Phillip was a struggling mathematician, then a linguist and now a neuroscientist, but always a hacker.
Simulation is the gold standard for power analysis of mixed-effects models. Traditionally, realistic simulations have been both hard to set up and computationally intractable. Here we present a Julia-based workflow for fast and easy power analysis of mixed-effects models. With MixedModelsSim.jl, we can easily simulate realistic datasets, which we can then quickly analyze with MixedModels.jl. Combined with Pluto.jl, users can explore the impact on power of design decisions interactively.
Permutation-based testing has become exceptionally popular in the analysis of neuroimaging data, but so far has depended on two-stage analyses to handle repeated measures data in order to be computationally tractable. We introduce MixedModelsPermutations to perform fast permutation-based inference on mixed models. Using mixed models instead of a two-stage approach enables properly representing crossed and nested designs, explicitly modelling multiple sources of variability and unbalanced designs
Regression models are useful but they can be tricky to interpret.
Variable centering and contrast coding can obscure the meaning of main effects.
Interaction terms, especially higher order ones, only increase the difficulty of interpretation.
Here, we introduce Effects.jl which translates the fitted model, including estimated uncertainty, back into data space.
Using Effects.jl, it is possible to generate effects plots that enable rapid visualization and interpretation of regression models.