Can you suggest some covariance based SEM software to analyse formative constructs? I used to use AMOS. But suggestions I got is that, AMOS is not built to deal with formative constructs. What do you think? Do you have any other suggestions? Thank you very much.
Regards, Catherine.
Hey Catherine,
using R with the Lavaan package is able to this.
Here is a link to R:
http://www.r-project.org/
And here is a useful link for the the Lavaan model syntax
http://www.inside-r.org/packages/cran/lavaan/docs/model.syntax
Hope this is what you are looking for.
Regards, Timo
Hi Timo, many thanks for your suggestion!
Best regards,
Catherine
I would suggest you to go for smartPLS. a very user friendly interface and easy to understand,
Hi Faizan,
Thank you for your suggestion. I am agree with you. SmartPLS is useful and easy to use esp. when having formative constructs. But it is a PLS-SEM software not covarience based SEM. I understand when having formative constructs then go for PLS-SEM. But I want to learn incorporating formatives into CB SEM. In theory, I know how to do it. But I used AMOS. A scholar told me AMOS is not built for dealing with formatives. So I m looking for suggestions for other CB-SEM softwares. Thank you Faizan for your input.
Best regards,
Catherine
Maybe the following paper will help you and maybe it also shows you that you would be better off using PLS over CB? Why do you want to stay with CB? Is there a certain reason for this?
Hi Charlott,
Thank you for the paper. The project i was working on was required to use CB not PLS. Attached two papers explain criticisms of PLS.
Dear Catherine,
Please check this page:
http://scaleresearch.siu.edu/petpeeve9.html
Many thanks for your contribution, Farimah. It helps a lot. Thank you!
In case you have a higher order construct such as overall job satisfaction which you might be thinking to conceptualise as being formatively related to its three or four lower order constructs, then what you do is to link those sub-dimensions upward to the higher order construct. This will enable you to relate subdimensions formatively to higher order construct. Now operationalise your higher order construct by a minimum of two or three reflective scale items. Your lower order constructs too can be operationalised reflectively in a similar way. Now test the model.
Wishing you a good luck.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling