Units of total effect in path with lo... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 victor china posted on Thursday, October 08, 2015 - 1:36 am
i have a path with a set of continuous variables that predict a categorical variable

for example : (a,b,c continuous and AA categorical)
a->c->AA
b->c->AA
b->AA

what are the units of the total effects? are they also odds ratio? what if they are negative? how can they be converted to odds ratio?
 victor china posted on Thursday, October 08, 2015 - 1:39 am
maybe it is log odds?
 Linda K. Muthen posted on Thursday, October 08, 2015 - 11:55 am
Any regression of AA on a covariate is either a probit or logistic regression depending on the estimator you use. All of the regressions of c ON a covariate are linear regressions.
 victor china posted on Thursday, October 08, 2015 - 9:06 pm
thanks!
so if for example:
c on a (a_c)
AA on c (c_AA)
AA on a (a_AA)

model constraint:
new (AA_total)
AA_total = a_c*c_AA + a_AA

would AA_total be a logistic regression coefficient? (given estimator=mlr)
 Bengt O. Muthen posted on Friday, October 09, 2015 - 8:51 am
A categorical outcome in a mediation analysis needs special attention - see the article on our website

Muthén, B. & Asparouhov, T. (2015). Causal effects in mediation modeling: An introduction with applications to latent variables. Structural Equation Modeling: A Multidisciplinary Journal, 22(1), 12-23. DOI:10.1080/10705511.2014.935843
 victor china posted on Monday, October 12, 2015 - 2:05 am
thank you!
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: