Hello Jourdan,

Is it essential to model simultaneously effects on more than one DV? If not, ordinary logistic regression software will address the kinds of variables you describe.

No, AMOS does not handle constrained variables very well. While you can generate bootstrapped estimates of errors, do note that these are not obtained by a mechanism which presumes that the true values cannot exceed the nominal limits (typically: 0,1 or 1, 2), This is a problem, for example, when trying to apply linear regression with a DV for which a logistic (or other, non-OLS) model would be called for.

The freely available R library, lavaan, can handle ordinal variables (and your DV may be considered as ordinal). The freely available program, Jamovi, incorporates lavaan and makes data entry and manipulation easier than via straight R (or even R studio interface).

Many will also mention Mplus, the commercially available software, as its ability to handle dichotomous and/or polytomous variables is one of its chief selling points. However, it's not free, nor is it low-cost (in my opinion).

Good luck with your work.

Thank you for your answer, David. I appreciate it!

While it would be interesting to examine the relationship between the dependent variables for theoretical reasons, I will use logistic regression to explore different models for the dependent variables. I will look into Jamovi and R for future projects. MPLUS seems like a great program. But, as you mentioned, it is just too expensive.

Hello Jourdan,

As an alternative to AMOS, you can use lavaan package on R for free. Many tutorial videos of applying lavaan in an online platform which have growing over the last years, and to add to this, JASP is a free software too, which has a lavaan based option to run your abovementioned SEM models within it.

Hope this helps.

hi,

my understanding is that AMOS uses parametric estimation techniques. Hence, it has not been developed to use categorical outcome variables. Mplus is ideal for this purpose.