SPSSX Discussion

regression with interaction?

Classic

List

Threaded

21 messages Options

Paola Chivers-2

regression with interaction?

Hi,

I apologise for this simplistic question.

I need to work out how to test a simple regression with an interaction using SPSS. That is I have dependent variable X, independent variable Y and then an interaction WY. X and W are both continuous variables, while Y can be categorical or continuous.

Do I use the UNIANOVA command? Where DV=X, fixed factor = Y and covariate=W. See below.

UNIANOVA X BY Y WITH W

/METHOD=SSTYPE(3)

/INTERCEPT=INCLUDE

/CRITERIA=ALPHA(0.05)

/DESIGN=W*Y.

Thanks.

Paula.

Marta Garcia-Granero

Re: regression with interaction?

Hi Paola:

Don't forget the main effects too (hierarchical rule: if an interaction
term is present, the main effects involved should be there too).

UNIANOVA X BY Y WITH W
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/CRITERIA=ALPHA(0.05)
/DESIGN= X Y W*Y.

Marta GG

--
For miscellaneous SPSS related statistical stuff, visit:
http://gjyp.nl/marta/

=====================
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

Andrew Lawrence-2

Re: regression with interaction?

In reply to this post by Paola Chivers-2

Hi there,

If your independent (or predictor) variable Y is categorical then the
UNIANOVA command will give you a sensible answer, but as the other other
poster says you should include your main effects of W and X in the
design (/DESIGN=W Y W*Y.)

If you would rather treat Y as continuous (which is often more powerful
than artificially making a continuous variable categorical) then you
could investigate the interaction in a linear regression framework
(using the REGRESSION command) by computing the cross-product of W and Y
(e.g. COMPUTE YW = W*Y.) and entering 'YW' as a third predictor. This
may produce problems with co-linearity, so check for this, and centre
your variables if necessary. More information:
http://faculty.chass.ncsu.edu/garson/PA765/regress.htm

Hope this helps!

Andrew

Paola Chivers wrote:

>
> Hi,
>
>
>
> I apologise for this simplistic question.
>
>
>
> I need to work out how to test a simple regression with an interaction
> using SPSS. That is I have dependent variable X, independent variable
> Y and then an interaction WY. X and W are both continuous variables,
> while Y can be categorical or continuous.
>
>
>
> Do I use the UNIANOVA command? Where DV=X, fixed factor = Y and
> covariate=W. See below.
>
>
>
>
>
> UNIANOVA X BY Y WITH W
>
> /METHOD=SSTYPE(3)
>
> /INTERCEPT=INCLUDE
>
> /CRITERIA=ALPHA(0.05)
>
> /DESIGN=W*Y.
>
>
>
> Thanks.
>
> Paula.
>
> * *
>
>
>

--
Andrew J. Lawrence
Research Psychologist
Centre for Clinical Neuroscience
St George's University of London
Cranmer Terrace
London
SW17 0RE

[hidden email]
tel: +44(0)20 8266 6468
fax: +44(0)20 8725 2950

=====================
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

Paola Chivers-2

Re: regression with interaction?

Hi Marta and Andrew.

Thanks for you both for your advice.

What's the difference between doing the univariate ANOVA and the regression using the computed YW?

Regards,
Paola

“Ours has become a time-poor society, fatigued by non-physical demands and trying to compartmentalize daily living tasks. It is small wonder that physical activity is discarded in this environment” p126 (Steinbeck, 2001)

 Please consider the environment before printing this email.

-----Original Message-----
From: Andrew Lawrence [mailto:[hidden email]]
Sent: Thursday, 30 July 2009 11:16 PM
To: Paola Chivers
Cc: [hidden email]
Subject: Re: regression with interaction?

Hi there,

If your independent (or predictor) variable Y is categorical then the
UNIANOVA command will give you a sensible answer, but as the other other
poster says you should include your main effects of W and X in the
design (/DESIGN=W Y W*Y.)

If you would rather treat Y as continuous (which is often more powerful
than artificially making a continuous variable categorical) then you
could investigate the interaction in a linear regression framework
(using the REGRESSION command) by computing the cross-product of W and Y
(e.g. COMPUTE YW = W*Y.) and entering 'YW' as a third predictor. This
may produce problems with co-linearity, so check for this, and centre
your variables if necessary. More information:
http://faculty.chass.ncsu.edu/garson/PA765/regress.htm

Hope this helps!

Andrew

Hi Paola:

Don't forget the main effects too (hierarchical rule: if an interaction term is present, the main effects involved should be there too).

UNIANOVA X BY Y WITH W
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/CRITERIA=ALPHA(0.05)
/DESIGN= X Y W*Y.

Marta GG

--
For miscellaneous SPSS related statistical stuff, visit:
http://gjyp.nl/marta/

Paola Chivers wrote:

SR Millis-3

Re: regression with interaction?

Regression and ANOVA are the same thing.

SR Millis

--- On Thu, 7/30/09, Paola Chivers <[hidden email]> wrote:

> From: Paola Chivers <[hidden email]>
> Subject: Re: regression with interaction?
> To: [hidden email]
> Date: Thursday, July 30, 2009, 7:03 PM
> Hi Marta and Andrew.
>
> Thanks for you both for your advice.
>
> What's the difference between doing the univariate ANOVA
> and the regression using the computed YW?
>
> Regards,
> Paola
>
> “Ours has become a time-poor society, fatigued by
> non-physical demands and trying to compartmentalize daily
> living tasks.� It is small wonder that physical
> activity is discarded in this environment” p126
> (Steinbeck, 2001)
>
>  Please consider the environment before printing this
> email.
>
>
> -----Original Message-----
> From: Andrew Lawrence [mailto:[hidden email]]
> Sent: Thursday, 30 July 2009 11:16 PM
> To: Paola Chivers
> Cc: [hidden email]
> Subject: Re: regression with interaction?
>
> Hi there,
>
> If your independent (or predictor) variable Y is
> categorical then the
> UNIANOVA command will give you a sensible answer, but as
> the other other
> poster says you should include your main effects of W and X
> in the
> design (/DESIGN=W Y W*Y.)
>
> If you would rather treat Y as continuous (which is often
> more powerful
> than artificially making a continuous variable categorical)
> then you
> could investigate the interaction in a linear regression
> framework
> (using the REGRESSION command) by computing the
> cross-product of W and Y
> (e.g. COMPUTE YW = W*Y.) and entering 'YW' as a third
> predictor. This
> may produce problems with co-linearity, so check for this,
> and centre
> your variables if necessary. More information:
> http://faculty.chass.ncsu.edu/garson/PA765/regress.htm
>
> Hope this helps!
>
> Andrew
>
> Hi Paola:
>
> Don't forget the main effects too (hierarchical rule: if an
> interaction term is present, the main effects involved
> should be there too).
>
> UNIANOVA X BY Y� WITH W
> /METHOD=SSTYPE(3)
> /INTERCEPT=INCLUDE
> /CRITERIA=ALPHA(0.05)
> /DESIGN= X Y W*Y.
>
> Marta GG
>
> --
> For miscellaneous SPSS related statistical stuff, visit:
> http://gjyp.nl/marta/
>
>
> Paola Chivers wrote:
> >
> > Hi,
> >
> >
> >
> > I apologise for this simplistic question.
> >
> >
> >
> > I need to work out how to test a simple regression
> with an interaction
> > using SPSS.� That is I have dependent variable X,
> independent variable
> >� Y and then an interaction WY.� X and W are
> both continuous variables,
> > while Y can be categorical or continuous.
> >
> >
> >
> > Do I use the UNIANOVA command? Where DV=X, fixed
> factor = Y and
> > covariate=W.� � � See below.
> >
> >
> >
> >
> >
> > UNIANOVA X BY Y� WITH W
> >
> > /METHOD=SSTYPE(3)
> >
> > /INTERCEPT=INCLUDE
> >
> > /CRITERIA=ALPHA(0.05)
> >
> > /DESIGN=W*Y.
> >
> >
> >
> > Thanks.
> >
> > Paula.
> >
> > * *
> >
> >
> >
>
> --
> Andrew J. Lawrence
> Research Psychologist
> Centre for Clinical Neuroscience
> St George's University of London
> Cranmer Terrace
> London
> SW17 0RE
>
> [hidden email]
> tel: +44(0)20 8266 6468
> fax: +44(0)20 8725 2950
>
> =====================
> 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
>

SR Millis-3

Re: regression with interaction?

In reply to this post by Paola Chivers-2

Indeed. Think of regression and ANOVA as manifestations of the generalized linear model---with the identity link.

Scott R Millis, PhD, ABPP (CN,CL,RP), CStat, CSci
Professor & Director of Research
Dept of Physical Medicine & Rehabilitation
Dept of Emergency Medicine
Wayne State University School of Medicine
261 Mack Blvd
Detroit, MI 48201
Email: [hidden email]
Tel: 313-993-8085
Fax: 313-966-7682

--- On Thu, 7/30/09, Paola Chivers <[hidden email]> wrote:

> From: Paola Chivers <[hidden email]>
> Subject: RE: regression with interaction?
> To: "'SR Millis'" <[hidden email]>
> Date: Thursday, July 30, 2009, 7:59 PM
> Thanks ... I thought so ... but then
> thought maybe I had missed something.�
>
> Regards,
> Paola
>
> “Ours has become a time-poor society, fatigued by
> non-physical demands and trying to compartmentalize daily
> living tasks.� It is small wonder that physical
> activity is discarded in this environment” p126
> (Steinbeck, 2001)
>
>  Please consider the environment before printing this
> email.
>
>
>
> -----Original Message-----
> From: SR Millis [mailto:[hidden email]]
>
> Sent: Friday, 31 July 2009 7:53 AM
> To: Paola Chivers; SPSS
> Subject: Re: regression with interaction?
>
> Regression and ANOVA are the same thing.
>
> SR Millis
>
> --- On Thu, 7/30/09, Paola Chivers <[hidden email]>
> wrote:
>
> > From: Paola Chivers <[hidden email]>
> > Subject: Re: regression with interaction?
> > To: [hidden email]
> > Date: Thursday, July 30, 2009, 7:03 PM
> > Hi Marta and Andrew.
> >
> > Thanks for you both for your advice.
> >
> > What's the difference between doing the univariate
> ANOVA
> > and the regression using the computed YW?
> >
> > Regards,
> > Paola
> >
> > “Ours has become a time-poor society, fatigued by
> > non-physical demands and trying to compartmentalize
> daily
> > living tasks.� It is small wonder that physical
> > activity is discarded in this environment” p126
> > (Steinbeck, 2001)
> >
> >  Please consider the environment before printing
> this
> > email.
> >
> >
> > -----Original Message-----
> > From: Andrew Lawrence [mailto:[hidden email]]
> > Sent: Thursday, 30 July 2009 11:16 PM
> > To: Paola Chivers
> > Cc: [hidden email]
> > Subject: Re: regression with interaction?
> >
> > Hi there,
> >
> > If your independent (or predictor) variable Y is
> > categorical then the
> > UNIANOVA command will give you a sensible answer, but
> as
> > the other other
> > poster says you should include your main effects of W
> and X
> > in the
> > design (/DESIGN=W Y W*Y.)
> >
> > If you would rather treat Y as continuous (which is
> often
> > more powerful
> > than artificially making a continuous variable
> categorical)
> > then you
> > could investigate the interaction in a linear
> regression
> > framework
> > (using the REGRESSION command) by computing the
> > cross-product of W and Y
> > (e.g. COMPUTE YW = W*Y.) and entering 'YW' as a third
> > predictor. This
> > may produce problems with co-linearity, so check for
> this,
> > and centre
> > your variables if necessary. More information:
> > http://faculty.chass.ncsu.edu/garson/PA765/regress.htm
> >
> > Hope this helps!
> >
> > Andrew
> >
> > Hi Paola:
> >
> > Don't forget the main effects too (hierarchical rule:
> if an
> > interaction term is present, the main effects
> involved
> > should be there too).
> >
> > UNIANOVA X BY Y� WITH W
> > /METHOD=SSTYPE(3)
> > /INTERCEPT=INCLUDE
> > /CRITERIA=ALPHA(0.05)
> > /DESIGN= X Y W*Y.
> >
> > Marta GG
> >
> > --
> > For miscellaneous SPSS related statistical stuff,
> visit:
> > http://gjyp.nl/marta/
> >
> >
> > Paola Chivers wrote:
> > >
> > > Hi,
> > >
> > >
> > >
> > > I apologise for this simplistic question.
> > >
> > >
> > >
> > > I need to work out how to test a simple
> regression
> > with an interaction
> > > using SPSS.� That is I have dependent
> variable X,
> > independent variable
> > >� Y and then an interaction WY.� X and W
> are
> > both continuous variables,
> > > while Y can be categorical or continuous.
> > >
> > >
> > >
> > > Do I use the UNIANOVA command? Where DV=X, fixed
> > factor = Y and
> > > covariate=W.� � � See below.
> > >
> > >
> > >
> > >
> > >
> > > UNIANOVA X BY Y� WITH W
> > >
> > > /METHOD=SSTYPE(3)
> > >
> > > /INTERCEPT=INCLUDE
> > >
> > > /CRITERIA=ALPHA(0.05)
> > >
> > > /DESIGN=W*Y.
> > >
> > >
> > >
> > > Thanks.
> > >
> > > Paula.
> > >
> > > * *
> > >
> > >
> > >
> >
> > --
> > Andrew J. Lawrence
> > Research Psychologist
> > Centre for Clinical Neuroscience
> > St George's University of London
> > Cranmer Terrace
> > London
> > SW17 0RE
> >
> > [hidden email]
> > tel: +44(0)20 8266 6468
> > fax: +44(0)20 8725 2950
> >
> > =====================
> > 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
> >
>
>

Andrew Lawrence-2

Re: regression with interaction?

