ABSTRACT
This chapter improves the models discussed in previous chapters by introducing the notion of random effects. Mixed-effects models, also known as hierarchical models, allow the researcher to take into account the grouped structure in their data. Simple models are blind to the fact that participants (or groups of participants) behave differently, or that experimental items may elicit different types of responses. This can be a major issue in studies with human subjects. Hierarchical models therefore account for the intrinsic variation observed in typical datasets in the field. Once the reader is introduced to hierarchical models, non-hierarchical models will be hardly ever recommended.
