canan muter posted on Thursday, April 17, 2008 - 6:20 am
Hi, In my model I have a latent variable with 6 observed variables (each formed by 5 items). But when I do CFA of these 6 factors (with 5 items for each) I get bad goodness-of-fit solutions. Do I have to parcell the items? If so, how will I decide the number of parcels? Can I do 6 parcels?