In reply to this post by Paola Chivers-2

The framework for the two is the same, but on a practical note the ANOVA
is set up for you to use if your Y variable is categorical (or has a
small number of discrete values, for example: gender, eye colour). The
regression is best if your Y variable is continuous (for example:
height, weight, age).

Another practical note, the ANOVA procedure uses type III sums of
squares as default, whilst regression uses type II. more on the
different sums of squares here: http://afni.nimh.nih.gov/sscc/gangc/SS.html

Regards,

Andrew

Paola Chivers wrote:

> Hi Marta and Andrew.
>
> Thanks for you both for your advice.
>
> What's the difference between doing the univariate ANOVA and the regression using the computed YW?
>
> Regards,
> Paola
>
> “Ours has become a time-poor society, fatigued by non-physical demands and trying to compartmentalize daily living tasks. It is small wonder that physical activity is discarded in this environment” p126 (Steinbeck, 2001)
>
>  Please consider the environment before printing this email.
>
>
> -----Original Message-----
> From: Andrew Lawrence [mailto:[hidden email]]
> Sent: Thursday, 30 July 2009 11:16 PM
> To: Paola Chivers
> Cc: [hidden email]
> Subject: Re: regression with interaction?
>
> Hi there,
>
> If your independent (or predictor) variable Y is categorical then the
> UNIANOVA command will give you a sensible answer, but as the other other
> poster says you should include your main effects of W and X in the
> design (/DESIGN=W Y W*Y.)
>
> If you would rather treat Y as continuous (which is often more powerful
> than artificially making a continuous variable categorical) then you
> could investigate the interaction in a linear regression framework
> (using the REGRESSION command) by computing the cross-product of W and Y
> (e.g. COMPUTE YW = W*Y.) and entering 'YW' as a third predictor. This
> may produce problems with co-linearity, so check for this, and centre
> your variables if necessary. More information:
> http://faculty.chass.ncsu.edu/garson/PA765/regress.htm
>
> Hope this helps!
>
> Andrew
>
> Hi Paola:
>
> Don't forget the main effects too (hierarchical rule: if an interaction term is present, the main effects involved should be there too).
>
> UNIANOVA X BY Y WITH W
> /METHOD=SSTYPE(3)
> /INTERCEPT=INCLUDE
> /CRITERIA=ALPHA(0.05)
> /DESIGN= X Y W*Y.
>
> Marta GG
>
> --
> For miscellaneous SPSS related statistical stuff, visit:
> http://gjyp.nl/marta/
>
>
> Paola Chivers wrote:
>
>> Hi,
>>
>>
>>
>> I apologise for this simplistic question.
>>
>>
>>
>> I need to work out how to test a simple regression with an interaction
>> using SPSS. That is I have dependent variable X, independent variable
>> Y and then an interaction WY. X and W are both continuous variables,
>> while Y can be categorical or continuous.
>>
>>
>>
>> Do I use the UNIANOVA command? Where DV=X, fixed factor = Y and
>> covariate=W. See below.
>>
>>
>>
>>
>>
>> UNIANOVA X BY Y WITH W
>>
>> /METHOD=SSTYPE(3)
>>
>> /INTERCEPT=INCLUDE
>>
>> /CRITERIA=ALPHA(0.05)
>>
>> /DESIGN=W*Y.
>>
>>
>>
>> Thanks.
>>
>> Paula.
>>
>> * *
>>
>>
>>
>>
>
> --
> Andrew J. Lawrence
> Research Psychologist
> Centre for Clinical Neuroscience
> St George's University of London
> Cranmer Terrace
> London
> SW17 0RE
>
> [hidden email]
> tel: +44(0)20 8266 6468
> fax: +44(0)20 8725 2950
>
> =====================
> 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
>
>

Mary Oliver

Re: regression with interaction?

In reply to this post by Paola Chivers-2

Hi All,

Paola, you might also want to check out this page for some very cool scripts
and macros developed by Andrew Hayes at OSU that will help you do simple
slopes analysis associated with a significant interaction. They're
wonderful, wonderful tools!

http://www.comm.ohio-state.edu/ahayes/SPSS%20programs/modprobe.htm

With kind regards,

Mary

On Fri, 31 Jul 2009 07:03:25 +0800, Paola Chivers <[hidden email]> wrote:

>Hi Marta and Andrew.
>
>Thanks for you both for your advice.
>
>What's the difference between doing the univariate ANOVA and the regression
using the computed YW?
>
>Regards,
>Paola
>
>âOurs has become a time-poor society, fatigued by non-physical demands and
trying to compartmentalize daily living tasks. It is small wonder that
physical activity is discarded in this environmentâ p126 (Steinbeck, 2001)

>
>ï Please consider the environment before printing this email.
>
>
>-----Original Message-----
>From: Andrew Lawrence [mailto:[hidden email]]
>Sent: Thursday, 30 July 2009 11:16 PM
>To: Paola Chivers
>Cc: [hidden email]
>Subject: Re: regression with interaction?
>
>Hi there,
>
>If your independent (or predictor) variable Y is categorical then the
>UNIANOVA command will give you a sensible answer, but as the other other
>poster says you should include your main effects of W and X in the
>design (/DESIGN=W Y W*Y.)
>
>If you would rather treat Y as continuous (which is often more powerful
>than artificially making a continuous variable categorical) then you
>could investigate the interaction in a linear regression framework
>(using the REGRESSION command) by computing the cross-product of W and Y
>(e.g. COMPUTE YW = W*Y.) and entering 'YW' as a third predictor. This
>may produce problems with co-linearity, so check for this, and centre
>your variables if necessary. More information:
>http://faculty.chass.ncsu.edu/garson/PA765/regress.htm
>
>Hope this helps!
>
>Andrew
>
>Hi Paola:
>
>Don't forget the main effects too (hierarchical rule: if an interaction

term is present, the main effects involved should be there too).

