SPSSX Discussion

Odds Rtio & Logistic Regression

Classic

List

Threaded

17 messages Options

Okasha

Odds Rtio & Logistic Regression

Hellow every body,

'am asking about two things:-

1- how can i get the Odds Ration from SPSS
2- how can i get Logistic Regression.

Pls tell me what is the specific steps can i do?

many thanks,

Okasha

John Amora

Re: Odds Rtio & Logistic Regression

Hi Okasha,

Consider the following variables: Y = a yes/no variable, which is the outcome variable; X1 and X2 are the predictor variables, where X1 is continuous and X2 is categorical. The syntax for the binary logistic regression looks like as follows:

LOGISTIC REGRESSION VARIABLES Y
/METHOD=ENTER X1 X2
/CONTRAST (X2)=Indicator
/CLASSPLOT
/PRINT=GOODFIT
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

Then, the column heading "exp(B)" in the logistic regression output is the odds ratio. You can find in the outputs the B coefficients to form the logistic regression.

Hope this helps.
Johnny

--- On Wed, 12/24/08, Okasha <[hidden email]> wrote:

From: Okasha <[hidden email]>
Subject: Odds Rtio & Logistic Regression
To: [hidden email]
Date: Wednesday, 24 December, 2008, 3:05 PM

Hellow every body,

'am asking about two things:-

1- how can i get the Odds Ration from SPSS
2- how can i get Logistic Regression.

Pls tell me what is the specific steps can i do?

many thanks,

Okasha
--
View this message in context:
http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21156112.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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

Greetings from Emoticarolers! Create and send a personalised holiday jingle with Yahoo! Messenger today! Try it at http://emoticarolers.com/

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

Johnny Amora

Re: Odds Rtio & Logistic Regression

In reply to this post by Okasha

Okasha

Re: Odds Rtio & Logistic Regression

Hi Johnny,

Many thanks for ur useful interpretation, now could u show me the steps which i have to use to obtain
"exp(B)" and and B coefficients , couse i use SPSS 12< but i can't find the correct menu..
Pls Show me the path like ( Analyze>> Regression >> ....)

Thanks again
Okasha

Johnny Amora wrote

Hi Okasha,

Consider the following variables: Y = a yes/no variable, which is the outcome variable; X1 and X2 are the predictor variables, where X1 is continuous and X2 is categorical. The syntax for the binary logistic regression looks like as follows:

LOGISTIC REGRESSION VARIABLES Y
/METHOD=ENTER X1 X2
/CONTRAST (X2)=Indicator
/CLASSPLOT
/PRINT=GOODFIT
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).

Then, the column heading "exp(B)" in the logistic regression output is the odds ratio. You can find in the outputs the B coefficients to form the logistic regression.

Hope this helps.
Johnny

--- On Wed, 12/24/08, Okasha <hasabu20000@HOTMAIL.COM> wrote:

From: Okasha <hasabu20000@HOTMAIL.COM>
Subject: Odds Rtio & Logistic Regression
To: SPSSX-L@LISTSERV.UGA.EDU
Date: Wednesday, 24 December, 2008, 3:05 PM

Hellow every body,

'am asking about two things:-

1- how can i get the Odds Ration from SPSS
2- how can i get Logistic Regression.

Pls tell me what is the specific steps can i do?

many thanks,

Okasha
--
View this message in context:
http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21156112.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (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

Greetings from Emoticarolers! Create and send a personalised holiday jingle with Yahoo! Messenger today! Try it at http://emoticarolers.com/

====================To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (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

Johnny Amora

Re: Odds Rtio & Logistic Regression

Try this website:

http://faculty.chass.ncsu.edu/garson/PA765/logispss.htm

Cheers,
Johnny

--- On Sat, 12/27/08, Okasha <[hidden email]> wrote:

From: Okasha <[hidden email]>
Subject: Re: Odds Rtio & Logistic Regression
To: [hidden email]
Date: Saturday, 27 December, 2008, 2:06 PM

Hi Johnny,

Many thanks for ur useful interpretation, now could u show me the steps
which i have to use to obtain
"exp(B)" and and B coefficients , couse i use SPSS 12< but i
can't find the
correct menu..
Pls Show me the path like ( Analyze>> Regression >> ....)

