A comparison of Bayesian vs frequentist MMRMs

[wlandau]

I lead the Bayesian MMRM sub-team in openstatsware, and we are working the brms.mmrm package. We have three package vignettes to support validation: a simulation-based calibration workflow, a comparison with the mmrm package on a small dataset, and a comparison with the mmrm package on a larger dataset. The last two feature the same datasets that mmrm uses to compare itself with SAS.

The sub-team is interested in contributing to CAMIS. However, in view of the current table at https://psiaims.github.io/CAMIS/#repository, our situation seems unusual. Our method is Bayesian and we are comparing against the mmrm R package instead of SAS (since mmrm and SAS already agree so closely). Is CAMIS the right place to share our method and validation materials?

[DrLynTaylor]

Hi @wlandau, Thanks so much for posting the above. @andreaczhang has reached out to you as we'd love to collaborate.