I recommend lavaan, which is an R package. It can run basic SEM and CFA analyses with only a very basic understanding of R. There are very good examples and tutorials on the website to help one along the way:

http://lavaan.ugent.be/start.html.

lavaan can be used to estimate advanced SEM models and is well equipped to deal with categorical data. There is good support and a lavaan discussion group where advice and support is given:

https://groups.google.com/d/forum/lavaan

LISREL is the classic structural equation modeling software. I believe the student version is free, but it is arbitrarily limited in terms of how many variables you can use and that sort of thing. Depending on familiarity with programming, I would recommend R. OpenMx and Lavaan are the R packages I'm most aware of for this purpose.

I use the latest LISREL software. I recommend you go to the site and download the free LISREL software, the student edution is an excellent free option too

Free software for SEM include Mx (stand alone program written by Mike Neale and colleagues), OpenMx (update of Mx, package through R; do a web search for the link to load the package; probably the most comprehensive package), sem (package through R), and lavaan (package through R; probably the most user friendly of the bunch).

In fact there is no free software (of quality) to perform strctural equation modelin. Although MPlus is the more complete, the analisys are performed by sintax. You'll find EQS more friendly to work with. Another suggestion is to use the AMOS program. All are paid. But I think that for the first and the last there maybe demo versions.

If you are not tied to co-variance-based modelling (for which I would recommend to use the commercial software packages AMOS or EQS), you could check out SmartPLS. SmartPLS uses a variance-based modelling approach, also has an easy-to-use graphical interface, and only requires a (free) registration at the corresponding forum (http://www.smartpls.de/forum/). Furthermore, LRP (Long Range Planning) just published a Special Issue on partial least square modelling (PLS) in management science, maybe you want to check that as a methodological introduction (http://www.sciencedirect.com/science/journal/00246301/45/5-6).

Onyx is a free SEM program. The graphical interface allows you to draw and estimate models easily. At the same time, Onyx lets you view matrix representations and generates equivalent model syntax (without introducing yet another model syntax) for various other free and commercial programs, e.g., OpenMx, lavaan, or Mplus on the fly. This allows you to transition to any of these script-based languages once your models become more complicated (e.g., need non-linear constraints, mixed continuous/ordinal estimation, etc.). As long as you are happy with maximum likelihood estimation and continuous outcomes (including multiple groups, definition variables, missing values), Onyx should be able to please you.

The program can be found here: http://onyx.brandmaier.de/

Thank you, downloaded onyx, how do i figure where the executable file is for me to run it? Thx

My students and I have used lavaan and it has worked very well for us.

Manuel

I think that SmartPLS is suitable for variance based estimation and not for the covariance structure analysis. For the latter Mplus, AMOS, EQS etc. may be more suitable. Further, the Partial Least Sqaure (PLS) based methods perhaps still lack indices for validation of the model being tested.

You can use SmartPLS. Its a free software with more than 50000 users in the world.

I personally use AMOS to analysis my data using two-steps approach (first run the measurement model and then the structural model)

I am trying this Amos Student Version to learn SEM. Wish me luck. =)

You should really consider using R and the lavaan package. It is very well maintained and evolving rapidly. The R community is very active and helping.

I recommend lavaan, which is an R package. It can run basic SEM and CFA analyses with only a very basic understanding of R. There are very good examples and tutorials on the website to help one along the way:

http://lavaan.ugent.be/start.html.

lavaan can be used to estimate advanced SEM models and is well equipped to deal with categorical data. There is good support and a lavaan discussion group where advice and support is given:

https://groups.google.com/d/forum/lavaan

AMOS which is part of SPSS but you need to purchase it separately unless your university or organization can do that.  

try Onyx http://onyx.brandmaier.de/

See this playlist at Youtube (60 videos in Portuguese):

https://www.youtube.com/playlist?list=PLhgzfaRGwuONbZ9xRpejcNmirbaeSW2ni

Best regards,

Bido

Hi.. I am doing my research using SPSS 24 and AMOS 25, but the AMOS was a trial version and it was expired before I finish my analysis. Can anyone help me out.

lavaan is fine for basic applications as long as you are comfortable with rudimentary output. Commercial software provides much better output. Caution. Some of the mentions might lead one to believe PLS and Analysis of Covariance Structure (CB SEM) are interchangeable. They are not. They have different purposes and objectives. The software may lead one to this mistaken impression.

We used lavaan for our work recently. Direct and simple effects were pretty straightforward..while latent interaction and categorical dep. variable took some reading and sifting. The lavaan community on Google groups was very helpful.

Lavaan, check it at Lavaan.org

I would recommend you SMART PLS for SEM analysis. you can down load free trial version for one month and can perform the analysis and its easier than other software.

Secondly, I will suggest you AMOS which is extension of SPSS and you can get the trial version from the website for 30 days and can download a crack version from google.

Best of luck

There are several R packages, all of them attain the 'free' criterion. Maybe you can try 'plspm' package (dedicated to partial least squares path modeling). All R-packages have instruction manuals to learn how to use them. If you don't know how to deal with R, still, you can learn to use 'plspm' from Gaston Sanchez's book ("Sanchez, G. (2013). PLS path modeling with R. Berkeley: Trowchez Editions, 383.") which is available online for free. I also recommend the youtube tutorial created by Diogenes de Souza Bido which teach how to use 'lavaan' (CB SEM) package(link in earlier responses)

Arnab Laha Dear Arnab, if you want to go free the lavaan package is the best. Although, after 6 months with that, I've installed AMOS. It's not free but in my opinion is the best you can get.

Gretl is also free. Definitively R is the best free software for statistics and data analysis in general. But the learning steep is high at the beginning. If you don't want and don't need to learn programming, you can use R via a GUI, for instance Rcommander (http://www.rcommander.com/) or Rattle (https://rattle.togaware.com/).

I recomeded AMOS

http://amosdevelopment.com/download.html