The help file and textbook are not useful in actually implementing SEM in practice. Neither can I find any Statistica tutorials or videos. Are there any good resources that others have found?
Hi, have you checked the online help document (also available in the e-manual) regarding Structural Equation Modeling (SEPATH) Analysis? Examples are also included there. Let me know if any additional assistance is needed. Thank you!
Yes, but those examples are really poorly documented. They only show the outcomes, not the process. There is a need for more completely documented tutorials that elaborate on the rationale for selecting the different types of variables (confusing!) and the process for doing so with Statistica (klunky user interface). The help file is not well written either.
There are online youtube videos for R, SPSS, AMOS, etc. But nothing for Statistica.
I also see the Statistica user interface for SEM has not had any software development in ages: seems a bit neglected by the developers. It seems one still has to type in variable numbers (can't select them, and the wizard is very confusing as it fails to adequately differentiate between the different types of variables.
Any suggestions for better tutorials?
In my dissertation, I utilized Structural Equation Modeling (SEM) for my dissertation on the epistemology of medical knowledge, and was able to verify the numerical accuracy of the Statistica platform for SEM. We could list the pros and cons for each applicaiton since I used AMOS, but Statistica has a robust and accurate SEM module.
I did use AMOS for my dissertation, which is an IBM product. The issue is, you don't get AMOS when you get SPSS. You have to buy them separately. AND AMOS data doesn't work with SPSS data. That is a huge disadvantage, especially if you want to save money, or be economically practical.
And yes, R has a module for AMOS, if you can code in R. Statistica does integrate with R however. See below KB articles about more details of R integration in Statistica:
Which versions of R to integrate with Statistica?
Writing R scripts in Statistica
It comes down to what you want within a platform. Given unlimited resources, I would imagine that Statistica would have the best resource for SEM out there. If you look at the roadmap, indatabase processing, data mining, edge analytics and HDFS compatabilities seem to be where development is focusing. If the customer base supported SEM, it would shift.
In closing, I was able to follow the help documentation very well, and while it isn't a textbook (nor should it be) it was pretty helpful in regards to SEM methodologies. Where do you see the limitations?
And here are two knowledge base articles about integrating R and Statistica: