🏆 Individual: @simine.com, psychologist at @unimelb.bsky.social & editor-in-chief of Psychological Science, is recognized for pioneering methodological rigor, reproducibility & collaborative research, driving initiatives such as @improvingpsych.org & the journal Collabra @ucpress.bsky.social. (2/5)

In mathematical psychology that happens on a regular basis; the work by Myung & Pitt is what drew me to Bayesian model comparisons.

I would also say that this argument holds for nested models only. Bayes factors allow the comparison of predictive performance for any two models.

A standard case where it's ambiguous is when the parameter space is discrete. How many different species of penguin are on island X?

I think it starts with the research question. If it calls for a BF answer, then a BF one ought to report. And priors on the nuisance parameters hardly matter, esp. if they are orthogonal.

Completely agree. Em-dashes (and semicolons) are very useful.

If the theoretical maximum is relatively low, that is also not a good thing