta