Re: Can factors identified using factor analysis on a smaller dataset be used on larger dataset?

[Kooij, A.J. van der]

You can do this using CATPCA if you have SPSS Categories. If you have
continuous variables you can
discretize them (by using CATPCA Discretization option multiply) and use
numerical scaling level.
Specify the unhealty persons as supplementary cases. These cases will
not be included in the analysis, but will be fitted into the factor
solution found for the healthy persons.
To examine relations between factors and demographic variables you
include the demographic variables in the CATPCA variable list, and in
the analysis variables list with Multiple Nominal scaling level, and
specify them as supplementary variables, then they are not included in
the PCA but are fitted into the found solution.
For relationships between the factors and demographic variables you can
inspect the loadings plot with centroids included (centroid points are
the points for the categories of the demographic variables). To examine
relations between persons and variables you can inspect the biplot
(including object(person) points and loadings) and the triplot
(including object(person) points, loadings and centroids).

Regards,
Anita van der Kooij
Data Theory Group
Leiden University

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Naila Baig-Ansari
Sent: 29 June 2006 12:36
To: [hidden email]
Subject: Can factors identified using factor analysis on a smaller
dataset be used on larger dataset?


I used exploratory factor analysis to find 3 factors describing eating
patterns in healthy persons in my dataset.  I also generated factor
scores and examined the relationship of the scores obtained to a number
of demographic variables such as age, ethnicity, socioeconomic status,
etc.

 I would now like to use the 3 factors identified to my entire dataset
that includes healthy and unhealthy persons and examine relationships to
demographic variables.  How do I do this in SPSS?  Is there any way of
generating factor scores for the remaining dataset based on the factors
identified through the exploratory factor analysis?

Thanks for any insight
Naila

**********************************************************************
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.
**********************************************************************

[Kooij, A.J. van der]

Prof. Swank wrote: "The meaningfullness of the factor scores on the second
sample really depends on wether the factor structure is accurate for
that sample."
 
If you use CATPCA as described below, you can look at the component scores (object scores table, plot or saved) for the second sample (the supplementary cases) to see if they fit well in the factor structure for the healthy persons: if majority of the unhealthy persons are close to the origin, the factor structure found for the healthy persons is not
very accurate for the unhealthy persons.
To see how well supplementary variables fit in the factor structure you can look at the VAF table or at loadings including centroids plot: if the category points for a variable are close to the origin, the variable has little or no relation to the factor structure.
 
Regards,
Anita
 

________________________________

From: Kooij, A.J. van der
Sent: Thu 29/06/2006 13:37
To: '[hidden email]'
Subject: RE: Can factors identified using factor analysis on a smaller dataset be used on larger dataset?




You can do this using CATPCA if you have SPSS Categories. If you have continuous variables you can
discretize them (by using CATPCA Discretization option multiply) and use numerical scaling level.
Specify the unhealty persons as supplementary cases. These cases will not be included in the analysis, but will be fitted into the factor solution found for the healthy persons.

To examine relations between factors and demographic variables you include the demographic variables in the CATPCA variable list, and in the analysis variables list with Multiple Nominal scaling level, and specify them as supplementary variables, then they are not included in the PCA but are fitted into the found solution.

For relationships between the factors and demographic variables you can inspect the loadings plot with centroids included (centroid points are the points for the categories of the demographic variables). To examine relations between persons and variables you can inspect the biplot (including object(person) points and loadings) and the triplot (including object(person) points, loadings and centroids).

Regards,
Anita van der Kooij
Data Theory Group
Leiden University

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Naila Baig-Ansari
Sent: 29 June 2006 12:36
To: [hidden email]
Subject: Can factors identified using factor analysis on a smaller dataset be used on larger dataset?


I used exploratory factor analysis to find 3 factors describing eating patterns in healthy persons in my dataset.  I also generated factor scores and examined the relationship of the scores obtained to a number of demographic variables such as age, ethnicity, socioeconomic status, etc.

 I would now like to use the 3 factors identified to my entire dataset that includes healthy and unhealthy persons and examine relationships to demographic variables.  How do I do this in SPSS?  Is there any way of generating factor scores for the remaining dataset based on the factors identified through the exploratory factor analysis?

Thanks for any insight
Naila


**********************************************************************
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.
**********************************************************************