Thanks again
Okasha

Johnny Amora wrote:

>
> Hi Okasha,
> Â
> Consider the following variables: Y =Â a yes/no variable,Â which is
> theÂ outcome variable; X1 and X2 are the predictor variables, where X1 is
> continuous and X2 is categorical.Â The syntax for the binary logistic
> regression looks like as follows:
> Â
> LOGISTIC REGRESSION VARIABLESÂ Y
> Â /METHOD=ENTERÂ X1 X2Â
> Â /CONTRAST (X2)=Indicator
> Â /CLASSPLOT
> Â /PRINT=GOODFIT
> Â /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).
> Â
> Then, the column heading "exp(B)" in the logistic regression

output is the

> odds ratio.Â You can find in the outputs the B coefficients to form the
> logistic regression.
> Â
> Hope this helps.
> Johnny
>
> --- On Wed, 12/24/08, Okasha <[hidden email]> wrote:
>
> From: Okasha <[hidden email]>
> Subject: Odds Rtio & Logistic Regression
> To: [hidden email]
> Date: Wednesday, 24 December, 2008, 3:05 PM
>
> Hellow every body,
>
> 'am asking about two things:-
>
> 1- how can i get the Odds Ration from SPSS
> 2- how can i get Logistic Regression.
>
> Pls tell me what is the specific steps can i do?
>
> many thanks,
>
> Okasha
> --
> View this message in context:
>

http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21156112.html

> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> 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
>
>
>
> Greetings from Emoticarolers! Create and send a personalised holiday
> jingle with Yahoo! Messenger today! Try it at http://emoticarolers.com/
>
> ====================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
>
>

paul wilson-7

Re: Odds Rtio & Logistic Regression

Logistic Regression is not available in SPSS base.
You need an add on module in addition to SPSS base to be able to run it.
I'm guessing you only have the base module which would be the reason you don't see Logistic Regression in your drop down menu.

________________________________
From: Johnny Amora <[hidden email]>
To: [hidden email]
Sent: Saturday, December 27, 2008 6:53:49 PM
Subject: Re: Odds Rtio & Logistic Regression

Try this website:

http://faculty.chass.ncsu.edu/garson/PA765/logispss.htm

Cheers,
Johnny

--- On Sat, 12/27/08, Okasha <[hidden email]> wrote:

From: Okasha <[hidden email]>
Subject: Re: Odds Rtio & Logistic Regression
To: [hidden email]
Date: Saturday, 27 December, 2008, 2:06 PM

Hi Johnny,

Many thanks for ur useful interpretation, now could u show me the steps
which i have to use to obtain
"exp(B)" and and B coefficients , couse i use SPSS 12< but i
can't find the
correct menu..
Pls Show me the path like ( Analyze>> Regression >> ....)

Thanks again
Okasha

Johnny Amora wrote:

output is the

http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21156112.html

--
View this message in context:
http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21181896.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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

Importing contacts has never been easier. Bring your friends over to Yahoo! Mail today! http://www.trueswitch.com/yahoo-ph

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

Okasha

Re: Odds Rtio & Logistic Regression

In reply to this post by Johnny Amora

Thanks a millions

Okasha

Johnny Amora wrote

Try this website:

http://faculty.chass.ncsu.edu/garson/PA765/logispss.htm

Cheers,
Johnny

--- On Sat, 12/27/08, Okasha <hasabu20000@HOTMAIL.COM> wrote:

From: Okasha <hasabu20000@HOTMAIL.COM>
Subject: Re: Odds Rtio & Logistic Regression
To: SPSSX-L@LISTSERV.UGA.EDU
Date: Saturday, 27 December, 2008, 2:06 PM

Hi Johnny,

Many thanks for ur useful interpretation, now could u show me the steps
which i have to use to obtain
"exp(B)" and and B coefficients , couse i use SPSS 12< but i
can't find the
correct menu..
Pls Show me the path like ( Analyze>> Regression >> ....)

