Mixed-effects modelling: Overrated strengths and underrecognized potential

Mixed-effects modelling has been widely used in psychology and neuroscience to analyze clustered data. It is used so frequently that it almost becomes a default option when data are clustered. However, I first argue that mixed-effects modelling can be substituted by simpler approaches in some situations (e.g., summary-statistics approach) and we need to understand when it is really needed. Understanding the equivalence with simpler approaches promotes our conceptual understanding of the model. I then argue that some real values of mixed-effects modelling are, on the other hand, often overlooked or not recognized. More specifically, when the data are crossed, mixed-effects modelling has significant advantages over other simpler approaches, but for several common types of data in psychology and neuroscience, researchers conventionally do not apply mixed-effects modelling with crossed random effects, jeopardizing the accuracy of statistical inference.