Is there a reason that risk difference isn't used more for time to event outcomes? For example, estimate baseline hazard and hazard at 5 years, then take the difference (i.e., using a Royston-Parmar model) #CausalSky #StatsSky

Draw a DAG to uncover your assumptions from the trenchcoat.

A very nice to initiative where you can post your DAG: opencausal.org And because they're machine readable, they're much easier to search.

I rounded up a few Claude Skills for #RStats users. Huge thanks to the creators who developed them. They share Skills for everything from tidyverse code to brand.yml files to learning while using AI. Hope the list is useful, and please let me know what I missed! 🧡 rworks.dev/posts/claude...

No amount of statistical gymnastics will save data that doesn't support what you're trying to do

This is great. I've given an example in class showing how priors can make some unidentified problems identifiable: a+b=6, solve for b. Prior: a=0~4

The more I learn about stats, the more I use these three things: 1/ Plots - a picture is worth a thousand words 2/ Probability can almost always guide you 3/ Simulation - when it doubt, simulate it out

Average treatment effect in the overlap can be a tricky causal estimand. Why? The ATO is a little different than other estimands. Often, it's not well defined before the analysis. This is because there are many ways to define the population. Instead, it's based on the statistical method. 1/2

The third installment of the “how should we actually construct our causal graphs anyway” series is out now! 👇🏼 Nick & I ask the question: can we just get an LLM to tell us what belongs on the graph?

A few papers I think worth reading. Mostly open access. Causal inference is hard: www.nature.com/articles/s41...

The more obscure a statistical analysis method, the more I question the design. Not saying it's wrong, but I'd have questions why a more "common" approach wasn't used.

Bootstrapping is sort of a semi-Bayesian approach when you think about it

Calling bullshit - a skill that every applied statistician should master. Unfortunately many of the younger statisticians I’ve worked with sometimes lack the bravery to do so. The book looks like a must-have. #Statistics #StatsSky @carlbergstrom.com @carlzimmer.bsky.social

A common critique of Bayesian methods is that priors are arbitrary. I think that's a good thing. It's an assumption, like much of science. Better to be explicit about assumptions (i.e., DAGs, priors, etc) than implicit

It can be tempting to think of propensity scores as a prediction problem. This is problematic. Why? In prediction models, any variable that helps can be included. In causal inference, this can cause bias, e.g., collider bias. Instead, use a directed acyclic graph (DAG) for variable selection.

Percentages > 100%...

Fantastic initiative! Especially useful for papers using simulations

Choosing a causal estimand is important. Why? To make sure the research question is answered! Certain methods can only estimate specific estimands. This is important when comparing methods. Let's use an example. Imagine we want to compare two methods: 1/3

The best way to improve your analysis: Plot your data

Too accurate

This is a good example of how Bayes & Frequentist methods are different paradigms of stats. Not unlike calculus vs linear algebra. Both useful, but mixing them is problematic.

Nominal coverage helped me with confidence intervals: If you repeat an analysis 1,000 times, nominal coverage is the % of intervals that capture the true effect. For 95% CIs, we'd expect ~950/1,000 to include the true value. It's a long-run frequency idea, not a guarantee for any single interval!

Censoring can result in selection bias for survival data. A solution is to use inverse probability of censoring weights. Why? IPCW creates a pseudo-population where censoring is independent of certain covariates! This is similar to how inverse probability of treatment weighting works. 1/n

Finally, another real advantage of causal inference is that you will better recognise situations where you cannot actually estimate the causal effect you are interested in, in any reliable way. So it is a good way to avoid doing things that are hopeless

#Statistics and #RStats thought of the day: Before analyzing data or doing multiple imputation, explore missingness extents and patterns in your data: hbiostat.org/rflow/missing - see especially the interactive NA Combinations tab. #StatsSky

Statistical Rethinking is such a fantastic name. Really has changed the way I think (at least so far, ~80% of the way through the book)

The {marginaleffects} package is an EXTREMELY useful tool for R & Python users that revolutionises the interpretation of models with complicated non-linear features. Be sure to check out this introduction from @vincentab.bsky.social, @noahgreifer.bsky.social. and @andrew.heiss.phd! #EpiSky

The more experience I get, the more I appreciate George Box's quote "All models are wrong, but some are useful"

Causal inference, DAGs especially, force you to be explicit about the assumptions you're making. Learning about Bayesian statistics, feels like it's similar

Does #randomization ensures balance of risk factors between groups? Consider this: In Denmark 860 individuals were randomly allocated to either intervention or control. Individuals were unaware of their allocation. No intervention took place. Mortality was higher in the intervention group (p=0.003)

My advice (for what little it's worth): pick a topic that interests you, pick statistics you can practice, and apply it there. Read widely and be judicious in who you follow and take advice from on social media. Importantly: have fun while doing it. If you aren't having fun, you won't pursue it

Simulation has now become my default to answer questions. Any stats question I'm curious about? Or situation I don't know about? Head straight to the sim-mobile

The purrr package is phenomenal. Have been using map_* for a while but only recently found out about reduce! Great for combining plots if you have a bunch stored in a list