Thanks again
Okasha

Johnny Amora wrote:
>
> Hi Okasha,
> Â
> Consider the following variables: Y =Â a yes/no variable,Â which is
> theÂ outcome variable; X1 and X2 are the predictor variables, where X1 is
> continuous and X2 is categorical.Â The syntax for the binary logistic
> regression looks like as follows:
> Â
> LOGISTIC REGRESSION VARIABLESÂ Y
> Â /METHOD=ENTERÂ X1 X2Â
> Â /CONTRAST (X2)=Indicator
> Â /CLASSPLOT
> Â /PRINT=GOODFIT
> Â /CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).
> Â
> Then, the column heading "exp(B)" in the logistic regression
output is the
> odds ratio.Â You can find in the outputs the B coefficients to form the
> logistic regression.
> Â
> Hope this helps.
> Johnny
>
> --- On Wed, 12/24/08, Okasha <hasabu20000@HOTMAIL.COM> wrote:
>
> From: Okasha <hasabu20000@HOTMAIL.COM>
> Subject: Odds Rtio & Logistic Regression
> To: SPSSX-L@LISTSERV.UGA.EDU
> Date: Wednesday, 24 December, 2008, 3:05 PM
>
> Hellow every body,
>
> 'am asking about two things:-
>
> 1- how can i get the Odds Ration from SPSS
> 2- how can i get Logistic Regression.
>
> Pls tell me what is the specific steps can i do?
>
> many thanks,
>
> Okasha
> --
> View this message in context:
>
http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21156112.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> LISTSERV@LISTSERV.UGA.EDU (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
>
>
>
> Greetings from Emoticarolers! Create and send a personalised holiday
> jingle with Yahoo! Messenger today! Try it at http://emoticarolers.com/
>
> ====================To manage your subscription to SPSSX-L, send a message
> to
> LISTSERV@LISTSERV.UGA.EDU (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
>
>

--
View this message in context:
http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21181896.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (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

Importing contacts has never been easier. Bring your friends over to Yahoo! Mail today! http://www.trueswitch.com/yahoo-ph

====================To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (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

Johnny Amora

Re: Odds Rtio & Logistic Regression

In reply to this post by paul wilson-7

Okasha,

Paul is right that logistic regression is not available on the SPSS Base Module. It is available in the SPSS Regression Model. To check if you have licence for the Regression Model, just run the following syntax.

show license.

Cheers,
Johnny

--- On Sun, 12/28/08, paul wilson <[hidden email]> wrote:

From: paul wilson <[hidden email]>
Subject: Re: Odds Rtio & Logistic Regression
To: [hidden email]
Date: Sunday, 28 December, 2008, 2:36 PM

Logistic Regression is not available in SPSS base.
You need an add on module in addition to SPSS base to be able to run it.
I'm guessing you only have the base module which would be the reason you
don't see Logistic Regression in your drop down menu.

________________________________
From: Johnny Amora <[hidden email]>
To: [hidden email]
Sent: Saturday, December 27, 2008 6:53:49 PM
Subject: Re: Odds Rtio & Logistic Regression

Try this website:

http://faculty.chass.ncsu.edu/garson/PA765/logispss.htm

Cheers,
Johnny

--- On Sat, 12/27/08, Okasha <[hidden email]> wrote:

From: Okasha <[hidden email]>
Subject: Re: Odds Rtio & Logistic Regression
To: [hidden email]
Date: Saturday, 27 December, 2008, 2:06 PM

Hi Johnny,

Many thanks for ur useful interpretation, now could u show me the steps
which i have to use to obtain
"exp(B)" and and B coefficients , couse i use SPSS 12< but i
can't find the
correct menu..
Pls Show me the path like ( Analyze>> Regression >> ....)

