ABSTRACT
This chapter provides an introduction to addressing the non-independence among a set of observations that is obtained within persons. Mixed-effects models are among a more recent generation of statistical models than many of the other models presented thus far. These models have received a great deal of attention in the methodological literature and are widely used for data analysis and theory testing in a wide variety of areas. An important part of the mixed-effects model is that not only does the model allow for different trajectories across individuals, but it also allows these differences in trajectories to be modeled in terms of individual characteristics. Trajectory plots can provide an essential supplement to the numerical information we obtain by fitting models. The mixed-effects approach treats random effects differently. In this approach, random effects are distinguished from fixed effects prior to parameter estimation. The mixed-effects approach used for mixed-effects models, on the other hand, requires a distributional assumption from the start.
