I have several significant 3-way interactions in a 2-level models (individuals nested in couples), and I am struggling with the computation of the simple slopes. Essentially, my categorical level-2 variable (couple is classified as distressed/violent or distress/non-violent) moderates the level-1 cross product of gender and satisfaction to predict conflict behavior. I am using HLM software, and I've always used Preacher's website to compute my simple slopes, but he doesn't have any options for this combination of level-2 and level-1 interactions.
I'm not finding a ton of information online. It seems like some (most?) people just run two sets of analyses for each of the groups, but that seems to be an inelegant solution at the very least.
Edit to clarify: See my model, and then my output. I have data from ~85 couples (170 individuals) who completed several questionnaires during a single lab visit.
The level-1 variables ending in "_S" are cross-products (computed in SPSS) between "PARTNER" (female or male) and the three continuous level-1 predictors (i.e., FLOOD, DAS, PCLEAD). "COU_TYPE" (violent or non-violent) is a between-level (level-2) predictor.
In several instances, "COU_TYPE" is a significant moderator of those level-1 cross-products (circled in red).
Is there a way to compute the simple slopes/effects for men in the violent group, men in the non-violent group, women in the violent group, and women in the non-violent group?