Thanks again
Okasha

Johnny Amora wrote:

output is the

http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21156112.html

--
View this message in context:
http://www.nabble.com/Odds-Rtio---Logistic-Regression-tp21156112p21181896.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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

Importing contacts has never been easier. Bring your friends over to
Yahoo! Mail today! http://www.trueswitch.com/yahoo-ph

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

Greetings from Emoticarolers! Create and send a personalised holiday jingle with Yahoo! Messenger today! Try it at http://emoticarolers.com/

====================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: Odds Rtio & Logistic Regression

Hi Okasha

>>
>> --- On Wed, 12/24/08, Okasha <[hidden email]> wrote:
>>
>> From: Okasha <[hidden email]>
>> Subject: Odds Rtio & Logistic Regression
>> To: [hidden email]
>> Date: Wednesday, 24 December, 2008, 3:05 PM
>>
>> Hellow every body,
>>
>> 'am asking about two things:-
>>
>> 1- how can i get the Odds Ration from SPSS
>> 2- how can i get Logistic Regression.
>>
>> Pls tell me what is the specific steps can i do?
>>
>>
>>

Since it looks like you don't have the regression module, you can try
with the freeware program Winpepi (available at
http://www.brixtonhealth.com/). Data can be copy-pasted from SPSS to
Winpepi easily.

HTH,
Mara García-Granero

--
For miscellaneous 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

Pettitt-Morris Caroline

Creating new variables from a string variable

Hello.

I am having some trouble with a piece of syntax that I can’t work out.

I have a string variable that is 1000 characters long (all in use).
I need to put this info into excel for other officers who do not use SPSS to view.
Excel has a limit of 255 characters per cell.

How do I split this variable into 4 new variables with 250 characters in each?

Along the lines of...

Text Variable (1 – 250) = Text1
Text Variable (251 – 500) = Text2
Text Variable (501 – 750) = Text3
Text Variable (751 -1000) = Text 4
EXECUTE.

But I need it to work!!!

Have tried using Visual Binning (numeric only) and the recode dialogues but am stuck.

Could you offer any advice or help please?
� � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �
Caroline PETTITT-MORRIS
Intelligence� Analyst
Greyfriars Police Station
Greyfriars
Bedford
MK40 1HR
�
E-mail: [hidden email]

**********************************************************************
'Bedfordshire police are now recruiting - visit www.bedfordshire.police.uk for more information

**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
This footnote also confirms that this email message has been swept for the presence of computer viruses.
********************************************************************************
Internet email is not to be treated as a secure means of communication.
Bedfordshire Police monitors all Internet email activity and content.
********************************************************************************

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

Pettitt-Morris Caroline

Re: Creating new variables from a string variable

In reply to this post by Marta Garcia-Granero

Thank you.

This has now been explained.

Caroline

-----Original Message-----
From: Pettitt-Morris Caroline
Sent: 29 December 2008 12:37
To: [hidden email]
Subject: Creating new variables from a string variable

Hello.

I am having some trouble with a piece of syntax that I can’t work out.

I have a string variable that is 1000 characters long (all in use).
I need to put this info into excel for other officers who do not use SPSS to view.
Excel has a limit of 255 characters per cell.

How do I split this variable into 4 new variables with 250 characters in each?

Along the lines of...

Text Variable (1 – 250) = Text1
Text Variable (251 – 500) = Text2
Text Variable (501 – 750) = Text3
Text Variable (751 -1000) = Text 4
EXECUTE.

But I need it to work!!!

Have tried using Visual Binning (numeric only) and the recode dialogues but am stuck.

Could you offer any advice or help please?
� � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �
Caroline PETTITT-MORRIS
Intelligence� Analyst
Greyfriars Police Station
Greyfriars
Bedford
MK40 1HR
�
E-mail: [hidden email]

**********************************************************************
'Bedfordshire police are now recruiting - visit www.bedfordshire.police.uk for more information

**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
This footnote also confirms that this email message has been swept for the presence of computer viruses.
********************************************************************************
Internet email is not to be treated as a secure means of communication.
Bedfordshire Police monitors all Internet email activity and content.
********************************************************************************

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

Clive Downs

Re: Creating new variables from a string variable

Hi

The following syntax should do this, if I have understood your question
correctly.

* bigstring is the original 1000 char string variable.
*---------------------------------------.

STRING unpack1 unpack2 unpack3 unpack4 (A250).
COMPUTE unpack1 = SUBSTR(bigstring,1,250) .
COMPUTE unpack2 = SUBSTR(bigstring,251,500) .
COMPUTE unpack3 = SUBSTR(bigstring,501,750) .
COMPUTE unpack4 = SUBSTR(bigstring,751,1000) .
EXECUTE .

Regards

Clive

On Mon, 29 Dec 2008 12:36:56 -0000, Pettitt-Morris Caroline
<[hidden email]> wrote:

>Hello.
>
>I am having some trouble with a piece of syntax that I canât work out.
>
>I have a string variable that is 1000 characters long (all in use).
>I need to put this info into excel for other officers who do not use SPSS
to view.
>Excel has a limit of 255 characters per cell.
>
>How do I split this variable into 4 new variables with 250 characters in
each?

>
>Along the lines of...
>
>Text Variable (1 â 250) = Text1
>Text Variable (251 â 500) = Text2
>Text Variable (501 â 750) = Text3
>Text Variable (751 -1000) = Text 4
>EXECUTE.
>
>But I need it to work!!!
>
>Have tried using Visual Binning (numeric only) and the recode dialogues

but am stuck.
>
>Could you offer any advice or help please?
>Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â
Â Â Â Â >Caroline PETTITT-MORRIS
>IntelligenceÂ Analyst
>Greyfriars Police Station
>Greyfriars
>Bedford
>MK40 1HR

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

J P-6

comparing weighted data to unweighted data

In reply to this post by Johnny Amora

Dear List,

I appreciate your patience with this non-SPSS question.

From a practical standpoint, what is the consequence of conducting a statistical test comparing means or frequencies when one data set is weighted and the other is not (data has been subsetted so weights are no longer accurate).

Thank you,

John

Neda Faregh

Re: comparing weighted data to unweighted data

Hi John, I think that would depend on how the weight is calculated. If the weight affects your effective N then you might be comparing two populations with different sample sizes, in which case one mean (for the larger sample) is more trustworthy than the other (smaller sample). This should affect your conclusion only minimally, as long the smaller sample size is adequate enough.

If the weight affects only your beta coefficients, then this should not influence your mean or frequency calculations.

Hope this helps,

Neda

From: J P <[hidden email]>
To: [hidden email]
Sent: Thursday, February 26, 2009 10:45:54 AM
Subject: comparing weighted data to unweighted data

Dear List,

I appreciate your patience with this non-SPSS question.

Thank you,

John

Hector Maletta

Re: comparing weighted data to unweighted data

In reply to this post by J P-6

Several consequences. First, if the weights were “inflationary”, i.e. they amplify sample size to the size of the population, SPSS would consider the expanded size as the size of the sample, and compute significance accordingly (thus overstating the significance of results). Besides, if different cases have different weights (i.e. weights are correcting proportions in the sample to proportions in the population), using the unweighted set gives some cases more weight than they deserve, and other cases less, again distorting the results.

