putting dummy variables into regression

[Courtney Nichols]

Hello All,
I apologize, I feel like this should be very simple to figure out, but I'm new to SPSS and regression analysis in general.  I need to run a regression where some of my predictor variables are continuous and others are categorical.  I'm having trouble figuring out how to include the categorical and need help with two things:
1) How do I get SPSS to code the categorical variables (ie education level) into a collection of dummy variables?  (I hear there is a "class statement" command in SAS that does this but don't see how to do it in SPSS and am also not using command syntax)
2) Once I have that, what regression in the analyze menu do I run to get output that gives me both an overall F-test that tells me whether the overall categorical variable (ie education level) should remain in the model and beta coefficients or t-tests to interpret for each dummy (ie high school, some college, graduate, etc. in relation to the omitted comparison group)?
Many thanks in advance!
Courtney
"The grand essentials to happiness in this life are something to do, something to love and something to hope for. " -- Joseph Addison
 
_________________________________________________________________
Get the power of Windows + Web with the new Windows Live.
http://www.windowslive.com?ocid=TXT_TAGHM_Wave2_powerofwindows_122007
====================To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Keith McCormick]

Hello Courtney,

My favorite way to dummy code is the following syntax:

COMPUTE dummy_var1 = (category_var=1).

In practice, it might look like this:

COMPUTE Biology_YN = (Major = 1).
COMPUTE Math_YN = (Major = 2).
etc.

If you need an F-test for the whole set of variables (like all the
majors), you can put the entire set in a "block". And request Change
in R Sq. The Delta R Sq can serve as an overall test for the dummy
variables. To use a block, click on the next button when you are in
regression. It sounds like you have more than one categorical var, so
you can add a block for each one. The sequence of the blocks will make
a difference in the interpretation, so you will want to think that
through.

Feel free to follow-up, but that should get you started.

Hope that helps.

Keith
www.keithmccormick.com

On Dec 27, 2007 2:42 PM, Courtney Nichols <[hidden email]> wrote:

> Hello All,
> I apologize, I feel like this should be very simple to figure out, but I'm new to SPSS and regression analysis in general.  I need to run a regression where some of my predictor variables are continuous and others are categorical.  I'm having trouble figuring out how to include the categorical and need help with two things:
> 1) How do I get SPSS to code the categorical variables (ie education level) into a collection of dummy variables?  (I hear there is a "class statement" command in SAS that does this but don't see how to do it in SPSS and am also not using command syntax)
> 2) Once I have that, what regression in the analyze menu do I run to get output that gives me both an overall F-test that tells me whether the overall categorical variable (ie education level) should remain in the model and beta coefficients or t-tests to interpret for each dummy (ie high school, some college, graduate, etc. in relation to the omitted comparison group)?
> Many thanks in advance!
> Courtney
> "The grand essentials to happiness in this life are something to do, something to love and something to hope for. " -- Joseph Addison
>
> _________________________________________________________________
> Get the power of Windows + Web with the new Windows Live.
> http://www.windowslive.com?ocid=TXT_TAGHM_Wave2_powerofwindows_122007
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Keith McCormick]

Hi,

You could ... but it is going to amount to the same work. Doing so
will still allow you to do the regression in the menus - as do I.

In Transform ... Compute, you would put the new name, dummyvar1, in
the box in the upper left and you would type the rest of the formula
in the formula box on the right. In this case, it would be Major = 1.
You will have to do this for each dummy variable.

cheers, Keith

On Dec 28, 2007 10:31 AM, Courtney Nichols <[hidden email]> wrote:

