Hi Jiwat,
Answer this question:
1. You say for Bido, you test outer model (CFA) using CB-SEM (AMOS) and inner model using PLS, it's possible? You think ML and LS is same (CB-SEM and PLS-SEM same)?
2. You duplicated latent variable become latent score for handle reciprocal in PLS. You think latent variable identical with latent score?
3. I argue for your connect Latent Variable E to other Latent Score A, B, C, D. You have reference for it?
4. You have latent variabel A, B, C, D and E indicators reflective. For latent score reciprocal must used mode B, It's same or different?
Regards,
Hengky
modeling and estimating two-way / reverse effects
Thanks Hengky.
1. Answer is No. When I am testing I am assessing both measurement model and structural model in the same application. So when I test my model I test both measurement and structrual model in PLS-PM.
Or when I test my model in CB-SEM, I test both my measurement and structural model in CB-SEM.
2. I am not sure how my duplicaed LV becomes latent score. I am running both original LV (A) and the duplicated LV ( say A1) in the same model at the same time. I am not using Latent scores of A to make a duplicated variable A1.
PLS modeling does not stop one to have two variables loaded with same indicator items. So in my case, both A and A1 have same indicator items. Only difference is that my A is directed to E and E is directed to A1. So I have two relationships, and one of which (E-A1) is reciprocal relationship.
3. That is not the issue at the moment, as I understand these relationships need to be derived by theory. Important thing to me is to decide whether PLS-PM is a good way to estimate reciprocal relationships.
4. All LVs have reflective indicators.
:)
Regards
Jiwat
1. Answer is No. When I am testing I am assessing both measurement model and structural model in the same application. So when I test my model I test both measurement and structrual model in PLS-PM.
Or when I test my model in CB-SEM, I test both my measurement and structural model in CB-SEM.
2. I am not sure how my duplicaed LV becomes latent score. I am running both original LV (A) and the duplicated LV ( say A1) in the same model at the same time. I am not using Latent scores of A to make a duplicated variable A1.
PLS modeling does not stop one to have two variables loaded with same indicator items. So in my case, both A and A1 have same indicator items. Only difference is that my A is directed to E and E is directed to A1. So I have two relationships, and one of which (E-A1) is reciprocal relationship.
3. That is not the issue at the moment, as I understand these relationships need to be derived by theory. Important thing to me is to decide whether PLS-PM is a good way to estimate reciprocal relationships.
4. All LVs have reflective indicators.
:)
Regards
Jiwat
-
Hengkov
- PLS Super-Expert
- Posts: 1599
- Joined: Sun Apr 24, 2011 10:13 am
- Real name and title: Hengky Latan
- Location: AMQ, Indonesia
- Contact:
Hi Jiwat,
1. For test outer model reciprocal using PLS not possible.
2. You used original latent variable for duplicate? I think latent score. So, you number of indicator and number of variable in model PLS different
with your model research. It's acceptable?
3. Because you used same items for original and duplicate variable (A, B, C, D) I think is bias method.
4. For your reason A => E = 0.3 and E => A = 0.54 not change if delete reverse link, because PLS algorithm remains the same for causal feedbacks.
1. For test outer model reciprocal using PLS not possible.
2. You used original latent variable for duplicate? I think latent score. So, you number of indicator and number of variable in model PLS different
with your model research. It's acceptable?
3. Because you used same items for original and duplicate variable (A, B, C, D) I think is bias method.
4. For your reason A => E = 0.3 and E => A = 0.54 not change if delete reverse link, because PLS algorithm remains the same for causal feedbacks.