I do not understand your remark about weights being no Langer achúrate because the data have been subsetted. If you have N cases, each with a weight, and you take one subset of M cases (M<N), you can still use the weights. For instance, súpose you extract a sample of 1000 from a population, with a complex sampling design mandating that each case has a specific weight in the sample; your weighted sample would yield results representative of the referente population. Then suppose you select one subset consisting of about one half of the samle (say, just all the males): if you keep the weights, your subset would represent the characteristics of the male population, just as the entire sample represented the entire population of both sexes.

Hector.

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of J P
Sent: 26 February 2009 13:46
To: [hidden email]
Subject: comparing weighted data to unweighted data

Dear List,

I appreciate your patience with this non-SPSS question.

Thank you,

John

J P-6

Re: comparing weighted data to unweighted data

Dear Hector and Neda,

Thank you so much for responding (Hector, I am in the process of reading your paper on weighting).

Here is the situation I am dealing with: I work at a small college and we have NSSE (http://nsse.iub.edu/index.cfm) data that departments would like to use to make comparisons between themself and the rest of the college, our Carnegie peer group, and NSSE as a whole. For the Carnegie and NSSE groups all I have are weighted frequencies, means, SDs, and Ns from a summary report.

Is it legitimate for departments to perform t-tests comparing their unweighted department data to the college-minus-department, Carnegie and NSSE groups? The response rate is around 30% for my college. The NSSE weight is based on status and gender which I believe makes them 'proportional' weights. My college has only full-time students and is about 90% male, so the weight variable tends to hover around 1. The department n = about 25, college n about 200, Caregie n = about 35,000, NSSE n = about 300,000.

I have an economist and engineer who say it's perfectly fine, however I do not understand the complexities and am hesitant to give my blessings. I know I should pose this question to NSSE, but am trying to first educate myself a little.

Thanks again!

From: Hector Maletta <[hidden email]>
To: J P <[hidden email]>; [hidden email]
Sent: Thursday, February 26, 2009 11:08:15 AM
Subject: RE: comparing weighted data to unweighted data

Hector.

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of J P
Sent: 26 February 2009 13:46
To: [hidden email]
Subject: comparing weighted data to unweighted data

Dear List,

I appreciate your patience with this non-SPSS question.

Thank you,

John

Hector Maletta

Re: comparing weighted data to unweighted data

You can compare the mean of your department or college to some “reference” mean, such as the NSSE. SPSS lets you compare a sample mean to an arbitrary number to check whether they are significantly different or their difference could be attributed to (your) sampling error.
If your cases reproduce the gender and status proportions of your population, you do not need weights. Otherwise you could apply the necessary weights, so that your mean reflects the correct gender-status composition of your student population (by department, if need be).
Your sample seems to be small. N=25, for instance, means that margins of error are rather wide, whatever the test you apply.
Suppose, however, you do the comparison, and find there is a significant difference between your sample and the reference (NSSE). What do you make of it? The difference may be due in part to differences in gender and status composition between the two populations (yours and NSSE), and in part to real differences in other variables (keeping gender and status constant). How do you distinguish between the two? Once way is to weight your sample with the NSSE weights (i.e. give your male and female students the weights both genders have in the NSSE data, and likewise for status): any remaining difference, if significant, would be due to OTHER factors, such as real differences between the two populations keeping gender and status constant.

Hope this helps.

Hector

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of J P
Sent: 26 February 2009 14:46
To: [hidden email]
Subject: Re: comparing weighted data to unweighted data

Dear Hector and Neda,

Thank you so much for responding (Hector, I am in the process of reading your paper on weighting).

Thanks again!

From: Hector Maletta <[hidden email]>
To: J P <[hidden email]>; [hidden email]
Sent: Thursday, February 26, 2009 11:08:15 AM
Subject: RE: comparing weighted data to unweighted data

Hector.

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of J P
Sent: 26 February 2009 13:46
To: [hidden email]
Subject: comparing weighted data to unweighted data

Dear List,

I appreciate your patience with this non-SPSS question.

Thank you,

John