SPSSX Discussion

Data Generation for Pre-determined correations

Classic

List

Threaded

7 messages Options

Sungeun Chung-2

Data Generation for Pre-determined correations

Dear SPSS Experts

I want to create a data set with five variables for pre-determined correlations.

1.000, .075, .400, .286, .309;

.075, 1.000, .411, .834, -.521;

.400, .411, 1.000, .846, .564;

.286, .834, .846, 1.000, .036;

.309, .521, .564, .036, 1.000

I tried to do it by modifying the syntax that David Marso posted (March 15, 2013).

Specifically, I replaced "3" with "5" in "CRITERIA FACTORS(3)." Then I put correlations for five variables.

But it didn't work. How can I fix the problem?

Thank you for your help in advance,

Sungeun

Sungeun Chung, Ph.D. Associate Professor
Department of Journalism and Mass Communication
40210 Faculty Hall Sungkyunkwan University
53 Myeongyun-Dong 3-Ga Jongno-Gu, Seoul, Korea 110-745
82-02-760-0398; [hidden email]; [hidden email]

**** David Marso Syntax*.

DEFINE CorrSim (CorrMat !CHAREND ("/") / N !TOKENS(1) / P !TOKENS(1)).

!LET !VLIST1 = !CONCAT('X1 TO X',!P ).

!LET !VLIST2 = !CONCAT('FS1 TO FS',!P ).

!LET !VLIST3 = !CONCAT('Sim1 TO Sim',!P ).

* Simulate raw data of desired dimensionality *.

INPUT PROGRAM.

+ LOOP #=1 TO !N.

+ DO REPEAT V=!VLIST1.

+ COMPUTE V=RV.NORMAL(0,1).

+ END REPEAT.

+ END CASE.

+ END LOOP.

+ END FILE.

END INPUT PROGRAM.

* Orthogonalize Vectors * .

FACTOR VAR !VLIST1 / CRITERIA FACTORS(5)/ SAVE REG (!P,FS).

* Generate Vectors with desired dimensionality *.

MATRIX.

GET RAW / VAR !VLIST1.

GET FS / VAR !VLIST2.

COMPUTE CORR={!CorrMat}.

SAVE {RAW,FS,FS*CHOL(CORR)} / OUTFILE * /VARIABLES !VLIST1 !VLIST2 !VLIST3.

END MATRIX.

** Check results (Optional) **.

CORRELATION !VLIST3 .

!ENDDEFINE.

SET MPRINT ON PRINTBACK ON.

* Self explanatory!*.

CorrSim CorrMat 1.000, .075, .400, .286, .309; .075, 1.000, .411, .834, -.521; .400, .411, 1.000, .846, .564; .286, .834, .846, 1.000, .036; .309, .521, .564, .036, 1.000

/ N=291 P=5.

Rich Ulrich

Re: Data Generation for Pre-determined correations

It usually helps if you are more specific than "it didn't work";
but this time I can tell you that you gave it a matrix that is
not symmetrical. And that would account for a failure.

"-.521" needs to be ".521" so that element (5,2) matches (2,5).

Using negative for both might work, or it might leave the matrix
non-positive definite. "Positive" looks like what was intended.

--
Rich Ulrich

Date: Thu, 28 Mar 2013 19:10:22 -0700
From: [hidden email]
Subject: Data Generation for Pre-determined correations
To: [hidden email]

Dear SPSS Experts

I want to create a data set with five variables for pre-determined correlations.

1.000, .075, .400, .286, .309;

.075, 1.000, .411, .834, -.521;

.400, .411, 1.000, .846, .564;

.286, .834, .846, 1.000, .036;

.309, .521, .564, .036, 1.000

I tried to do it by modifying the syntax that David Marso posted (March 15, 2013).

Specifically, I replaced "3" with "5" in "CRITERIA FACTORS(3)." Then I put correlations for five variables.

But it didn't work. How can I fix the problem?

Thank you for your help in advance,

Sungeun

**** David Marso Syntax*.

DEFINE CorrSim (CorrMat !CHAREND ("/") / N !TOKENS(1) / P !TOKENS(1)).

!LET !VLIST1 = !CONCAT('X1 TO X',!P ).

!LET !VLIST2 = !CONCAT('FS1 TO FS',!P ).

!LET !VLIST3 = !CONCAT('Sim1 TO Sim',!P ).

* Simulate raw data of desired dimensionality *.

INPUT PROGRAM.

+ LOOP #=1 TO !N.

+ DO REPEAT V=!VLIST1.

+ COMPUTE V=RV.NORMAL(0,1).

+ END REPEAT.

+ END CASE.

+ END LOOP.

+ END FILE.

END INPUT PROGRAM.

* Orthogonalize Vectors * .

FACTOR VAR !VLIST1 / CRITERIA FACTORS(5)/ SAVE REG (!P,FS).

* Generate Vectors with desired dimensionality *.

MATRIX.

GET RAW / VAR !VLIST1.

GET FS / VAR !VLIST2.

COMPUTE CORR={!CorrMat}.

SAVE {RAW,FS,FS*CHOL(CORR)} / OUTFILE * /VARIABLES !VLIST1 !VLIST2 !VLIST3.

END MATRIX.

** Check results (Optional) **.

CORRELATION !VLIST3 .

!ENDDEFINE.

SET MPRINT ON PRINTBACK ON.

* Self explanatory!*.

CorrSim CorrMat 1.000, .075, .400, .286, .309; .075, 1.000, .411, .834, -.521; .400, .411, 1.000, .846, .564; .286, .834, .846, 1.000, .036; .309, .521, .564, .036, 1.000

/ N=291 P=5.

Ryan

Re: Data Generation for Pre-determined correations

In reply to this post by Sungeun Chung-2

A little OT...I noticed that the code here:

http://www.uvm.edu/~dhowell/StatPages/More_Stuff/CorrData.SPS

has a RECODE statement with a corresponding note that does not make
any sense to me:

* Now you have your correlated variables. The next lines adjust the
means and variances.
RECODE
nr5 (-10 thru 0=1) (.0001 thru 10=2) .

If anyone plans on using this code, I suggest you grey out this
RECODE. I hope I am corrected if I'm wrong.

Ryan

On Thu, Mar 28, 2013 at 10:10 PM, Sungeun Chung <[hidden email]> wrote:

> Dear SPSS Experts
>
> I want to create a data set with five variables for pre-determined
> correlations.
>
> 1.000, .075, .400, .286, .309;
> .075, 1.000, .411, .834, -.521;
> .400, .411, 1.000, .846, .564;
> .286, .834, .846, 1.000, .036;
> .309, .521, .564, .036, 1.000
>
> I tried to do it by modifying the syntax that David Marso posted (March 15,
> 2013).
> Specifically, I replaced "3" with "5" in "CRITERIA FACTORS(3)." Then I put
> correlations for five variables.
> But it didn't work. How can I fix the problem?
> Thank you for your help in advance,
>
> Sungeun
>
>
> Sungeun Chung, Ph.D. Associate Professor
> Department of Journalism and Mass Communication
> 40210 Faculty Hall Sungkyunkwan University
> 53 Myeongyun-Dong 3-Ga Jongno-Gu, Seoul, Korea 110-745
> 82-02-760-0398; [hidden email]; [hidden email]
>
> **** David Marso Syntax*.
>
> DEFINE CorrSim (CorrMat !CHAREND ("/") / N !TOKENS(1) / P !TOKENS(1)).
>
> !LET !VLIST1 = !CONCAT('X1 TO X',!P ).
> !LET !VLIST2 = !CONCAT('FS1 TO FS',!P ).
> !LET !VLIST3 = !CONCAT('Sim1 TO Sim',!P ).
>
> * Simulate raw data of desired dimensionality *.
> INPUT PROGRAM.
> + LOOP #=1 TO !N.
> + DO REPEAT V=!VLIST1.
> + COMPUTE V=RV.NORMAL(0,1).
> + END REPEAT.
> + END CASE.
> + END LOOP.
> + END FILE.
> END INPUT PROGRAM.
>
> * Orthogonalize Vectors * .
> FACTOR VAR !VLIST1 / CRITERIA FACTORS(5)/ SAVE REG (!P,FS).
>
> * Generate Vectors with desired dimensionality *.
> MATRIX.
> GET RAW / VAR !VLIST1.
> GET FS / VAR !VLIST2.
> COMPUTE CORR={!CorrMat}.
> SAVE {RAW,FS,FS*CHOL(CORR)} / OUTFILE * /VARIABLES !VLIST1 !VLIST2 !VLIST3.
> END MATRIX.
>
> ** Check results (Optional) **.
> CORRELATION !VLIST3 .
> !ENDDEFINE.
> SET MPRINT ON PRINTBACK ON.
>
> * Self explanatory!*.
> CorrSim CorrMat 1.000, .075, .400, .286, .309; .075, 1.000, .411, .834,
> -.521; .400, .411, 1.000, .846, .564; .286, .834, .846, 1.000, .036; .309,
> .521, .564, .036, 1.000
> / N=291 P=5.
>
>

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

David Marso

Re: Data Generation for Pre-determined correations

Administrator

In reply to this post by Sungeun Chung-2

Note that I edited my post later and replaced the hard coded 3 with !P .
The 3 was possibly a premacrofication artifact of my home brewed DoppleBock ;-).
--
FACTOR VAR !VLIST1 / CRITERIA FACTORS(!P)/ SAVE REG (!P,FS).

-- and as Rich noted, the matrix needs to be symmetrical (fix that) --.

"But it didn't work.".
Best to post specific errors so we don't need to crank up (TAFKA)SPSS and replicate.
I can typically sort this type of thing without having to run code if I see the train wreck from the run.
--
Hope this helps, David

Sungeun Chung-2 wrote

Dear SPSS Experts

I want to create a data set with five variables for pre-determined correlations.

1.000, .075, .400, .286, .309;
.075, 1.000, .411, .834, -.521;
.400, .411, 1.000, .846, .564;
.286, .834, .846, 1.000, .036;
.309, .521, .564, .036, 1.000

I tried to do it by modifying the syntax that David Marso posted (March 15, 2013).
Specifically, I replaced "3" with "5" in "CRITERIA FACTORS(3)." Then I put correlations for five variables.
But it didn't work. How can I fix the problem?
Thank you for your help in advance,

Sungeun

Sungeun Chung, Ph.D. Associate Professor
Department of Journalism and Mass Communication
40210 Faculty Hall Sungkyunkwan University
53 Myeongyun-Dong 3-Ga Jongno-Gu, Seoul, Korea 110-745
82-02-760-0398; [hidden email]; [hidden email]

**** David Marso Syntax*.

DEFINE CorrSim (CorrMat !CHAREND ("/") / N !TOKENS(1) / P !TOKENS(1)).

!LET !VLIST1 = !CONCAT('X1 TO X',!P ).
!LET !VLIST2 = !CONCAT('FS1 TO FS',!P ).
!LET !VLIST3 = !CONCAT('Sim1 TO Sim',!P ).

* Simulate raw data of desired dimensionality *.
INPUT PROGRAM.
+ LOOP #=1 TO !N.
+ DO REPEAT V=!VLIST1.
+ COMPUTE V=RV.NORMAL(0,1).
+ END REPEAT.
+ END CASE.
+ END LOOP.
+ END FILE.
END INPUT PROGRAM.

* Orthogonalize Vectors * .
FACTOR VAR !VLIST1 / CRITERIA FACTORS(5)/ SAVE REG (!P,FS).

* Generate Vectors with desired dimensionality *.
MATRIX.
GET RAW / VAR !VLIST1.
GET FS / VAR !VLIST2.
COMPUTE CORR={!CorrMat}.
SAVE {RAW,FS,FS*CHOL(CORR)} / OUTFILE * /VARIABLES !VLIST1 !VLIST2 !VLIST3.
END MATRIX.

** Check results (Optional) **.
CORRELATION !VLIST3 .
!ENDDEFINE.
SET MPRINT ON PRINTBACK ON.

* Self explanatory!*.
CorrSim CorrMat 1.000, .075, .400, .286, .309; .075, 1.000, .411, .834, -.521; .400, .411, 1.000, .846, .564; .286, .834, .846, 1.000, .036; .309, .521, .564, .036, 1.000
/ N=291 P=5.

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"

Sungeun Chung-2

Re: Data Generation for Pre-determined correations

Thanks Ryan, Rich Ulrich, and David for finding errors and giving solutions.

Both Ryan's and David' syntax are working well (with different correlation matrices). It seems that my matrix has a problem (In fact, I don't have raw data but the author of a paper gave me correlations).

Thank again.

Sungeun

From: David Marso <[hidden email]>
To: [hidden email]
Sent: Friday, March 29, 2013 1:08 PM
Subject: Re: Data Generation for Pre-determined correations

Note that I edited my post later and replaced the hard coded 3 with !P .
The 3 was possibly a premacrofication artifact of my home brewed DoppleBock
;-).
--
FACTOR VAR !VLIST1 / CRITERIA FACTORS(!P)/ SAVE REG (!P,FS).

-- and as Rich noted, the matrix needs to be symmetrical (fix that) --.

"But it didn't work.".
Best to post specific errors so we don't need to crank up (TAFKA)SPSS and
replicate.
I can typically sort this type of thing without having to run code if I see
the train wreck from the run.
--
Hope this helps, David

Sungeun Chung-2 wrote

> Dear SPSS Experts
>
> I want to create a data set with five variables for pre-determined
> correlations. Â
>
> Â 1.000, .075, .400, .286, .309;Â
> .075, 1.000, .411, .834, -.521;Â
> .400, .411, 1.000, .846, .564;Â
> .286, .834, .846, 1.000, .036;Â
> .309, .521, .564, .036, 1.000
>
>
> I tried to do it by modifying the syntax that David Marso posted (March
> 15, 2013).
> Specifically, I replaced "3" with "5" in "CRITERIA FACTORS(3)." Then I put
> correlations for five variables.
> But it didn't work. How can I fix the problem?
> Thank you for your help in advance,
>
> Sungeun
>
>
> Sungeun Chung, Ph.D. Associate Professor
> Department of Journalism and Mass Communication
> 40210 Faculty Hall Sungkyunkwan University
> 53 Myeongyun-Dong 3-Ga Jongno-Gu, Seoul, Korea 110-745
> 82-02-760-0398;Â

> chseun@

> ;Â

> chseun@

>
>
> **** Â David Marso Syntax*.
>
> DEFINE CorrSim (CorrMat !CHAREND ("/") / N !TOKENS(1) / P !TOKENS(1)).
>
> !LET !VLIST1 = !CONCAT('X1 TO X',!P ).
> !LET !VLIST2 = !CONCAT('FS1 TO FS',!P ).
> !LET !VLIST3 = !CONCAT('Sim1 TO Sim',!P ).
>
> * Simulate raw data of desired dimensionality *.
> INPUT PROGRAM.
> + Â LOOP #=1 TO !N.
> + Â Â DO REPEAT V=!VLIST1.
> + Â Â Â COMPUTE V=RV.NORMAL(0,1).
> + Â Â END REPEAT.
> + Â Â END CASE.
> + Â END LOOP.
> + Â END FILE.
> END INPUT PROGRAM.
>
> * Orthogonalize Vectors * .
> FACTOR VAR !VLIST1 / CRITERIA FACTORS(5)/ SAVE REG (!P,FS).
>
> * Generate Vectors with desired dimensionality *.
> MATRIX.
> GET RAW Â / VAR !VLIST1.
> GET FS Â Â / VAR !VLIST2.
> COMPUTE CORR={!CorrMat}.
> SAVE {RAW,FS,FS*CHOL(CORR)} / OUTFILE * /VARIABLES !VLIST1 !VLIST2
> !VLIST3.
> END MATRIX.
>
> ** Check results (Optional) **.
> CORRELATION !VLIST3 .
> !ENDDEFINE.
> SET MPRINT ON PRINTBACK ON.
>
> * Self explanatory!*.
> CorrSim CorrMat 1.000, .075, .400, .286, .309; .075, 1.000, .411, .834,
> -.521; .400, .411, 1.000, .846, .564; .286, .834, .846, 1.000, .036; .309,
> .521, .564, .036, 1.000
> Â Â Â Â Â Â Â Â / N=291 Â P=5.

-----
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-Generation-for-Pre-determined-correations-tp5719160p5719163.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

Ryan

Re: Data Generation for Pre-determined correations

For the record, that link is not directed to syntax I developed. Since it could be used for the OP's question, along with the fact that I found a potential error, I figured I'd chime in.

Ryan

On Fri, Mar 29, 2013 at 2:47 AM, Sungeun Chung <[hidden email]> wrote:

Thanks Ryan, Rich Ulrich, and David for finding errors and giving solutions.

Both Ryan's and David' syntax are working well (with different correlation matrices). It seems that my matrix has a problem (In fact, I don't have raw data but the author of a paper gave me correlations).