>
> Hi Keith,
>  Thanks so much for your help.  Do you happen to know how to do the dummy
> coding without using command syntax.  I've never used the syntax and just
> use the point and click interface with the dropdown menus and such.  I can
> learn how to use the program with syntax if needed, but thought I'd ask if
> there is a way to do so with the menus.
>  Many thanks,
>
>  Courtney
>
>
>
> "The grand essentials to happiness in this life are something to do,
> something to love and something to hope for. " -- Joseph Addison
>
>
>  ________________________________
>  > Date: Fri, 28 Dec 2007 03:04:03 -0500
> > From: [hidden email]
> > Subject: Re: putting dummy variables into regression
> > To: [hidden email]
>
>
> >
> > Hello Courtney,
> >
> > My favorite way to dummy code is the following syntax:
> >
> > COMPUTE dummy_var1 = (category_var=1).
> >
> > In practice, it might look like this:
> >
> > COMPUTE Biology_YN = (Major = 1).
> > COMPUTE Math_YN = (Major = 2).
> > etc.
> >
> > If you need an F-test for the whole set of variables (like all the
> > majors), you can put the entire set in a "block". And request Change
> > in R Sq. The Delta R Sq can serve as an overall test for the dummy
> > variables. To use a block, click on the next button when you are in
> > regression. It sounds like you have more than one categorical var, so
> > you can add a block for each one. The sequence of the blocks will make
> > a difference in the interpretation, so you will want to think that
> > through.
> >
> > Feel free to follow-up, but that should get you started.
> >
> > Hope that helps.
> >
> > Keith
> > www.keithmccormick.com
> >
> > On Dec 27, 2007 2:42 PM, Courtney Nichols <[hidden email]>
> wrote:
> > > Hello All,
> > > I apologize, I feel like this should be very simple to figure out, but
> I'm new to SPSS and regression analysis in general. I need to run a
> regression where some of my predictor variables are continuous and others
> are categorical. I'm having trouble figuring out how to include the
> categorical and need help with two things:
> > > 1) How do I get SPSS to code the categorical variables (ie education
> level) into a collection of dummy variables? (I hear there is a "class
> statement" command in SAS that does this but don't see how to do it in SPSS
> and am also not using command syntax)
> > > 2) Once I have that, what regression in the analyze menu do I run to get
> output that gives me both an overall F-test that tells me whether the
> overall categorical variable (ie education level) should remain in the model
> and beta coefficients or t-tests to interpret for each dummy (ie high
> school, some college, graduate, etc. in relation to the omitted comparison
> group)?
> > > Many thanks in advance!
> > > Courtney
> > > "The grand essentials to happiness in this life are something to do,
> something to love and something to hope for. " -- Joseph Addison
> > >
> > > _________________________________________________________________
> > > Get the power of Windows + Web with the new Windows Live.
> > > http://www.windowslive.com?ocid=TXT_TAGHM_Wave2_powerofwindows_122007
> > > To manage your subscription to SPSSX-L, send a message to
> > > [hidden email] (not to SPSSX-L), with no body text except the
> > > command. To leave the list, send the command
> > > SIGNOFF SPSSX-L
> > > For a list of commands to manage subscriptions, send the command
> > > INFO REFCARD
> > >
> >
> > =====================
>
> > To manage your subscription to SPSSX-L, send a message to
> > [hidden email] (not to SPSSX-L), with no body text except the
> > command. To leave the list, send the command
> > SIGNOFF SPSSX-L
> > For a list of commands to manage subscriptions, send the command
> > INFO REFCARD
>
> ________________________________
> i'm is proud to present Cause Effect, a series about real people making a
> difference. Learn more

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Henrik Lolle]

Hi Courtney,

You could also use the GLM/UNIANOVA procedure where you will find a BY
statement which corresponds to the CLASS statement in SAS. You can't
ask for standardized regression coefficients here, but you probably
don't need them anyway in a situation with categorical predictors.

Best,
Henrik

************************************************************
Henrik Lolle
Department of Economics, Politics and Public Administration
Aalborg University
Fibigerstraede 1
9200 Aalborg East
Phone: [phone]
************************************************************



Quoting Keith McCormick <[hidden email]>:

