Hi Anirban,

in this thread, I posted a few suggestions and recommendations. https://www.researchgate.net/post/Dear_scholars_how_to_design_a_Structural_Equation_Modeling_surveys

Beyond that I would start with Judea Pearl's Book of why followed by Rex Kline's Principles and Practice of Structural Equation Modeling (important: get the 4th edition).

Bill Shipley's Cause and Correlation in Biology is also great (don't get distracted by the biology-focus).

And if you allow me a personal note: Learn R (good book is Hadley Wickhams's "R for data science") and do SEM in the lavaan-package of R (www.lavaan.org). You have to invest a bit in the beginning but get out very much. Avoid SPSS and AMOS :)

Best,

Holger

Holger Steinmetz Dear Sir, thank you for the suggestions. Yes, I will be doing it in R.

I recommend the book of Structural Equations of Gana y Broc (2019).

Hi Anirban,

As Holger has indicated, there are lots of good books, journals and web resources that are easily available. Indeed, there is an Mplus channel on youtube. What I would have given for these resources 20 years ago!

I would add that attending training courses on SEM/LVM is a valuable way to learn. There's nothing better than spending a few days, uninterpreted, getting exposure to this material. SEM is a combination of regression and factor analysis (crudely put) so there is no natural starting or end point - you just have to keep learning things until it all comes together and makes sense.

There used to be a list of courses/workshops on the web that was maintained by Jeremy Miles - this may be long dead as there are too many such courses to list. However, I hear that the summer school courses at Ulster University on SEM and other multivariate methods are particularly good ;-)

Best wishes

mark

Cristian Ramos-Vera and Mark Shevlin thank you for the suggestions. I will surely follow these.

Dear Anirban,

I personality think that you don´t have to in advance descart any software, first of all, because there may be viewed as statistics tools, tha can be useful in some context. This does not imply that software such as R, M-plus, and FACTOR (for alternative CFA approach) have wider (and robust) options according to most of the data that is usually worked with. In fact, I recommend this. Nonetheless, some soft my be more friendly for beginners, because not use syntax format. This is the case of AMOS. One book that you can use for this last program is "Structural Equation Modeling with AMOS: Basic Concepts, Applications, and Programming" from Byrne (2001). In any case, all of your desition (including software, that is, basically, methods for estimations) have to be properly justified.