Malcolm Barrett
Malcolm Barrett
probably less sample size for when you marginalize over it but you want to be careful of the curse of dimensionality even when not estimating the interaction term for sake...
That code should use `inherits()`, either way
This sounds cool, but I don't like the name. Variance is just one aspect of selecting an adjustment set, and I think a user seeing "optimal" will think it's doing...
Yes, I reckon that's the idea. But that's true only if the data are perfect. Real life adjustment sets offer different levels of bias reduction despite being theoretically equivalent. Not...
I have no problem with that idea. In fact, I like it. But again, that assumes that even the finite data are perfect. In real life, adjustment sets are not...
This is probably the section where we show target trial design in practice: filter exclusions, transform or select exposure that is specific to the problem, etc
Related to https://github.com/malcolmbarrett/causal_inference_r_workshop/issues/19 where we need data with loss to follow up
Could this type of approach be useful here? https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3935334/
Some data sources: https://github.com/hfrick/cetaceans (has unknown category, so could be case for loss to followup) `censored::time_to_million` `parmsurvfit::oscars` (this package has some other datasets, too)
Time to divorce: https://grodri.github.io/survival/project. number of kids as time-varying exposure? SurvSet (collection of time to event data): https://arxiv.org/pdf/2203.03094.pdf, https://github.com/ErikinBC/SurvSet/tree/main/SurvSet/_datagen/output