> Hi,
>
> You could ... but it is going to amount to the same work. Doing so
> will still allow you to do the regression in the menus - as do I.
>
> In Transform ... Compute, you would put the new name, dummyvar1, in
> the box in the upper left and you would type the rest of the formula
> in the formula box on the right. In this case, it would be Major = 1.
> You will have to do this for each dummy variable.
>
> cheers, Keith
>
> On Dec 28, 2007 10:31 AM, Courtney Nichols
> <[hidden email]> wrote:
>>
>> Hi Keith,
>>  Thanks so much for your help.  Do you happen to know how to do the dummy
>> coding without using command syntax.  I've never used the syntax and just
>> use the point and click interface with the dropdown menus and such.  I can
>> learn how to use the program with syntax if needed, but thought I'd ask if
>> there is a way to do so with the menus.
>>  Many thanks,
>>
>>  Courtney
>>
>>
>>
>> "The grand essentials to happiness in this life are something to do,
>> something to love and something to hope for. " -- Joseph Addison
>>
>>
>>  ________________________________
>>  > Date: Fri, 28 Dec 2007 03:04:03 -0500
>> > From: [hidden email]
>> > Subject: Re: putting dummy variables into regression
>> > To: [hidden email]
>>
>>
>> >
>> > Hello Courtney,
>> >
>> > My favorite way to dummy code is the following syntax:
>> >
>> > COMPUTE dummy_var1 = (category_var=1).
>> >
>> > In practice, it might look like this:
>> >
>> > COMPUTE Biology_YN = (Major = 1).
>> > COMPUTE Math_YN = (Major = 2).
>> > etc.
>> >
>> > If you need an F-test for the whole set of variables (like all the
>> > majors), you can put the entire set in a "block". And request Change
>> > in R Sq. The Delta R Sq can serve as an overall test for the dummy
>> > variables. To use a block, click on the next button when you are in
>> > regression. It sounds like you have more than one categorical var, so
>> > you can add a block for each one. The sequence of the blocks will make
>> > a difference in the interpretation, so you will want to think that
>> > through.
>> >
>> > Feel free to follow-up, but that should get you started.
>> >
>> > Hope that helps.
>> >
>> > Keith
>> > www.keithmccormick.com
>> >
>> > On Dec 27, 2007 2:42 PM, Courtney Nichols <[hidden email]>
>> wrote:
>> > > Hello All,
>> > > I apologize, I feel like this should be very simple to figure out, but
>> I'm new to SPSS and regression analysis in general. I need to run a
>> regression where some of my predictor variables are continuous and others
>> are categorical. I'm having trouble figuring out how to include the
>> categorical and need help with two things:
>> > > 1) How do I get SPSS to code the categorical variables (ie education
>> level) into a collection of dummy variables? (I hear there is a "class
>> statement" command in SAS that does this but don't see how to do it in SPSS
>> and am also not using command syntax)
>> > > 2) Once I have that, what regression in the analyze menu do I run to get
>> output that gives me both an overall F-test that tells me whether the
>> overall categorical variable (ie education level) should remain in the model
>> and beta coefficients or t-tests to interpret for each dummy (ie high
>> school, some college, graduate, etc. in relation to the omitted comparison
>> group)?
>> > > Many thanks in advance!
>> > > Courtney
>> > > "The grand essentials to happiness in this life are something to do,
>> something to love and something to hope for. " -- Joseph Addison
>> > >
>> > > _________________________________________________________________
>> > > Get the power of Windows + Web with the new Windows Live.
>> > > http://www.windowslive.com?ocid=TXT_TAGHM_Wave2_powerofwindows_122007
>> > > To manage your subscription to SPSSX-L, send a message to
>> > > [hidden email] (not to SPSSX-L), with no body text except the
>> > > command. To leave the list, send the command
>> > > SIGNOFF SPSSX-L
>> > > For a list of commands to manage subscriptions, send the command
>> > > INFO REFCARD
>> > >
>> >
>> > =====================
>>
>> > To manage your subscription to SPSSX-L, send a message to
>> > [hidden email] (not to SPSSX-L), with no body text except the
>> > command. To leave the list, send the command
>> > SIGNOFF SPSSX-L
>> > For a list of commands to manage subscriptions, send the command
>> > INFO REFCARD
>>
>> ________________________________
>> i'm is proud to present Cause Effect, a series about real people making a
>> difference. Learn more
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Chen Wei-2]

Hello all,

I have a dataset of individual income from 10 cities, I would like to
compute Gini coefficient by city. How can I compute Gini coefficient in SPSS
and also by city? Anyone can help? Thanks very much.

Wayne

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Albert-Jan Roskam]

Hi Wayne,

Have a look at Ray's site:
http://www.spsstools.net/Syntax/Inequality/ManyTestsOfInequalityV5.txt

Cheers!!
Albert-Jan


--- Chen Wei <[hidden email]> wrote:

> Hello all,
>
> I have a dataset of individual income from 10
> cities, I would like to
> compute Gini coefficient by city. How can I compute
> Gini coefficient in SPSS
> and also by city? Anyone can help? Thanks very much.
>
> Wayne
>
> =====================
> To manage your subscription to SPSSX-L, send a
> message to
> [hidden email] (not to SPSSX-L), with no
> body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send
> the command
> INFO REFCARD
>

      ____________________________________________________________________________________
Never miss a thing.  Make Yahoo your home page.
http://www.yahoo.com/r/hs

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD