Hello @Sara If your sole objective is to do CFA and validation then Smart-PLS wouldn't be right choice as it's not covariance based package and doesn't run a reflective model alone. Given your high number of first-order factors (15), I guess there are plenty of observed items as well to load with respect latent factors, hence AMOS possibly wouldn't also work well. I suggest you to go with R as it's open source and one of the most powerful statistical softwares as present and doesn't also suffer with the limitations of sample size, and large model identification. Wish you luck with your work.

I also would suggest R cause you will have a more flexible set of estimators. Mplus is very good too, but it's not free

I would prefer mplus, although R is not inferior. Both e.g. have strength of the Wlsmv-estimation, which is excellent for non-normal or ordinal data. If you only want to estimate CFA models and apply Maximum likelihood estimates i really like Amos, especially if you have no prior experience in SEM. Amos has excellent features, e.g. bootstrapping . I don't agree, that Amos might not work well in case of many variables, Amos has excellent estimation procedures. But Mplus and R allow to adopt interesting alternative estimations and to combine cfa with e.g. mixture modelling or iten response models.

Thank you so much fr your guidance Imran Anwar and Euclides José de Mendonça Filho

I see. Thank you so much for your insightful comment Markus Antonius Wirtz

Depends what the other people in your research group use, and therefore can help with questions on. When using any new software you are likely to have questions that are much more easily answered by people you know and understand your research topic than others, like folks on RG or stackexchange.

R is great. There are around 20 packages for CFA in R to satisfy the different requirement you likely face.

The advice of having someone to cooperate with when using software is important. Mplus have a lot of very good instruction stuff on their web site www.statmodel.com, including systematic teaching programs on videos. There is a demo which you could try. I think it is the most versatile, have the most possible data formats and possibilities. However, just as Amos, Mplus are not for free. Choosing among the two I really would favor Mplus.

R is a toolbox where anything is possible, but complicated. Very many researchers have constructed methods from R. However, you will not know if procedures wil be updated. Lavaan is one of these. JASP is an interesting and easy, but limited, alternative.

I liked SPSS & Amos. Of course those are costly but I'm sure that your U has the licences. The most important issue is that you should get help when you are getting problems - as everybody gets ;) Find out for good theory books which guiding you!

Thanks, Jan Ivanouw . That was quite helpful. Well, I have started working and so far so good. I would love to have a collaborator, but I'm afraid finding someone would be hard. I've been sending emails to a few people, haven't received a positive response yet. DO you know any type of forum or website I can find a probable collaborator?

Thanks Otto Ilmari Burman for your insightful suggestion.

Use JASP. It is open source. You don't have to spend hours making relations as in Amos or write any code. It is click and drop.

One important thing is to understand the data and the correlations in it. First of all you have to have enough large data!

Click and drop do not give you the opportunity to inspect the real data... And you can read a bit more deeper things related to factors and correlations conserning MTMM-martrix

Henseler, J., Ringle, C.M. & Sarstedt, M. J. 2015. Acad. Mark. Sci. 2015: 43: 115.

https://doi.org/10.1007/s11747-014-0403-8

This gives you some good hints to understand the factorial thinking :)

You can get my excel-file from:

https://ottoburman.fi/test/htmt.xlsx