There are different software to carry out confirmatory analyzes, such as SPSS AMOS, which sometimes are insufficient for the publication requirements of some journals. What would be the most recommended software today?
Lavaan is operationalised in GUI software JAMOVI and JASP. both are free and built on top of R. Alternately package sem has been built into RCommander another GUI for R. All are good enough for publication and FREE.
Free: The R-package lavaan (http://lavaan.ugent.be/tutorial/index.html ). It improved a lot over the years and allows for CFA, multilevel CFA, invariance analysis, etc. Great package and open source
Paid: MPlus is a very powerful program and the authors provide ample online support. Syntax is relatively easy to learn and flexible.
Support: Onyx is a nice and simple program able to help build models graphically and to generate lavaan/mplus-syntax. http://onyx.brandmaier.de/
These two are popular and reliable. I don't really understand why AMOS should be seen as insufficient for publication, but given that LISREL and EQS are not really supported anymore, MPlus (and lavaan) stand out as strong options
Firstly, I am not a fan of IBM. However, AMOS is sufficient enough for publication. Some values such as CR and AVE may need to be calculated manually (I am not sure whether it was recently added or not).
If you are looking for alternative programs to AMOS, you can use JASP. JASP is R-based program and easy to use. LISREL is an analytical tool, but it cannot be said to be user friendly.
To the best of my knowledge, AMOS, Smart Pls and LISREL are good enough for confirmatory factor analysis, however if your data is not normally distributed, Smart pls is recommended.
neither PLS nor smart PLS specify common factor models (but composites). Neither (to my knowledge) can they test the the implications of the models (it may be that the newer form--"consistent PLS" can do that, please check).
In the case of non-normal data, SEM (as implemented in lavaan and Mplus, but I do not know whether this is the case in AMOS, anyone?) allows to apply corrections of the test and standard errors (e.g., Satorra-Bentler correction or Yuan-Bentler corrections). Again no need to switch to PLS :)
Lavaan is operationalised in GUI software JAMOVI and JASP. both are free and built on top of R. Alternately package sem has been built into RCommander another GUI for R. All are good enough for publication and FREE.
When running complex factor models and path analysis with large amounts of data, does anyone recommend using a particular computer / laptop processor (e.g., intel core i5 or i7)?
I do not think that there are differences between established latent variable software packages (proprietary, e.g. LISREL, EQS, AMOS, Mplus; in R: e.g., lavaan, openMX, sem) in terms of meeting journal requirements. Given the identical data and specification of the analysis, all of them will deliver consistent estimates. There are differences in the packages' analytic capabilities, though (e.g., available estimators). A detailled comparison can be found, e.g., in Rick Hoyle's (2014) "Handbook of Structural Equation Modeling". For publication, its more important to follow reporting standards.