SPSSX Discussion

matrix data

Classic

List

Threaded

3 messages Options

Neil McClelland

matrix data

Please help me. I am new to using syntax - and not doing very well. What I
want to do is run a factor analysis on a correlation matrix. I attempted to
write the following syntax - which does not seem to work (two options are
included below - /contents and rowtype_ versions). Can anyone tell me what
I am doing wrong. Any help would be much appreciated.

With Regards.

Neil McClelland
University of Kitakyushu

MATRIX DATA VARIABLES=F1 TO F6 /CONTENTS =CORR.
BEGIN DATA
1.00
.00 1.00
.43 -.02 1.00
.55 .05 .23 1.00
.62 -.14 .56 .51 1.00
-.39 -.34 -.24 -.32 -.25 1.00
END DATA.
LIST VAR = ALL.
FACTOR
/MATRIX IN(*)
/MISSING LISTWISE
/ANALYSIS F1 TO F6
/PRINT INITIAL EXTRACTION ROTATION
/FORMAT SORT BLANK(.4)
/PLOT EIGEN
/CRITERIA MINEIGEN(1) ITERATE(25)
/EXTRACTION ML
/CRITERIA ITERATE(25) DELTA(0)
/ROTATION OBLIMIN.

MATRIX DATA VARIABLES=ROWTYPE_ F1 TO F6.
BEGIN DATA
CORR 1.00
CORR .00 1.00
CORR .43 -.02 1.00
CORR .55 .05 .23 1.00
CORR .62 -.14 .52 .51 1.00
CORR -.39 -.34 -.24 -.32 -.25 1.00
END DATA.
LIST VAR = ALL.
FACTOR
/MATRIX IN(*)
/MISSING LISTWISE
/ANALYSIS F1 TO F6
/PRINT INITIAL EXTRACTION ROTATION
/FORMAT SORT BLANK(.4)
/PLOT EIGEN
/CRITERIA MINEIGEN(1) ITERATE(25)
/EXTRACTION ML
/CRITERIA ITERATE(25) DELTA(0)
/ROTATION OBLIMIN.

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

Bruce Weaver

Re: matrix data

Administrator

To create a model of a matrix data file that will work as input for FACTOR, run some correlations with MATRIX OUT. E.g.,

CORRELATIONS
/VARIABLES=educ paeduc maeduc speduc prestg80
/PRINT=TWOTAIL NOSIG
/MATRIX=out(*)
/MISSING=LISTWISE.

The resulting matrix file looks like this:

ROWTYPE_ VARNAME_ educ paeduc maeduc speduc prestg80

MEAN 13.7211538 10.7961538 10.8346154 13.4153846 45.1673077
STDDEV 2.7463647 3.9124307 3.1895110 2.9031499 13.1678393
N 520.000000 520.000000 520.000000 520.000000 520.000000
CORR educ 1.0000000 .3983482 .4117756 .5694070 .5422923
CORR paeduc .3983482 1.0000000 .6487286 .3804975 .2382280
CORR maeduc .4117756 .6487286 1.0000000 .4221452 .2160521
CORR speduc .5694070 .3804975 .4221452 1.0000000 .3257418
CORR prestg80 .5422923 .2382280 .2160521 .3257418 1.0000000

Number of cases read: 8 Number of cases listed: 8

Judging by that, I'd try adding means, standard deviations and Ns to your matrix file. Indeed, the following command runs on the matrix file shown above, but fails when I delete the first 3 rows.

FACTOR
/MATRIX IN(COR=*)
/MISSING LISTWISE
/ANALYSIS educ paeduc maeduc speduc prestg80
/PRINT INITIAL EXTRACTION
/PLOT EIGEN
/CRITERIA MINEIGEN(1) ITERATE(25)
/EXTRACTION PAF
/ROTATION NOROTATE
/METHOD=CORRELATION.

Neil McClelland wrote

Please help me. I am new to using syntax - and not doing very well. What I
want to do is run a factor analysis on a correlation matrix. I attempted to
write the following syntax - which does not seem to work (two options are
included below - /contents and rowtype_ versions). Can anyone tell me what
I am doing wrong. Any help would be much appreciated.

With Regards.

