Dear all,

please forgive me if this has been asked many times before, but I
couldn't find any other info about it. BTW, is there a FAQ section?

My problem is this:
In market research we deal with many data that are batteries of
questionnaire items but where the items are coded as dichotomous
variables (e.g., 1 for "applies", 0 for "doesn't apply"). From time to
time I hear that it is possible to treat dichotomous variables as
metric variables which would allow me to make use of Pearson
correlation coefficients or even run PCA or Factor Analysis on such
data. However, I haven't found more detailed information on this.

Thus, my question(s):
(1) Is it indeed possible to treat dichotomous variables in the same
way as metric variables? I know that there are probably special
techniques in factor analysis and/or correlation (tetrachoric), but I'd
rather like to know if I can use the standard techniques without too
much loss of interpretability (so that I can use standard stats
packages).
(2) Can you point me to any references (books, articles) where this
issue is addressed?

Any comments/input would be truly appreciated.

Thanks in advance,

Stefan Ahrens
IVE Research International,
Hamburg, Germany


Sent via Deja.com http://www.deja.com/
Before you buy.


=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
                  http://jse.stat.ncsu.edu/
=================================================================

Reply via email to