Here’s an inspiring presentation by Lisa DeBruine, featuring the “faux” package for data simulation for design of studies.
I really really really like simulated-data experimentation, and I can’t stop talking about it. (More here.)
Here’s an inspiring presentation by Lisa DeBruine, featuring the “faux” package for data simulation for design of studies.
I really really really like simulated-data experimentation, and I can’t stop talking about it. (More here.)
Nice! Cool package. However, I think one really helpful thing (maybe the main thing) about simulation for me has been the understanding about the problem and potential model(s) that I gain from having to program the simulation myself. (This is also mentioned in the slides). So, while it could be convenient to have a package if one knows what is really going on behind the package functions, I think it is actually best to have to code up the simulation from scratch for oneself.
For the modern era, I wonder if using LLMs to simulate interview and survey responses might be similarly helpful.