In the real world, many effects are non-linear. This new paper by @urisohn.bsky.social offers an intuitive explanation of the challenges with standard methods of testing interactions, along with a revised toolbox. In addition to changing a few lines in my R code, this changes how I think about data.
Reposted from
PsyArXivBot
Interacting With Curves: How to Validly Test and Probe Interactions in the Real (Non-linear) World: http://osf.io/xnzc3/
Comments