forum.smartpls.com

Dear all,

I really want to apply PLS, but my sample size on the current study is approxomately 5000.
I will do multigroup analysis, where the segments consist of 1700 respondents in each.

Are there any other papers using PLS on such large sample?

And are there any reasons for not using Smart pls in this case?

I am very much looking forward to your respons.
Best regards, Maria

PLS works well on large samples, but depending on your model and research goals you might also consider CB-SEM.

There are ASCI datasets that are analyzed with PLS and have 10,000 observations. But I don't have a specific reference at hand. Maybe look at the Fornell studies and some of the paper on unobserved heterogeneity that use the ASCI as an example dataset.

Thank you so much for your reply.

I chose PLS sem because some of my variables was skewed, and consequently, the data was not normally distributed. Is this good enough reason to justify the use of PLS contra CB-SEM?

I appriciate any comments:)

Well, there are also CB-SEM estimators that can handle non-normal data, especially if you have very large sample sizes. Depending on the reviewer of your paper you might get problems if that is your only reason to use PLS.

Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. European Business Review. doi:10.1108/EBR-11-2018-0203