I am running several multi-level models. The multilevel aspect is not that important here though, this question would be the same even if it was simple regression. Basically, randomly assigned condition (control, intervention 1, or intervention 2) is the predictor for a series of quantitative mental health outcomes (e.g., depression; one model per outcome).
My reference group is the control. I have my results comparing control vs intervention 1, and control vs intervention 2, but it would not be possible to also calculate results for intervention 1 vs intervention 2, because of limitations posed by my degrees of freedom correct? Just wanted to double check I am not misunderstanding regression/reference groups/dummy coding.
This post shares how to get the confidence levels for pairwise comparisons: Categorical variable coding to compare all levels to all levels However, I don't want the confidence interval, I just want the beta-coefficient and p-value.
