ApolloChoiceModelling forum

Dear all,
I have run a Hybrid Choice Model with a latent attitude.
I have used items from a validated attitudinal scale (using Ordered logit estimation for the latent var). Only 4 out of 6 items are included in the HCM (with all 4 thresholds for each item are significant).
Do I need to report the results of the CFA of the latent variable?
Thanks,
Anat

Hi Anat

some papers do, some don't, so this depends on you and also the target audience, I guess

Stephane

Thanks!

Dear Stephane

For simultaneously estimated HCM, do the latent variables need to meet CFA criteria (e.g., CR, AVE)? What methods can be used to demonstrate the construct reliability of latent variables?
Thanks a lot!

Hi

some authors pay attention to this and report it, others do not. Of course, you need to also think about whether paying too much attention to the grouping of indicators on their own (rather than looking at it jointly with the choice data) reduces the quality of your findings for the hybrid model itself

Stephane

Thank you very much!