Experimental conditions predict dichotomous outcome with mediation and moderation

32 views

dichotomous_outcomelogistic_regressionmediationmoderation

Skip to first unread message

Carina Hörst

unread,

Jun 13, 2019, 9:06:20 AM6/13/19

to lavaan

Hi all,

I try to use the following:

I have an experimental design (0,1) predict a dichotomous outcome (0,1) via logistic regression. That is the main effect. However, I also need to apply a mediation and a moderation on that model. So far I only found the "ordered" function which doesnt seem to make much sense in my model. Also, I have not found any practical guide how to apply the mediation/ moderation in the next step.

Please, I am happy for any tips.

Thank you.

Carina

Terrence Jorgensen

unread,

Jun 17, 2019, 4:58:42 PM6/17/19

to lavaan

So far I only found the "ordered" function which doesnt seem to make much sense in my model.

That is how you tell lavaan an outcome is binary/ordered. Binary is a special case that can be arbitrarily interpreted as nominal or ordinal. In order to have binary outcomes in SEM, you must assume it is a dichotomized realization of a normal latent variable (i.e., 0 and 1 are ordered categories, such that people scoring 0 have more of a quality than people scoring 1).

Also, I have not found any practical guide how to apply the mediation/ moderation in the next step.

https://www.statmodel.com/Mplus_Book.shtml

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam