[CogSci] Call for papers: Special issue of Computational Brain and Behavior, Bayesian inference for hierarchical models

[Scott Brown]

Recent years have seen rapid adoption of mixed models (also known as "multi-level models" or "hierarchical models") in psychological science. These approaches have many strengths, particularly the ability to coherently address the ever-present differences between individuals and items in typical psychological experiments. 

When mixed models are addressed in a Bayesian framework, it is not clear which is the most appropriate Bayes factor hypothesis test to quantify support the presence or absence of experimental effects. Different choices for both the null model and the alternative model are possible, and each choice constitutes a different definition of an effect resulting in a different test outcome. 

The target article, by van Doorn, Aust, Haaf, Stefan, and Wagenmakers, addresses this problem. They outline the common approaches and illustrate the impacts on hypothesis tests from aggregation of data, measurement error, choice of prior distribution, and the detection of interactions. You can read a pre-print of the target article here: https://psyarxiv.com/y65h8  (note that it is still working its way through the publication process, so may change in some ways before final publication). The target article also includes simulation scripts and other analytic tools, which are also all available online (URL in the article). 

We will use the pages of Computational Brain and Behavior (https://www.springer.com/journal/42113) to help advance this issue. We propose publishing the target article in a special issue, alongside different approaches and perspectives on the problem, to more carefully contrast the strengths and weakness of each approach. Authors of the other articles in the special issue, should illustrate their alternative perspectives in a detailed and concrete manner. The minimal requirements for this are: (1) the use of Bayes factors to test hypotheses in a mixed model; and (2) directly addressing the questions raised in the target article using the same data (available online from OSF) or simulation environment. 

If you are interested in submitting an article on this topic, or have an idea for whom you would like to see a contribution from, please let me know. If you plan submit a manuscript for consideration in this special issue, please let me know that, also. If you are unsure about the scope of the issue, or have other questions, don't be afraid to ask! Finally, feel free to distribute this email further.

Thanks for your support,

Professor Scott Brown
Editor, Computational Brain and Behavior.