Conversation in #SpeedLab:
Student: I programmed an LLM with all of your publicly-available papers, and now I ask it for advice instead of bothering you.
Me: ...That's like a child building a robot to give out hugs because their parents are absent.
@lakens Effect sizes for group-by-group comparisons in regression/ANCOVA models. If the central premise of a study is: Does Group A and Group B differ? The effect size emphasized shouldn't be R2 (b/c uneven group size will impact it), authors should report Standardized Mean Differences.