Michael Schomaker
@mfschomaker.bsky.social
Heisenberg Professor for Biostatistics at the Department of Statistics, LMU München | causal inference - missing data - HIV
michaelschomaker.github.io
michaelschomaker.github.io
3/ Additionally, we quantify positivity violations by looking at "high-density regions" of the conditional treatment density, using a novel diagnostic: the non-overlap ratio.
March 21, 2025 at 6:14 AM
3/ Additionally, we quantify positivity violations by looking at "high-density regions" of the conditional treatment density, using a novel diagnostic: the non-overlap ratio.
2/ The key idea is to stick as closely as possible to the estimand of interest (dose-response curve), deviating from it only when really necessary, i.e. for units (and time points) where conditional support is too low; and only as much as needed.
March 21, 2025 at 6:14 AM
2/ The key idea is to stick as closely as possible to the estimand of interest (dose-response curve), deviating from it only when really necessary, i.e. for units (and time points) where conditional support is too low; and only as much as needed.