>
>UNIANOVA X BY Y WITH W
>/METHOD=SSTYPE(3)
>/INTERCEPT=INCLUDE
>/CRITERIA=ALPHA(0.05)
>/DESIGN= X Y W*Y.
>
>Marta GG
>
>--
>For miscellaneous SPSS related statistical stuff, visit:
>http://gjyp.nl/marta/
>
>
>Paola Chivers wrote:
>>
>> Hi,
>>
>>
>>
>> I apologise for this simplistic question.
>>
>>
>>
>> I need to work out how to test a simple regression with an interaction
>> using SPSS. That is I have dependent variable X, independent variable
>> Y and then an interaction WY. X and W are both continuous variables,
>> while Y can be categorical or continuous.
>>
>>
>>
>> Do I use the UNIANOVA command? Where DV=X, fixed factor = Y and
>> covariate=W. See below.
>>
>>
>>
>>
>>
>> UNIANOVA X BY Y WITH W
>>
>> /METHOD=SSTYPE(3)
>>
>> /INTERCEPT=INCLUDE
>>
>> /CRITERIA=ALPHA(0.05)
>>
>> /DESIGN=W*Y.
>>
>>
>>
>> Thanks.
>>
>> Paula.
>>
>> * *
>>
>>
>>
>
>--
>Andrew J. Lawrence
>Research Psychologist
>Centre for Clinical Neuroscience
>St George's University of London
>Cranmer Terrace
>London
>SW17 0RE
>
>[hidden email]
>tel: +44(0)20 8266 6468
>fax: +44(0)20 8725 2950
>
>=====================
>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

SR Millis-3

Re: regression with interaction?

In reply to this post by Paola Chivers-2

Excellent observation, Michael. I like your characterization of "surface" and "deep" levels. When teaching, I like to introduce the concept of the generalized linear model as the central unifying theme for all models--as early as possible.

Scott Millis

--- On Fri, 7/31/09, Granaas, Michael <[hidden email]> wrote:

> From: Granaas, Michael <[hidden email]>
> Subject: RE: regression with interaction?
> To: "SR Millis" <[hidden email]>
> Date: Friday, July 31, 2009, 10:53 AM
>
> They are indeed the same thing at a "deep" level. But
> they have profoundly different surface structures. I
> find that those training is based only on the surface
> structure can be challenged by the notion that they are the
> same thing.
>
> I prefer to tell folks that ANOVA is an algebraic special
> case of the multiple regression model and go from
> there.
>
> To the OP, if you are using a regression framework and your
> categorical predictor has more than two levels you'll want
> to use some sort of coding scheme such as effect, dummy, or
> orthogonal in order to pull the multiple predictors out of
> the predictor.
>
> Michael
>
>
>
> ****************************************************
> Michael Granaas
> [hidden email]
> Assoc. Prof.
> Phone: 605 677 5295
> Dept. of Psychology
> FAX: 605 677 3195
> University of South Dakota
> 414 E. Clark St.
> Vermillion, SD 57069
> *****************************************************
>
>
>
>
> -----Original Message-----
> From: SPSSX(r) Discussion on behalf of SR Millis
> Sent: Thu 7/30/09 6:53 PM
> To: [hidden email]
> Subject: Re: regression with interaction?
>
> Regression and ANOVA are the same thing.
>
> SR Millis
>
> --- On Thu, 7/30/09, Paola Chivers <[hidden email]>
> wrote:
>
> > From: Paola Chivers <[hidden email]>
> > Subject: Re: regression with interaction?
> > To: [hidden email]
> > Date: Thursday, July 30, 2009, 7:03 PM
> > Hi Marta and Andrew.
> >
> > Thanks for you both for your advice.
> >
> > What's the difference between doing the univariate
> ANOVA
> > and the regression using the computed YW?
> >
> > Regards,
> > Paola
> >
> > "Ours has become a time-poor society, fatigued by
> > non-physical demands and trying to compartmentalize
> daily
> > living tasks. It is small wonder that physical
> > activity is discarded in this environment" p126
> > (Steinbeck, 2001)
> >
> > ? Please consider the environment before printing
> this
> > email.
> >
> >
> > -----Original Message-----
> > From: Andrew Lawrence [mailto:[hidden email]]
> > Sent: Thursday, 30 July 2009 11:16 PM
> > To: Paola Chivers
> > Cc: [hidden email]
> > Subject: Re: regression with interaction?
> >
> > Hi there,
> >
> > If your independent (or predictor) variable Y is
> > categorical then the
> > UNIANOVA command will give you a sensible answer, but
> as
> > the other other
> > poster says you should include your main effects of W
> and X
> > in the
> > design (/DESIGN=W Y W*Y.)
> >
> > If you would rather treat Y as continuous (which is
> often
> > more powerful
> > than artificially making a continuous variable
> categorical)
> > then you
> > could investigate the interaction in a linear
> regression
> > framework
> > (using the REGRESSION command) by computing the
> > cross-product of W and Y
> > (e.g. COMPUTE YW = W*Y.) and entering 'YW' as a third
> > predictor. This
> > may produce problems with co-linearity, so check for
> this,
> > and centre
> > your variables if necessary. More information:
> > http://faculty.chass.ncsu.edu/garson/PA765/regress.htm
> >
> > Hope this helps!
> >
> > Andrew
> >
> > Hi Paola:
> >
> > Don't forget the main effects too (hierarchical rule:
> if an
> > interaction term is present, the main effects
> involved
> > should be there too).
> >
> > UNIANOVA X BY Y WITH W
> > /METHOD=SSTYPE(3)
> > /INTERCEPT=INCLUDE
> > /CRITERIA=ALPHA(0.05)
> > /DESIGN= X Y W*Y.
> >
> > Marta GG
> >
> > --
> > For miscellaneous SPSS related statistical stuff,
> visit:
> > http://gjyp.nl/marta/
> >
> >
> > Paola Chivers wrote:
> > >
> > > Hi,
> > >
> > >
> > >
> > > I apologise for this simplistic question.
> > >
> > >
> > >
> > > I need to work out how to test a simple
> regression
> > with an interaction
> > > using SPSS. That is I have dependent
> variable X,
> > independent variable
> > > Y and then an interaction WY. X and W
> are
> > both continuous variables,
> > > while Y can be categorical or continuous.
> > >
> > >
> > >
> > > Do I use the UNIANOVA command? Where DV=X, fixed
> > factor = Y and
> > > covariate=W. See below.
> > >
> > >
> > >
> > >
> > >
> > > UNIANOVA X BY Y WITH W
> > >
> > > /METHOD=SSTYPE(3)
> > >
> > > /INTERCEPT=INCLUDE
> > >
> > > /CRITERIA=ALPHA(0.05)
> > >
> > > /DESIGN=W*Y.
> > >
> > >
> > >
> > > Thanks.
> > >
> > > Paula.
> > >
> > > * *
> > >
> > >
> > >
> >
> > --
> > Andrew J. Lawrence
> > Research Psychologist
> > Centre for Clinical Neuroscience
> > St George's University of London
> > Cranmer Terrace
> > London
> > SW17 0RE
> >
> > [hidden email]
> > tel: +44(0)20 8266 6468
> > fax: +44(0)20 8725 2950
> >
> > =====================
> > 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
>
>

johnleeholmes@gmail.com

Re: odd problem in spss with variable name

There's a variable in our dataset called path. Spss (version 15) refuses to work with that variable unless I select it in the Variable View and rename it manually first. I cannot manage to rename it with syntax. I did not think "path" was a system reserved command and SPSS should automatically rename reserved names. Perhaps it's a glitch in my copy of spss?

