Overview

This will be a very brief introduction into custom contrasts that can be used on estimated marginal means. The examples referenced in this page are from custom contrasts in emmeans

Background

We should already be familiar with contrast coding, if not review Effect Coding in R. Essentially, we can now change when in the analysis we do contrast coding. Initially, we were taught we can do dummy coding, for example, on a factor before running a model and the output from the model would produce a y-intercept that reflects the mean of the reference level of the factor, with each beta estimate representing the effect of the level compared to the reference.

• This page is not completed- I realized there is not a need to fully understand this yet so I will plan on completing this on a later date.