Neil McClelland
University of Kitakyushu

MATRIX DATA VARIABLES=F1 TO F6 /CONTENTS =CORR.
BEGIN DATA
1.00
.00 1.00
.43 -.02 1.00
.55 .05 .23 1.00
.62 -.14 .56 .51 1.00
-.39 -.34 -.24 -.32 -.25 1.00
END DATA.
LIST VAR = ALL.
FACTOR
/MATRIX IN(*)
/MISSING LISTWISE
/ANALYSIS F1 TO F6
/PRINT INITIAL EXTRACTION ROTATION
/FORMAT SORT BLANK(.4)
/PLOT EIGEN
/CRITERIA MINEIGEN(1) ITERATE(25)
/EXTRACTION ML
/CRITERIA ITERATE(25) DELTA(0)
/ROTATION OBLIMIN.

MATRIX DATA VARIABLES=ROWTYPE_ F1 TO F6.
BEGIN DATA
CORR 1.00
CORR .00 1.00
CORR .43 -.02 1.00
CORR .55 .05 .23 1.00
CORR .62 -.14 .52 .51 1.00
CORR -.39 -.34 -.24 -.32 -.25 1.00
END DATA.
LIST VAR = ALL.
FACTOR
/MATRIX IN(*)
/MISSING LISTWISE
/ANALYSIS F1 TO F6
/PRINT INITIAL EXTRACTION ROTATION
/FORMAT SORT BLANK(.4)
/PLOT EIGEN
/CRITERIA MINEIGEN(1) ITERATE(25)
/EXTRACTION ML
/CRITERIA ITERATE(25) DELTA(0)
/ROTATION OBLIMIN.

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

--
Bruce Weaver
bweaver@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

PLEASE NOTE THE FOLLOWING:
1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above.
2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/).

Art Kendall

Re: matrix data

In reply to this post by Neil McClelland

The syntax in green shows changes that made your syntax work.
This line is new.n 500 500 500 500 500 500

this line changed the file specification.
/MATRIX IN(cor=*)

If you are new to syntax you may be new to factor analysis. OBLIMIN allows the factors to be correlated. The correlation of.14 between the factors is fairly small so it may not be a problem. If you are trying to create scales, you would do better to stick with varimax rotation in order to have divergent validity.
If the grouping of variables makes sense scale A would be the sum of F5, F1, F4, and F3. Scale B would be the sum of F6 and F2 reflected. But beware F6 is a "splitter", so depending on your goals you may want to stick with scale A only.

An edited set of tested syntax follows.

MATRIX DATA VARIABLES=ROWTYPE_ F1 TO F6. BEGIN DATA n 500 500 500 500 500 500 CORR 1.00 CORR .00 1.00 CORR .43 -.02 1.00 CORR .55 .05 .23 1.00 CORR .62 -.14 .52 .51 1.00 CORR -.39 -.34 -.24 -.32 -.25 1.00 END DATA. LIST VAR = ALL. FACTOR /MATRIX IN(cor=*) /MISSING LISTWISE /ANALYSIS F1 TO F6 /PRINT INITIAL EXTRACTION ROTATION /FORMAT SORT BLANK(.40) /PLOT EIGEN /CRITERIA MINEIGEN(1) ITERATE(25) /EXTRACTION ML /CRITERIA ITERATE(25) DELTA(0) /ROTATION OBLIMIN.
Art Kendall
Social Research Consultants

On 11/25/2010 7:57 PM, Neil McClelland wrote:

Please help me.  I am new to using syntax - and not doing very well.  What I
want to do is run a factor analysis on a correlation matrix.  I attempted to
write the following syntax - which does not seem to work (two options are
included below - /contents and rowtype_ versions).  Can anyone tell me what
I am doing wrong.  Any help would be much appreciated.

With Regards.

Neil McClelland
University of Kitakyushu


<snip>

MATRIX DATA VARIABLES=ROWTYPE_ F1 TO F6.
BEGIN DATA
CORR 1.00
CORR .00   1.00
CORR .43   -.02   1.00
CORR .55   .05   .23   1.00
CORR .62   -.14   .52   .51   1.00
CORR -.39   -.34   -.24   -.32   -.25   1.00
 END DATA.
LIST VAR = ALL.
FACTOR
/MATRIX IN(*)
/MISSING LISTWISE
/ANALYSIS F1 TO F6
  /PRINT INITIAL EXTRACTION ROTATION
  /FORMAT SORT BLANK(.4)
  /PLOT EIGEN
  /CRITERIA MINEIGEN(1) ITERATE(25)
  /EXTRACTION ML
  /CRITERIA ITERATE(25) DELTA(0)
  /ROTATION OBLIMIN.

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

Art Kendall
Social Research Consultants