I am trying to conduct a path analysis with 1 binary outcome variable (DV), a continuous mediating variable, a dichotomized mediating variable, 3 continuous IVs, 3 dichotomized IVs, and 4 ordinal/nominal IVs. Which statistical software to use in this case?
Hello Habib,
A number of software packages can estimate SEM model parameters for ordinal/nominal variables. Have a look at some of these:
1. Mplus
2. LISREL
Article Structural Equation Modeling with Ordinal Variables using LISREL
3. Up to a point, lavaan (in the R system): http://lavaan.ugent.be/tutorial/cat.html
4. If it makes conceptual sense to do so, PLS-SEM (partial least squares sem) may also be used.
Here's a link to previous thread asking a similar question: https://www.researchgate.net/post/What_is_the_best_way_to_conduct_path_analysis_with_an_ordinal_end_variable
Good luck with your work!
If you are an ancient guy like me a logistic regression program should work. That's how Wright would have done it if logistic regression had existed in 1918. See his 1918 paper referenced at this link: https://en.wikipedia.org/wiki/Path_analysis_(statistics)
Best, D. Booth
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics