I'm running a model using glmmadmb and I'm not sure how to do multiple comparisons on an interaction between categorical and continuous predictors.
Here are details on the model and data:
My dependent variable is count data (based on participants responses), and I'm using glmmadmb to use a Poisson distribution with zero inflation.
The predictors are:
Score_A, Score_B, Score_C, Score_D: four continuous predictors each corresponding to participants' scores on four tasks, ranging from 1 to 9.
Condition: a categorical variable indicating the condition participants were in (three levels: cond_1, cond_2, cond_3).
I set up my model to check for the main effects of each Score on the dependent variable, as well as the interaction between each Score and Condition (I don't care about an interaction between different scores). My model is:
mod <- glmmadmb(dv ~ Score_A*Condition + Score_B*Condition + Score_C*Condition + Score_D*Condition, data = d, family="poisson", zeroInflation=TRUE)
First I tested the significance of the interactions by comparing the model above to a model without interactions:
mod_no_interaction <- glmmadmb(dv ~ Score_A + Score_B + Score_C + Score_D, data = d, family="poisson", zeroInflation=TRUE) anova(mod_no_interaction , mod)
This led to a significant difference, so I now want to check multiple comparisons to see how the effect of each Score variable varies by Condition (e.g., is the effect of Score_A different for cond_1 vs cond_2, for cond_1 vs cond_3, etc).
However I'm not sure how to do this. I found this resource on multiple comparisons for a continuous vs categorical predictor interaction:https://mailman.ucsd.edu/pipermail/ling-r-lang-l/2012-November/000393.html
But since my model is much more complicated and has multiple continuous predictors I'm not sure if I can use the same method.
Any advice is welcome, including if you think I'm using the wrong model or should use a different package.