The underlying issue is the non-collapsibility of ORs and HRs. Non-collapsibility means that the conditional ratio is different from the marginal (unadjusted) ratio even in the complete absence of confounding (as in our example dataset below). By the way, don’t make the mistake of concluding that non-collapsibility is undesirable. Any measure that has the potential for summarizing a treatment effect with one constant for all types of patients will be non-collapsible when the outcome is categorical or represents time to event1. Collapsible measures such as absolute risk reduction and relative risk reduction must vary over risk factors (creating mathematical but not subject-matter-relevant interactions), otherwise probabilities will arise that are outside the allowable range of [0,1]. Log odds and log hazard ratios have an unlimited ranges and can possibly apply to everyone. This makes them good bases for studying heterogeneity of treatment effect. — Unadjusted Odds Ratios are Conditional.

Ubuntu VM on macOS with libvirt + QEMU.

Evening: PJ Harvey.

Invent with Python. #python

Contravariant functors are Weird. #haskell

We all know that a languages like Ruby or Python are designed explicitly to hide this sort of complexity from us and let us get on with the dirty business of munging data blobs or serving web requests or solving sudokus or whatever, and thank goodness for that, but wow that is quite a lot, isn’t it? — Hello “Hello world!”

First, new features just keep being added, because how else would you justify releases? VS Code releases every month with new stuff! They are well past the point of shipping the essentials (and have been for a couple of years now), but releases are still been shipped. Because they have a team, that team has a huge budget to spare, and there’s no power stopping them from shipping anything, no matter important or not. There’s no filter. — The most important feature of Sublime Text

Not sure what data-oriented programming is but I bookmarked two posts the other day and I thought I could share them:

• Data-oriented Programming in Python
• An introduction to Data Oriented Design with Rust

The many Flavours of Missing Values. The eternal problem with NA values… #rstats

Introspective Emacs. #emacs