Can I use AMOS to perform SEM model with a dummy dependent variable? if NOT would you please suggest any another statistical package?
My dependent variable is a binary variable consists on 0 and 1 only.
A similar question was asked in this forum a few years ago:
https://www.researchgate.net/post/Can_anyone_help_with_path_analysis_with_a_logistic_regression_model
Responses suggested using a different SEM package, MPlus.
In short, if you use the basic settings in your SEM program then you are attempting the equivalent of OLS Regression on your binary dependent variable. You'll get an answer, but predicted values can be less than Zero or greater than One. What you need is a model that gives you an S-shaped curve ranging from Zero to One, with predicted values effectively being probabilities of One. A Logistic Regression.
The discussion on Bayesian estimation and Categorical Response Models here is very good:
http://www.structuralequations.com/AmosTutorials.html
You just have to conduct a logistic regression. It is easy with MPlus but it is possible with Amos. See this link and this article: http://www.e-ijaet.org/media/43I6-IJAET0612656.pdf
Hi Rebean ,
Also you may be use Structural Equation Modeling with Partial Least Square - The SmartPLS is very friendly.
If you need i will give more details.
Hope that help
Best Regards
Dirceu
Dear Hume, thank you very much for your quick comment and helpful link.
I have a sample of 600 observations, and my dependent variable is binary, it is not continuous (1= has the intention, 0= has no intention).
Does this mean that I can still use the Bayesian Estimation?
Dear Claude, thank you for the paper.
I have no experience with MPlus, that is why I am trying to do it with either Amos or STATA.
Dear Dirceu, I would be grateful if you could provide me with some more details on how to apply that in the SmartPLS.
Although SmartPLS is not provided at my University, I will use the Free 30-Day Trial to see how it works.
Dear Rebean,
Another way is to use R software (free and open source). All analysis are possible, especially with the package LAVAAN (http://lavaan.ugent.be/tutorial/tutorial.pdf): CFA, SEM, logistic regression... It is a little bit difficult at the beginning because the software design is totally different than SPSS, AMOS, LISREL or others. But finally, it is really powerful and adapted to every kind of analysis.
For future researches, there are a lot of different packages and impossible don't find one for your need.
Best
Hi Rebean,
You can use the SmartPLS 2.0 version - It is free! - go to the bottom of the download page. There you will find the link.
Going to your problema. You can generate a new construct in your model and should put the dummy variable in this new cosntruct. See the figure added. It is very simple.
To use the smartPLS:
https://www.researchgate.net/publication/281448905_STRUCTURAL_EQUATION_MODELING_WITH_THE_SMARTPLS?_iepl%5BviewId%5D=b7uuM03VWfKVZyoBscAg2nFp&_iepl%5BprofilePublicationItemVariant%5D=default&_iepl%5Bcontexts%5D%5B0%5D=prfpi&_iepl%5BinteractionType%5D=publicationTitle
Best Regards
Dirceu
Dear Claude Houssemand, please share the link to the article you suggested again or provide the reference.
You guys may read this paper. It's very helpful: Marketing research using single-item indicators in structural equation models https://link.springer.com/content/pdf/10.1057/jma.2013.7.pdf
- Badges
- Science topic
- Similar topics
- Psychology