question on factor analysis

[Rod Turner]

I am having a discussion with colleagues over the inclusion of dichotomous variables and Likert scale variables in factor analysis. I feel uncomfortable with the inclusion of different scales in the same analysis but I seek advice on this.

Thanks

Rod

====================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 Greenberg]

Technically, the observed variables in factor analysis should be interval-level. Many people ignore this. Some programs, such as Latent Gold, allow you to specify the kind of variable you have, including dichotomous and ordinal variables. David Greenberg, Sociology Department, New York University

----- Original Message -----
From: Rod Turner <[hidden email]>
Date: Thursday, November 8, 2007 4:01 pm
Subject: question on factor analysis
To: [hidden email]

> I am having a discussion with colleagues over the inclusion of
> dichotomous variables and Likert scale variables in factor analysis. I
> feel uncomfortable with the inclusion of different scales in the same
> analysis but I seek advice on this.
>
> Thanks
>
> Rod
>
> fffffff
> 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

[Kooij, A.J. van der]

With CATPCA (nonlinear PCA or PCA for Categorical data) in SPSS Categories you can do PCA if the variables are categorical or of mixed measurement level (unordered and ordered categorical and numeric)

________________________________

From: SPSSX(r) Discussion on behalf of Rod Turner
Sent: Thu 08/11/2007 21:47
To: [hidden email]
Subject: question on factor analysis



I am having a discussion with colleagues over the inclusion of dichotomous variables and Likert scale variables in factor analysis. I feel uncomfortable with the inclusion of different scales in the same analysis but I seek advice on this.

Thanks

Rod

=======
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



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

====================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]

         Rod,
         There is a tradition of using Likert scales without compunction,
and many people also use dichotomous variables in factor analysis. Both
practices are objectionable: Likert scales are ordinal, not interval;
dummies are interval, but the residuals are not normally distributed.
         A better alternative is using Categorical Principal Component
Analysis (CATCPA in SPSS, Categories Module), based on an Alternating Least
Squares algorithm. CATCPA accepts all sorts of variable (interval, ordinal,
multi-category, dummies). It estimates numerical values for categories of
ordinal and categorical variables, and computes factor analysis. Alternating
least squares proceeds by iteration: it takes whatever initial numerical
codes are given to categories (say 1,2,3,4,5) and computes an initial
solution, then estimates new values for variables based on factor scores,
and produces a second factor analysis based on this new set of variables'
values, and so on until stabilization.

         Hector

         -----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Rod
Turner
Sent: 08 November 2007 17:48
To: [hidden email]
Subject: question on factor analysis

         I am having a discussion with colleagues over the inclusion of
dichotomous variables and Likert scale variables in factor analysis. I feel
uncomfortable with the inclusion of different scales in the same analysis
but I seek advice on this.

         Thanks

         Rod

         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

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