Page 1 of 1

Cross-validated redundancies and cross-validated communaliti

Posted: Mon Oct 30, 2006 4:47 pm
by yusofahmad
Hi all,

I had a few questions and would be grateful for any answers to the same

1. is the cross-validated redundancy for an endogenous variable the same as the Stone-Geisser test and therefore should exceed zero if the var is to be retained

2. What if the cross-validated redundancy for an indicator is less than zero? What of the CV-redundancy for exogenous LVs?

3. What is the cross validated communality for indicators and constructs and how is it to be used?

I have access to very limited literature and cannot look up the material in the "literature" section on this. Would be grateful if anyone answers this ASAP.

Thanks in adnvance,
Yusof

Posted: Wed Nov 08, 2006 7:00 pm
by cringle
Hi,

check the article by Tenenhaus et al. (2005) in the literature section (viewforum.php?f=7). You will find very good answers to your questions and a lot more! This aricle can be orderd online or you might find a similiar HEC Research Paper in the Internet.

Best
Christian

cross-loading ??? minus

Posted: Tue Nov 14, 2006 3:14 am
by JAFLAH
Hi All
I have problem with the cross loading one of may construct (LV - latent construct) have a minus (ex. -0.78,) cross loading. However with SPSS the loading was plus and not high (.657, .66, .7.....). The measure that I’m using is well established in the literature.

Posted: Tue Nov 14, 2006 10:11 pm
by cringle
Hi,

this usually turns out to be a user problem. There should not be any differences. Check this topic:

viewtopic.php?t=268&highlight=cross+loadings

recommended reading

Posted: Thu Nov 16, 2006 9:59 am
by yusofahmad
Does the recommended reading also deal with indicator CV red. and CV com.?