Any ideas?

JL

ViAnn Beadle

Re: odd problem in SPSS with variable name

Path is not a reserved word. What do you mean when you say that SPSS refuses to work with it—what are the symptoms? Select it in the Descriptives dialog box, paste the syntax and run it. What happens?

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of [hidden email]
Sent: Friday, July 31, 2009 4:05 PM
To: [hidden email]
Subject: Re: odd problem in spss with variable name

johnleeholmes@gmail.com

Re: odd problem in SPSS with variable name

Wait I realize why now: we are using a path macro called path. So, hand smack to the forehead, that explains it! Sorry about that, but thanks for your kind offer of assistance, and have a splendid weekend.

JL

ViAnn Beadle wrote:

Path is not a reserved word. What do you mean when you say that SPSS refuses to work with it—what are the symptoms? Select it in the Descriptives dialog box, paste the syntax and run it. What happens?

From: SPSSX(r) Discussion [[hidden email]] On Behalf Of [hidden email]
Sent: Friday, July 31, 2009 4:05 PM
To: [hidden email]
Subject: Re: odd problem in spss with variable name

There's a variable in our dataset called path. Spss (version 15) refuses to work with that variable unless I select it in the Variable View and rename it manually first. I cannot manage to rename it with syntax. I did not think "path" was a system reserved command and SPSS should automatically rename reserved names. Perhaps it's a glitch in my copy of spss?

Any ideas?

JL

Okay, doing exactly as you specified, spss pastes this into the syntax window:
DESCRIPTIVES
VARIABLES=path
/STATISTICS=MEAN STDDEV MIN MAX .

When I run it the output tosses out some version of this message:
Warnings
Text: \\PROJECTS\projname\Data\Pretest 1\
Unrecognized text was found where a variable list was expected.
This command not executed.

No Variables subcommand.

So it is as if it is inserting the path of the file instead of recognizing it as a variable.

Note that the normal error message would be something like this:
Warnings
Text: path33
An undefined variable name, or a scratch or system variable was specified in a variable list which accepts only standard variables. Check spelling and verify the existence of this variable.
This command not executed.

No Variables subcommand.

Daciuk, Tim

Re: odd problem in spss with variable name

In reply to this post by johnleeholmes@gmail.com

I am currently working with a dataset that has a variable called path. I have not encountered any problems using it in SPSS 17 and 18. What do you get if you run a frequencies from the dialog boxes?

Tim Daciuk

Director, Worldwide Demo Resources

SPSS Inc.

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of [hidden email]
Sent: Friday, July 31, 2009 6:05 PM
To: [hidden email]
Subject: Re: odd problem in spss with variable name

Albert-Jan Roskam

Can anybody help me with this Spanish text? (NHS)

Hi, or buenos dias!

I am trying to find documentation on the Spanish NHS of 2001, in particular on the (non-)response rate. I found the text fragment below somewhere. Does it indicate that 15 % of all participants did not respond? I hardly understand any Spanish (just some French, which helps a little). Can anybody help me with this? The people I'd normally ask are on holidays, I think.

Los datos proceden de la Encuesta Nacional de Salud (ENS) realizada por el Ministerio de Sanidad y Consumo en 2001. En esta encuesta se entrevistó a una muestra domiciliaria de 21.120 personas representativa de la población española no institucionalizada ≥ 16 años. El procedimiento muestral fue polietápico, estratificado por conglomerados, con selección de las unidades de primera etapa -los municipios- y las de segunda etapa -las secciones censales- de forma aleatoria proporcional al tamaño de la población. Los individuos fueron elegidos por cuotas de edad y sexo. Un 15% de los sujetos inicialmente seleccionados no contestó a la encuesta. El error muestral máximo para las estimaciones estatales para un intervalo de confianza (IC) del 95% fue del 1,38%.

Thanks very much in advance!

Albert-Jan

--- On Sat, 8/1/09, Daciuk, Tim <[hidden email]> wrote:

> From: Daciuk, Tim <[hidden email]>
> Subject: Re: odd problem in spss with variable name
> To: [hidden email]
> Date: Saturday, August 1, 2009, 12:32 AM
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> I am
> currently working with a dataset that
> has a variable called path. I have not encountered
> any problems using it in
> SPSS 17 and 18. What do you get if you run a
> frequencies from the dialog
> boxes?
>
>
>
>
>
>
> Tim
> Daciuk
>
> Director, Worldwide
> Demo Resources
>
> SPSS
> Inc.
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> From:
> SPSSX(r) Discussion [mailto:[hidden email]]
> On Behalf Of
> [hidden email]
>
> Sent: Friday,
> July 31, 2009 6:05
> PM
>
> To:
> [hidden email]
>
> Subject: Re:
> odd problem in spss
> with variable name
>
>
>
>
>
> There's a variable in our
> dataset called path. Spss (version 15) refuses to
> work with that variable
> unless I select it in the Variable View and rename it
> manually first. I
> cannot manage to rename it with syntax. I did not
> think "path"
> was a system reserved command and SPSS should automatically
> rename reserved
> names. Perhaps it's a glitch in my copy of spss?
>
>
>
> Any ideas?
>
>
>
> JL
>
>
>
>
>
>
>

Kylie

Re: Can anybody help me with this Spanish text? (NHS)

Hi Albert-Jan,

Hopefully a Spanish speaker will be able to confirm, but a web translation
(babelfish) seems to support your interpretation:

The data come from the National Survey of Salud (ENS) realised by the
Ministry of Health and Consumption in 2001. In this survey entrevistó to
representative a domiciliary sample of 21,120 people of noninstitutionalized
the Spanish population & #8805; 16 years. The procedure sample was
multi-stage, stratified by conglomerates, with selection of the units of
first stage - the municipalities and those of second stage - the sections
censales- of proportional random form to the size of the population. The
individuals were chosen by quotas of age and sex. A 15% of the subjects
initially selected did not answer the survey. The error maximum sample for
the state estimations for a confidence interval (IC) of 95% was of 1.38%.

Hope this helps.

