Orthogonality argumentation

[Christian]

Hi all,

When building a scale, I used component transformation matrix as confirmation of orthogonality between the factors, and it came up with a rather high value. Trying to argue for orthogonality, I turned to the factor scores coefficient matrix as the scores are standardized values, and should tell me if factors are right angled vectors right?But when running af a covariance display, there seemed to be a trivial value for covariance but nevertheless it shows. Why is that? To me very low coefficients for covariance between the two toploading variables in each factor should mean the factors are uncorrelated and thus orthogonal or?

Christian Kobbernagel
Communication Roskilde University Denmark

[statisticsdoc]

Christian,

What rotation did you use?  What factor score computation method?

Steve

For personalized and professional consultation in statistics and research
design, visit
www.statisticsdoc.com


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of
chko
Sent: Friday, December 29, 2006 12:02 PM
To: [hidden email]
Subject: Orthogonality argumentation


Hi all,

When building a scale, I used component transformation matrix as
confirmation of orthogonality between the factors, and it came up with a
rather high value. Trying to argue for orthogonality, I turned to the factor
scores coefficient matrix as the scores are standardized values, and should
tell me if factors are right angled vectors right?But when running af a
covariance display, there seemed to be a trivial value for covariance but
nevertheless it shows. Why is that? To me very low coefficients for
covariance between the two toploading variables in each factor should mean
the factors are uncorrelated and thus orthogonal or?

Christian Kobbernagel
Communication Roskilde University Denmark

--
View this message in context:
http://www.nabble.com/Orthogonality-argumentation-tf2895461.html#a8089680
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

[statisticsdoc]

Christian,

I am assuming that you used an orthogonal rotation of the factors, and that
you are wondering why the factor scores show a small correlation.  While the
orthogonal factors are in principle not correlated, the factor scores are
estimates, and under many methods of estimation they do show a very small
correlation.

HTH,

Stephen Brand

For personalized and professional consultation in statistics and research
design, visit
www.statisticsdoc.com


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of
chko
Sent: Friday, December 29, 2006 12:02 PM
To: [hidden email]
Subject: Orthogonality argumentation


Hi all,

When building a scale, I used component transformation matrix as
confirmation of orthogonality between the factors, and it came up with a
rather high value. Trying to argue for orthogonality, I turned to the factor
scores coefficient matrix as the scores are standardized values, and should
tell me if factors are right angled vectors right?But when running af a
covariance display, there seemed to be a trivial value for covariance but
nevertheless it shows. Why is that? To me very low coefficients for
covariance between the two toploading variables in each factor should mean
the factors are uncorrelated and thus orthogonal or?

Christian Kobbernagel
Communication Roskilde University Denmark

--
View this message in context:
http://www.nabble.com/Orthogonality-argumentation-tf2895461.html#a8089680
Sent from the SPSSX Discussion mailing list archive at Nabble.com.