Dear Qasim,

repeated indicators approach: the indicators of the first-order constructs are used as indicators for the second-order construct.

In constrast, the two-stage approach estimates the model in the first stage without out the second-order construct. In the second-stage, the construct scores of the first-order constrcuts are used as indicators for the second-order construct to estimate the model parameters.

Best regards,

Florian

Check this out:

https://www.smartpls.com/resources/ebook_on_pls-sem.pdf

@florian

Dear Dr. Florian ,

Could you further clarify, when using two stage approach in the second stage should the indicators be connected with the variable in formative way or reflective way.

Two first order constructs (one formative and other reflective) make up the final order construct

Dear Shweta,

It depends on the assumed relationship. I guess if you refer to formative you mean a concept that is modeled as a composite, while in the reflective case, you model the concept by a common factor. Just wanted to mention that because Mode A and B are still confused with reflective and formative. From your description, it sound that you second order construct is modeled as a composite and the two first-order construct are modeled as common factor and composite.

Recently, Van Riel et al. (2017) proposed an approach to consistently estimate models containing a second-order composite formed by common factors which requires a third step. To estimate your model which contains both, you can follow their approach but set the reliability of the first-order composite to 1. We have also implemented the approach in the cSEM package:

https://github.com/M-E-Rademaker/cSEM

However, it is still under development. If you try it and face problems please let me know.

Best regards,

Florian

Van Riel, A. C., Henseler, J., Kemény, I., & Sasovova, Z. (2017). Estimating hierarchical constructs using consistent partial least squares: The case of second-order composites of common factors. Industrial management & data systems, 117(3), 459-477.

Thanks a lot Dr Florian for the clarification and resource guidance.

Much appreciated.

Will go through it and come back to you if need be.

Regards

Shweta

Dr Florian on what criterion we can decide if our model is best suited for repeated indicator approach or two stage model to run the model? like in my case I have two higher order constructs and the whole model including these two higher order construct are reflective-reflective. so Can I choose one of the approaches or there must be some guidelines that I need to follow to decide which method will be better for me. Looking forward for your reply. thanking you in advance

Dear Shoaib,

If your model contains a second-order model specified as a common factor measured by common factors and moreover no composites, I would estimate the model by maximum likelihood or a related estimator. These are consistent and (asymptotically) efficient. Why do you want to employ PLS?

That having said of course the decision also depends on the kind of research, i.e., confirmatory vs. predictive research. What kind of research are you doing?

Best regards,

Florian

Dr. Florian,

I really appreciate for your kindness and replying to my earlier post. Sorry for my delayed reply as I was struck with some work which I had to finish in a month's time. I think in your reply you suggested that according to my model and constructs composition I would be better of to use softwares such as CB-SEM or AMOS that give more accurate and consistent results for such models as compared to PLS SEM. In whole Malaysia the trend is that mostly Universities and institutions prefer PLS SEM over CB-SEM or for that matter any other software or package. Dr. I already inbox you a separate message related to two-stage model or repeated indicator approach in terms of my research. my research is more related to explanatory and predictive in nature so thats why i am using PLS SEM and besides that the other reason i already mentioned in this reply.

Once again Prof thanks for your earlier reply and apology for not replying in time.

Awaiting your kind response to my this and earlier inbox message.

Have a great day ahead

Thank you and warm regards

Florian Schuberth

Dear Dr. Florian,

I use repeat indicator for second order formative (reflective-formative), how I can know the weight of first order contribute to the second order? Because the outer weight only show the weight of indicator.

Thank you

Dear Adista,

in case of the repeated indicators approach, the weight of the FOCs building the SOC are the path coefficients of the FOC on SOC. However, if you care about consistency, you shoudl use the thee-stage approach of VanRiel et al (2017) if you have a composuite build by common factor.

Best regards,

Florian

Thank you for your answer Dr Florian Schuberth I really apreciated that