Thank again.

Sungeun

From: David Marso <[hidden email]>
To: [hidden email]
Sent: Friday, March 29, 2013 1:08 PM
Subject: Re: Data Generation for Pre-determined correations

Note that I edited my post later and replaced the hard coded 3 with !P .
The 3 was possibly a premacrofication artifact of my home brewed DoppleBock
;-).
--
FACTOR VAR !VLIST1 / CRITERIA FACTORS(!P)/ SAVE REG (!P,FS).

-- and as Rich noted, the matrix needs to be symmetrical (fix that) --.

"But it didn't work.".
Best to post specific errors so we don't need to crank up (TAFKA)SPSS and
replicate.
I can typically sort this type of thing without having to run code if I see
the train wreck from the run.
--
Hope this helps, David

Sungeun Chung-2 wrote
> Dear SPSS Experts
>
> I want to create a data set with five variables for pre-determined
> correlations. Â
>

> Â 1.000, .075, .400, .286, .309;Â
> .075, 1.000, .411, .834, -.521;Â
> .400, .411, 1.000, .846, .564;Â
> .286, .834, .846, 1.000, .036;Â
> .309, .521, .564, .036, 1.000
>
>
> I tried to do it by modifying the syntax that David Marso posted (March
> 15, 2013).
> Specifically, I replaced "3" with "5" in "CRITERIA FACTORS(3)." Then I put
> correlations for five variables.
> But it didn't work. How can I fix the problem?
> Thank you for your help in advance,
>
> Sungeun
>
>
> Sungeun Chung, Ph.D. Associate Professor
> Department of Journalism and Mass Communication
> 40210 Faculty Hall Sungkyunkwan University
> 53 Myeongyun-Dong 3-Ga Jongno-Gu, Seoul, Korea 110-745
> 82-02-760-0398;Â

> chseun@

> ;Â

> chseun@

>
>
> **** Â David Marso Syntax*.
>
> DEFINE CorrSim (CorrMat !CHAREND ("/") / N !TOKENS(1) / P !TOKENS(1)).
>
> !LET !VLIST1 = !CONCAT('X1 TO X',!P ).
> !LET !VLIST2 = !CONCAT('FS1 TO FS',!P ).
> !LET !VLIST3 = !CONCAT('Sim1 TO Sim',!P ).
>
> * Simulate raw data of desired dimensionality *.
> INPUT PROGRAM.
> + Â LOOP #=1 TO !N.
> + Â Â DO REPEAT V=!VLIST1.
> + Â Â Â COMPUTE V=RV.NORMAL(0,1).
> + Â Â END REPEAT.
> + Â Â END CASE.
> + Â END LOOP.
> + Â END FILE.
> END INPUT PROGRAM.
>
> * Orthogonalize Vectors * .
> FACTOR VAR !VLIST1 / CRITERIA FACTORS(5)/ SAVE REG (!P,FS).
>
> * Generate Vectors with desired dimensionality *.
> MATRIX.
> GET RAW Â / VAR !VLIST1.
> GET FS Â Â / VAR !VLIST2.
> COMPUTE CORR={!CorrMat}.
> SAVE {RAW,FS,FS*CHOL(CORR)} / OUTFILE * /VARIABLES !VLIST1 !VLIST2
> !VLIST3.

> END MATRIX.
>
> ** Check results (Optional) **.
> CORRELATION !VLIST3 .
> !ENDDEFINE.
> SET MPRINT ON PRINTBACK ON.
>
> * Self explanatory!*.
> CorrSim CorrMat 1.000, .075, .400, .286, .309; .075, 1.000, .411, .834,
> -.521; .400, .411, 1.000, .846, .564; .286, .834, .846, 1.000, .036; .309,
> .521, .564, .036, 1.000
> Â Â Â Â Â Â Â Â / N=291 Â P=5.

-----
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-Generation-for-Pre-determined-correations-tp5719160p5719163.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

David Marso

Re: Data Generation for Pre-determined correations

Administrator

In reply to this post by David Marso

FWIW:
Here is a link to my original post (note the 3 had already been corrected to !P on March 15?).
Also cited Howell, but I am unimpressed with code which requires multiple user interventions.
http://spssx-discussion.1045642.n5.nabble.com/Re-Application-of-Rule-set-in-SPSS-Logistic-Regression-tt5718607.html#a5718648