Computing analytic Bayes factors from summary statistics in repeated-measures designs

Summary Bayes factors are an increasingly popular tool for indexing evidence from experiments. For two competing population models, the factor reflects relative likelihood of observing some data under one model compared to other. Computing can be difficult, requiring integrate product and a prior distribution on parameter(s) both models. Previous work has obviated this difficulty independent-groups designs. In paper, we develop new analytic formula computing directly minimal summary statistics in repeated-measures This is improvement previous methods (e.g., BIC method), which produce that violate Sellke upper bound smaller sample sizes. The approach taken paper requires knowing only F -statistic degrees freedom, commonly reported most empirical work. addition providing computational examples, report simulation study benchmarks against other Our method provides easy way researchers compute set statistics, allowing users index evidential value their own data, as well published studies.