Heteroscedasticity Test

[David Clemson]

I need to run heteroscedasticity test for my linear regression models. I
have searched the archives and I could not find a way to test
heteroscedasticity
of the model. Only information ı could find was using plots but I need some
numerical way to test it. Could you please help me on this? I need to
test heteroscedasticity
of my linear regression models by applying extra tests available in SPSS (I
currently use SPSS 15). Since ı am not an expert on coding nd macros I need
to use menus and parameter screens.

Thanks*
*

[Matthias Spörrle]

David,

since you are looking for an easy procedure:
a very simple test concerning heteroscedasticity would be the Levene test.
The procedure is described via video at
http://www.stat.tamu.edu/spss.php
[Regression Analysis -> Levene Test]

HTH
Cheers
Matthias



On 4/5/07, David Clemson <[hidden email]> wrote:

>
> I need to run heteroscedasticity test for my linear regression models. I
> have searched the archives and I could not find a way to test
> heteroscedasticity
> of the model. Only information ı could find was using plots but I need
> some
> numerical way to test it. Could you please help me on this? I need to
> test heteroscedasticity
> of my linear regression models by applying extra tests available in SPSS
> (I
> currently use SPSS 15). Since ı am not an expert on coding nd macros I
> need
> to use menus and parameter screens.
>
> Thanks*
> *
>

[Muir Houston]

See also Raynald's SPSS site where details may be found for White's test for heteroscedasticity and details of how to apply it
at:
http://www.spsstools.net/spss.htm#Heteroscedasticity
 
 
 
Muir Houston
Research Fellow
CRLL
Institute of Education
University of Stirling
FK9 4LA
01786-46-7615

________________________________

From: SPSSX(r) Discussion on behalf of Matthias Spörrle
Sent: Fri 06/04/2007 10:44
To: [hidden email]
Subject: Re: Heteroscedasticity Test



David,

since you are looking for an easy procedure:
a very simple test concerning heteroscedasticity would be the Levene test.
The procedure is described via video at
http://www.stat.tamu.edu/spss.php
[Regression Analysis -> Levene Test]

HTH
Cheers
Matthias



On 4/5/07, David Clemson <[hidden email]> wrote:

>
> I need to run heteroscedasticity test for my linear regression models. I
> have searched the archives and I could not find a way to test
> heteroscedasticity
> of the model. Only information i could find was using plots but I need
> some
> numerical way to test it. Could you please help me on this? I need to
> test heteroscedasticity
> of my linear regression models by applying extra tests available in SPSS
> (I
> currently use SPSS 15). Since i am not an expert on coding nd macros I
> need
> to use menus and parameter screens.
>
> Thanks*
> *
>

--
The University of Stirling is a university established in Scotland by
charter at Stirling, FK9 4LA.  Privileged/Confidential Information may
be contained in this message.  If you are not the addressee indicated
in this message (or responsible for delivery of the message to such
person), you may not disclose, copy or deliver this message to anyone
and any action taken or omitted to be taken in reliance on it, is
prohibited and may be unlawful.  In such case, you should destroy this
message and kindly notify the sender by reply email.  Please advise
immediately if you or your employer do not consent to Internet email
for messages of this kind.

[Peck, Jon]

I would like to point out that in SPSS 15, the GENLIN procedure, which is part of the Advanced Models option and can be found in the menus under Analyze/Generalized Linear Models, provides the White HCCM estimator that is referred to in the cited paper.  See the "Estimation" tab of the dialog and select "Robust Estimator" radio button in the Covariance Matrix group.

c) Whites Standard Errors
White (op cit) developed an algorithm for correcting the standard errors in OLS when
heteroscedasticity is present. The correction procedure does not assume any particular form of
heteroscedasticity and so in some ways White has “solved” the heteroscedasticity problem.


For comparison, here are the parameter estimation tables from GENLIN for a regression equation on the employee data.sav file.  using, first, the model based estimator and then the robust estimator.  The dependent variable is salary, and the predictors are salbegin, educ, and gender and jobcat dummy variables.  I have extracted just the coefficients and their standard errors in order to make the plain text email readable.  As you can see, the parameter estimates are identical, but in this case, the standard errors are substantially different.  One might well expect heteroscedasticity in an equation like this.

Parameter       B       Std. Error
 (Intercept)    13795.745       3247.0083
[gender=f]      -1757.046       797.0853
[gender=m]      0       .
[jobcat=1]      -12236.096      1430.1509
[jobcat=2]      -9217.461       2137.2856
[jobcat=3]      0       .
salbegin                1.217   .0748
educ            786.599 163.2750
(Scale) 51566354.525    3349594.6723


Parameter       B       Std. Error
 (Intercept)    13795.745       3769.5888
[gender=f]      -1757.046       709.1054
[gender=m]      0       .
[jobcat=1]      -12236.096      2054.5460
[jobcat=2]      -9217.461       2246.6021
[jobcat=3]      0       .
salbegin                1.217   .1176
educ            786.599 141.9034
(Scale) 51566354.525    3349594.6723

GENLIN has lots of other capabilities, too.

HTH,
Jon Peck
SPSS


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Muir Houston
Sent: Friday, April 06, 2007 5:29 AM
To: [hidden email]
Subject: Re: [SPSSX-L] Heteroscedasticity Test

See also Raynald's SPSS site where details may be found for White's test for heteroscedasticity and details of how to apply it
at:
http://www.spsstools.net/spss.htm#Heteroscedasticity



Muir Houston
Research Fellow
CRLL
Institute of Education
University of Stirling
FK9 4LA
01786-46-7615

________________________________

From: SPSSX(r) Discussion on behalf of Matthias Spörrle
Sent: Fri 06/04/2007 10:44
To: [hidden email]
Subject: Re: Heteroscedasticity Test



David,

since you are looking for an easy procedure:
a very simple test concerning heteroscedasticity would be the Levene test.
The procedure is described via video at
http://www.stat.tamu.edu/spss.php
[Regression Analysis -> Levene Test]

HTH
Cheers
Matthias



On 4/5/07, David Clemson <[hidden email]> wrote:

>
> I need to run heteroscedasticity test for my linear regression models. I
> have searched the archives and I could not find a way to test
> heteroscedasticity
> of the model. Only information i could find was using plots but I need
> some
> numerical way to test it. Could you please help me on this? I need to
> test heteroscedasticity
> of my linear regression models by applying extra tests available in SPSS
> (I
> currently use SPSS 15). Since i am not an expert on coding nd macros I
> need
> to use menus and parameter screens.
>
> Thanks*
> *
>

--
The University of Stirling is a university established in Scotland by
charter at Stirling, FK9 4LA.  Privileged/Confidential Information may
be contained in this message.  If you are not the addressee indicated
in this message (or responsible for delivery of the message to such
person), you may not disclose, copy or deliver this message to anyone
and any action taken or omitted to be taken in reliance on it, is
prohibited and may be unlawful.  In such case, you should destroy this
message and kindly notify the sender by reply email.  Please advise
immediately if you or your employer do not consent to Internet email
for messages of this kind.