Cheers,
Kylie.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Albert-Jan Roskam
Sent: Saturday, 1 August 2009 8:41 PM
To: [hidden email]
Subject: Can anybody help me with this Spanish text? (NHS)

Hi, or buenos dias!

I am trying to find documentation on the Spanish NHS of 2001, in particular
on the (non-)response rate. I found the text fragment below somewhere. Does
it indicate that 15 % of all participants did not respond? I hardly
understand any Spanish (just some French, which helps a little). Can anybody
help me with this? The people I'd normally ask are on holidays, I think.

Los datos proceden de la Encuesta Nacional de Salud (ENS) realizada por el
Ministerio de Sanidad y Consumo en 2001. En esta encuesta se entrevistó a
una muestra domiciliaria de 21.120 personas representativa de la población
española no institucionalizada ≥ 16 años. El procedimiento muestral
fue polietápico, estratificado por conglomerados, con selección de las
unidades de primera etapa -los municipios- y las de segunda etapa -las
secciones censales- de forma aleatoria proporcional al tamaño de la
población. Los individuos fueron elegidos por cuotas de edad y sexo. Un 15%
de los sujetos inicialmente seleccionados no contestó a la encuesta. El
error muestral máximo para las estimaciones estatales para un intervalo de
confianza (IC) del 95% fue del 1,38%.

Thanks very much in advance!

Albert-Jan

--- On Sat, 8/1/09, Daciuk, Tim <[hidden email]> wrote:

Hector Maletta

Re: Can anybody help me with this Spanish text? (NHS)

In reply to this post by Albert-Jan Roskam

A small style correction in the translation: From where you put "The
procedure sample" onwards, a better rendering would be:
"A multistage, stratified clustered sampling design was used, with random
selection proportional to size of first-order units (municipalities) and
second order units (census tracts). Individuals were selected according to
age-sex quotas. Fifteen percent of initially selected subjects failed to
respond. Maximum margin of error for the State-level estimates was 1.38% (at
a 95% confidence level)."

The last sentence is not altogether clear in the Spanish original. It seems
to refer to the maximum margin of error (i.e. the interval representing 1.96
standard errors around the mean) for various variables, stating that the
maximum of those margins was just 1.38%. It is not clear whether this means
+/- 1.38 percentage points around a percentage mean, or +/- 1.38% of the
mean value. For a mean of 50%, the first interpretation puts the maximum
interval from 50-1.38=48.62 to 50+1.38=51.38; the second interpretation
would imply an interval from 50*(1-0.0138) to 50*(1+0.0138), i.e. from 49.31
to 50.69. You may not care about this ambiguity, and leave it in your
translation as ambiguous as it is in the original.

Hector
-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Kylie Lange
Sent: 02 August 2009 19:18
To: [hidden email]
Subject: Re: Can anybody help me with this Spanish text? (NHS)

Hi Albert-Jan,

Hopefully a Spanish speaker will be able to confirm, but a web translation
(babelfish) seems to support your interpretation:

The data come from the National Survey of Salud (ENS) realised by the
Ministry of Health and Consumption in 2001. In this survey entrevistó to
representative a domiciliary sample of 21,120 people of noninstitutionalized
the Spanish population & #8805; 16 years. The procedure sample was
multi-stage, stratified by conglomerates, with selection of the units of
first stage - the municipalities and those of second stage - the sections
censales- of proportional random form to the size of the population. The
individuals were chosen by quotas of age and sex. A 15% of the subjects
initially selected did not answer the survey. The error maximum sample for
the state estimations for a confidence interval (IC) of 95% was of 1.38%.

Hope this helps.

Cheers,
Kylie.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Albert-Jan Roskam
Sent: Saturday, 1 August 2009 8:41 PM
To: [hidden email]
Subject: Can anybody help me with this Spanish text? (NHS)

Hi, or buenos dias!

I am trying to find documentation on the Spanish NHS of 2001, in particular
on the (non-)response rate. I found the text fragment below somewhere. Does
it indicate that 15 % of all participants did not respond? I hardly
understand any Spanish (just some French, which helps a little). Can anybody
help me with this? The people I'd normally ask are on holidays, I think.

Los datos proceden de la Encuesta Nacional de Salud (ENS) realizada por el
Ministerio de Sanidad y Consumo en 2001. En esta encuesta se entrevistó a
una muestra domiciliaria de 21.120 personas representativa de la población
española no institucionalizada ≥ 16 años. El procedimiento muestral
fue polietápico, estratificado por conglomerados, con selección de las
unidades de primera etapa -los municipios- y las de segunda etapa -las
secciones censales- de forma aleatoria proporcional al tamaño de la
población. Los individuos fueron elegidos por cuotas de edad y sexo. Un 15%
de los sujetos inicialmente seleccionados no contestó a la encuesta. El
error muestral máximo para las estimaciones estatales para un intervalo de
confianza (IC) del 95% fue del 1,38%.

Thanks very much in advance!

Albert-Jan

--- On Sat, 8/1/09, Daciuk, Tim <[hidden email]> wrote:

=====================
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
No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 8.5.392 / Virus Database: 270.13.37/2273 - Release Date: 08/02/09
17:56:00

=====================
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

Marta Garcia-Granero

Re: Can anybody help me with this Spanish text? (NHS)

My twopence

Hector Maletta wrote:

> A small style correction in the translation: From where you put "The
> procedure sample" onwards, a better rendering would be:
> "A multistage, stratified clustered sampling design was used, with random
> selection proportional to size of first-order units (municipalities) and
> second order units (census tracts). Individuals were selected according to
> age-sex quotas. Fifteen percent of initially selected subjects failed to
> respond. Maximum margin of error for the State-level estimates was 1.38% (at
> a 95% confidence level)."
>
> The last sentence is not altogether clear in the Spanish original. It seems
> to refer to the maximum margin of error (i.e. the interval representing 1.96
> standard errors around the mean) for various variables, stating that the
> maximum of those margins was just 1.38%. It is not clear whether this means
> +/- 1.38 percentage points around a percentage mean, or +/- 1.38% of the
> mean value. For a mean of 50%, the first interpretation puts the maximum
> interval from 50-1.38=48.62 to 50+1.38=51.38; the second interpretation
> would imply an interval from 50*(1-0.0138) to 50*(1+0.0138), i.e. from 49.31
> to 50.69. You may not care about this ambiguity, and leave it in your
> translation as ambiguous as it is in the original.
>

From my experience (being used to reading these kind of reports), the
first interpretation Hector gives is the correct one. The maximum error
is usually computed as 100*(1.96*SQRT(1/4*n)), assuming the maximum
variance (when p=q=0.5, therefore p(1-p)=0.25=1/4).

