gianna posted on Sunday, August 22, 2010 - 10:41 pm
I am currently doing the cfa of a questionnaire with 130 dichotomous items which are known to be divided under 9 factors.
These 9 factors are divided in three clusters.
I try to do the cfa of these one-two-three cluster models. There is no problem when I take the overall mean of the factors as observed variables but when considering that the items are dichotomous, i need to parcel them.
My question is how will I compose a model with factors which items are parcelled and which belong to different latent clusters.