HTH,
Marta GG

--
For miscellaneous SPSS related statistical stuff, visit:
http://gjyp.nl/marta/

=====================
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

Re: Can anybody help me with this Spanish text? (NHS)

In reply to this post by Hector Maletta

Dear Hector, Marta, Kylie, Pablo and others who replied:

Thank you very much [muchos gracios] for helping me. Much appreciated!

Best wishes,
Albert-Jan

--- On Mon, 8/3/09, Hector Maletta <[hidden email]> wrote:

> From: Hector Maletta <[hidden email]>
> Subject: Re: Can anybody help me with this Spanish text? (NHS)
> To: [hidden email]
> Date: Monday, August 3, 2009, 3:04 AM
> A small style correction in the
> translation: From where you put "The
> procedure sample" onwards, a better rendering would be:
> "A multistage, stratified clustered sampling design was
> used, with random
> selection proportional to size of first-order units
> (municipalities) and
> second order units (census tracts). Individuals were
> selected according to
> age-sex quotas. Fifteen percent of initially selected
> subjects failed to
> respond. Maximum margin of error for the State-level
> estimates was 1.38% (at
> a 95% confidence level)."
>
> The last sentence is not altogether clear in the Spanish
> original. It seems
> to refer to the maximum margin of error (i.e. the interval
> representing 1.96
> standard errors around the mean) for various variables,
> stating that the
> maximum of those margins was just 1.38%. It is not clear
> whether this means
> +/- 1.38 percentage points around a percentage mean, or +/-
> 1.38% of the
> mean value. For a mean of 50%, the first interpretation
> puts the maximum
> interval from 50-1.38=48.62 to 50+1.38=51.38; the second
> interpretation
> would imply an interval from 50*(1-0.0138) to
> 50*(1+0.0138), i.e. from 49.31
> to 50.69. You may not care about this ambiguity, and leave
> it in your
> translation as ambiguous as it is in the original.
>
> Hector
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]]
> On Behalf Of
> Kylie Lange
> Sent: 02 August 2009 19:18
> To: [hidden email]
> Subject: Re: Can anybody help me with this Spanish text?
> (NHS)
>
> Hi Albert-Jan,
>
> Hopefully a Spanish speaker will be able to confirm, but a
> web translation
> (babelfish) seems to support your interpretation:
>
> The data come from the National Survey of Salud (ENS)
> realised by the
> Ministry of Health and Consumption in 2001. In this survey
> entrevistó to
> representative a domiciliary sample of 21,120 people of
> noninstitutionalized
> the Spanish population & #8805; 16 years. The procedure
> sample was
> multi-stage, stratified by conglomerates, with selection of
> the units of
> first stage - the municipalities and those of second stage
> - the sections
> censales- of proportional random form to the size of the
> population. The
> individuals were chosen by quotas of age and sex. A 15% of
> the subjects
> initially selected did not answer the survey. The error
> maximum sample for
> the state estimations for a confidence interval (IC) of 95%
> was of 1.38%.
>
> Hope this helps.
>
> Cheers,
> Kylie.
>
>
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]]
> On Behalf Of
> Albert-Jan Roskam
> Sent: Saturday, 1 August 2009 8:41 PM
> To: [hidden email]
> Subject: Can anybody help me with this Spanish text? (NHS)
>
> Hi, or buenos dias!
>
> I am trying to find documentation on the Spanish NHS of
> 2001, in particular
> on the (non-)response rate. I found the text fragment below
> somewhere. Does
> it indicate that 15 % of all participants did not respond?
> I hardly
> understand any Spanish (just some French, which helps a
> little). Can anybody
> help me with this? The people I'd normally ask are on
> holidays, I think.
>
> Los datos proceden de la Encuesta Nacional� de Salud
> (ENS) realizada por el
> Ministerio de Sanidad y Consumo en 2001. En esta�
> encuesta se entrevistó a
> una muestra domiciliaria de 21.120 personas�
> representativa de la población
> española no institucionalizada ≥ 16 años. El
> procedimiento muestral
> fue polietápico,� estratificado por conglomerados,
> con selección de las
> unidades de primera etapa� -los municipios- y las de
> segunda etapa -las
> secciones censales-� de forma aleatoria proporcional
> al tamaño de la
> población. Los individuos fueron� elegidos por cuotas
> de edad y sexo. Un 15%
> de los sujetos inicialmente� seleccionados no
> contestó a la encuesta. El
> error muestral máximo para las� estimaciones
> estatales para un intervalo de
> confianza (IC) del 95% fue del� 1,38%.
>
> Thanks very much in advance!
>
> Albert-Jan
>
> --- On Sat, 8/1/09, Daciuk, Tim <[hidden email]>
> wrote:
>
> > From: Daciuk, Tim <[hidden email]>
> > Subject: Re: odd problem in spss with variable name
> > To: [hidden email]
> > Date: Saturday, August 1, 2009, 12:32 AM
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > I am
> > currently working with a dataset that
> > has a variable called path.� I have not
> encountered
> > any problems using it in
> > SPSS 17 and 18.� What do you get if you run a
> > frequencies from the dialog
> > boxes?
> >
> >
> >
> >
> >
> >
> > Tim
> > Daciuk
> >
> > Director, Worldwide
> > Demo Resources
> >
> > SPSS
> > Inc.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > From:
> > SPSSX(r) Discussion [mailto:[hidden email]]
> > On Behalf Of
> > [hidden email]
> >
> > Sent: Friday,
> > July 31, 2009 6:05
> > PM
> >
> > To:
> > [hidden email]
> >
> > Subject: Re:
> > odd problem in spss
> > with variable name
> >
> >
> >
> >
> >
> > There's a variable in our
> > dataset called path.� Spss (version 15) refuses
> to
> > work with that variable
> > unless I select it in the Variable View and rename it
> > manually first.� I
> > cannot manage to rename it with syntax.� I did
> not
> > think "path"
> > was a system reserved command and SPSS should
> automatically
> > rename reserved
> > names.� Perhaps it's a glitch in my copy of
> spss?
> >
> >
> >
> > Any ideas?
> >
> >
> >
> > JL
> >
> >
> >
> >
> >
> >
> >
>
> =====================
> 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
> No virus found in this incoming message.
> Checked by AVG - www.avg.com
> Version: 8.5.392 / Virus Database: 270.13.37/2273 - Release
> Date: 08/02/09
> 17:56:00
>
> =====================
> 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
>

Hector Maletta

Re: Can anybody help me with this Spanish text? (NHS)

Albert-Jan, you are welcome. However, you better say "muchas gracias" and not "muchos gracios"...
Hector

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Albert-Jan Roskam
Sent: 03 August 2009 05:48
To: [hidden email]
Subject: Re: Can anybody help me with this Spanish text? (NHS)

Dear Hector, Marta, Kylie, Pablo and others who replied:

Thank you very much [muchos gracios] for helping me. Much appreciated!

Best wishes,
Albert-Jan

--- On Mon, 8/3/09, Hector Maletta <[hidden email]> wrote:

> From: Hector Maletta <[hidden email]>
> Subject: Re: Can anybody help me with this Spanish text? (NHS)
> To: [hidden email]
> Date: Monday, August 3, 2009, 3:04 AM
> A small style correction in the
> translation: From where you put "The
> procedure sample" onwards, a better rendering would be:
> "A multistage, stratified clustered sampling design was
> used, with random
> selection proportional to size of first-order units
> (municipalities) and
> second order units (census tracts). Individuals were
> selected according to
> age-sex quotas. Fifteen percent of initially selected
> subjects failed to
> respond. Maximum margin of error for the State-level
> estimates was 1.38% (at
> a 95% confidence level)."
>
> The last sentence is not altogether clear in the Spanish
> original. It seems
> to refer to the maximum margin of error (i.e. the interval
> representing 1.96
> standard errors around the mean) for various variables,
> stating that the
> maximum of those margins was just 1.38%. It is not clear
> whether this means
> +/- 1.38 percentage points around a percentage mean, or +/-
> 1.38% of the
> mean value. For a mean of 50%, the first interpretation
> puts the maximum
> interval from 50-1.38=48.62 to 50+1.38=51.38; the second
> interpretation
> would imply an interval from 50*(1-0.0138) to
> 50*(1+0.0138), i.e. from 49.31
> to 50.69. You may not care about this ambiguity, and leave
> it in your
> translation as ambiguous as it is in the original.
>
> Hector
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]]
> On Behalf Of
> Kylie Lange
> Sent: 02 August 2009 19:18
> To: [hidden email]
> Subject: Re: Can anybody help me with this Spanish text?
> (NHS)
>
> Hi Albert-Jan,
>
> Hopefully a Spanish speaker will be able to confirm, but a
> web translation
> (babelfish) seems to support your interpretation:
>
> The data come from the National Survey of Salud (ENS)
> realised by the
> Ministry of Health and Consumption in 2001. In this survey
> entrevistó to
> representative a domiciliary sample of 21,120 people of
> noninstitutionalized
> the Spanish population & #8805; 16 years. The procedure
> sample was
> multi-stage, stratified by conglomerates, with selection of
> the units of
> first stage - the municipalities and those of second stage
> - the sections
> censales- of proportional random form to the size of the
> population. The
> individuals were chosen by quotas of age and sex. A 15% of
> the subjects
> initially selected did not answer the survey. The error
> maximum sample for
> the state estimations for a confidence interval (IC) of 95%
> was of 1.38%.
>
> Hope this helps.
>
> Cheers,
> Kylie.
>
>
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]]
> On Behalf Of
> Albert-Jan Roskam
> Sent: Saturday, 1 August 2009 8:41 PM
> To: [hidden email]
> Subject: Can anybody help me with this Spanish text? (NHS)
>
> Hi, or buenos dias!
>
> I am trying to find documentation on the Spanish NHS of
> 2001, in particular
> on the (non-)response rate. I found the text fragment below
> somewhere. Does
> it indicate that 15 % of all participants did not respond?
> I hardly
> understand any Spanish (just some French, which helps a
> little). Can anybody
> help me with this? The people I'd normally ask are on
> holidays, I think.
>
> Los datos proceden de la Encuesta Nacional de Salud
> (ENS) realizada por el
> Ministerio de Sanidad y Consumo en 2001. En esta
> encuesta se entrevistó a
> una muestra domiciliaria de 21.120 personas
> representativa de la población
> española no institucionalizada ≥ 16 años. El
> procedimiento muestral
> fue polietápico, estratificado por conglomerados,
> con selección de las
> unidades de primera etapa -los municipios- y las de
> segunda etapa -las
> secciones censales- de forma aleatoria proporcional
> al tamaño de la
> población. Los individuos fueron elegidos por cuotas
> de edad y sexo. Un 15%
> de los sujetos inicialmente seleccionados no
> contestó a la encuesta. El
> error muestral máximo para las estimaciones
> estatales para un intervalo de
> confianza (IC) del 95% fue del 1,38%.
>
> Thanks very much in advance!
>
> Albert-Jan
>
> --- On Sat, 8/1/09, Daciuk, Tim <[hidden email]>
> wrote:
>
> > From: Daciuk, Tim <[hidden email]>
> > Subject: Re: odd problem in spss with variable name
> > To: [hidden email]
> > Date: Saturday, August 1, 2009, 12:32 AM
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > I am
> > currently working with a dataset that
> > has a variable called path. I have not
> encountered
> > any problems using it in
> > SPSS 17 and 18. What do you get if you run a
> > frequencies from the dialog
> > boxes?
> >
> >
> >
> >
> >
> >
> > Tim
> > Daciuk
> >
> > Director, Worldwide
> > Demo Resources
> >
> > SPSS
> > Inc.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > From:
> > SPSSX(r) Discussion [mailto:[hidden email]]
> > On Behalf Of
> > [hidden email]
> >
> > Sent: Friday,
> > July 31, 2009 6:05
> > PM
> >
> > To:
> > [hidden email]
> >
> > Subject: Re:
> > odd problem in spss
> > with variable name
> >
> >
> >
> >
> >
> > There's a variable in our
> > dataset called path. Spss (version 15) refuses
> to
> > work with that variable
> > unless I select it in the Variable View and rename it
> > manually first. I
> > cannot manage to rename it with syntax. I did
> not
> > think "path"
> > was a system reserved command and SPSS should
> automatically
> > rename reserved
> > names. Perhaps it's a glitch in my copy of
> spss?
> >
> >
> >
> > Any ideas?
> >
> >
> >
> > JL
> >
> >
> >
> >
> >
> >
> >
>
> =====================
> 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
> No virus found in this incoming message.
> Checked by AVG - www.avg.com
> Version: 8.5.392 / Virus Database: 270.13.37/2273 - Release
> Date: 08/02/09
> 17:56:00
>
> =====================
> 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
No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 8.5.392 / Virus Database: 270.13.37/2273 - Release Date: 08/02/09 17:56:00

=====================
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

Bibel, Daniel (POL)

font for printing 'old' output?

In reply to this post by Paola Chivers-2

An odd question, but...
I have an older document, circa 1999, which has a number of tables which
were produced by SPSS Tables using the default font. The document is a
WordPerfect file currently, and the tables do not look very good - the
border characters do not print out well - they do not 'line up'
correctly.

Does anyone have an idea of a font which could be used to print these
out properly?

Thanks in advance for any assistance.

Dan Bibel
Massachusetts State Police

=====================
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