Michael Schomaker
LightNews LightNews
banner
mfschomaker.bsky.social
Michael Schomaker
@mfschomaker.bsky.social
85 followers 96 following 13 posts
Heisenberg Professor for Biostatistics at the Department of Statistics, LMU München | causal inference - missing data - HIV
michaelschomaker.github.io
Posts Replies Media Videos
Reposted by Michael Schomaker
pausalz.bsky.social
Going to start referring to myself as a causalologist
November 21, 2025 at 1:07 PM
Reposted by Michael Schomaker
spotlightnsp.bsky.social
DIVE IN | Spotlight has launched a new HIV dashboard and graph generator based on Thembisa model data, making key HIV statistics easier to access for the public and healthcare professionals. Find out more: shorturl.at/GrZyF

Join our mailing list: shorturl.at/Ntzoa
November 14, 2025 at 7:42 AM
Reposted by Michael Schomaker
dagophile.bsky.social
This week's episode is on the Manipulation Theorem, which is not well-known in the broader causal community, but has hugely influenced philosophers. I cover hard vs soft interventions, why causation doesn't require human agents, and whether causation requires open systems.

youtu.be/fg1ApidilVM
E9: Using Causal Graphs to Change the World
YouTube video by Causal Foundations
youtu.be
November 14, 2025 at 5:57 PM
Reposted by Michael Schomaker
juhakarvanen.bsky.social
In longitudinal studies, dropout leads to a monotone missing data pattern. We show in a new article that monotonicity sometimes enables and sometimes prevents the identification of  the full law, i.e., the joint distribution of actual variables and response indicators. openreview.net/pdf?id=kVthd...
Two graphs where monotonicity prevents identification
November 13, 2025 at 11:39 AM
mfschomaker.bsky.social
Happening at 4.15pm at our department today: @herbps10.bsky.social and @idiaz.bsky.social presenting on strategies to overcome positivity violations for complex, possibly longitudinal exposures 👇
caslmu.bsky.social Center for Advanced Studies @caslmu.bsky.social · 11d
We warmly welcome Herb Susmann @herbps10.bsky.social (@nyu.edu Grossman School of Medicine), as a #CASFellow. Michael Schomaker @mfschomaker.bsky.social invited him to come to Munich in the context of the #CASResearchFocus “Medical Data Processing”.
November 12, 2025 at 1:01 PM
Reposted by Michael Schomaker
caslmu.bsky.social
🗓 Today, at 4:15 pm he gives a joint talk with Iván Díaz @idiaz.bsky.social (NYU Grossman School of Medicine) on Causal Inference Based on Machine Learning for Complex Longitudinal Exposures at the Institute for Statistics @lmumuenchen.bsky.social

www.cas.lmu.de/de/veranstal...
Causal Inference Based on Machine Learning for Complex Longitudinal Exposures
Referenten: Prof. Iván Díaz, Ph.D. and Herbert Susmann, Ph.D. (CAS Fellows/NYU) | Moderation: Prof. Dr. Michael Schomaker (CAS Young Center/LMU)
www.cas.lmu.de
November 12, 2025 at 12:25 PM
mfschomaker.bsky.social
How should we deal with missing data when using TMLE? Find out in our new working paper, led by Christoph Wiederkehr:
arxiv.org/pdf/2510.22202

Pro Tip: Check out Box 1 for a summary! (...more surprising than you may think...)
arxiv.org
October 29, 2025 at 1:48 PM
Reposted by Michael Schomaker
statsepi.bsky.social
My own TLDR for the message from this paper:

statsepi.substack.com/p/sorry-what...
You aren't allowed to say “risk factor” People often say their goal is to identify “risk factors”. But what does that mean? Some people use the term to indicate potential causes of outcomes. Then just say cause. Others use it to identify predictors of outcomes. Then just say predict. And, sadly, too many others use it as shorthand for factors that are “statistically associated” with outcomes. In this case, say nothing at all, since this “goal” has no clinical utility whatsoever (beyond what it might suggest about causation or prediction). So once you have framed your research question as description, prediction, causation or measurement, there is no longer a need to talk about risk factors. It's basically just a catch-all phrase to cover up muddy thinking.
October 28, 2025 at 11:34 AM
Reposted by Michael Schomaker
munichcenterml.bsky.social
𝗖𝗮𝗹𝗹 𝗳𝗼𝗿 𝗳𝘂𝗹𝗹𝘆 𝗳𝘂𝗻𝗱𝗲𝗱 𝗣𝗵𝗗 𝗣𝗼𝘀𝗶𝘁𝗶𝗼𝗻𝘀: We are offering several PhD positions across our various research areas, open to highly qualified candidates.
‼️ The application portal will be open from 15 October to 14 November 2025.

Find out more: mcml.ai/opportunitie...
October 10, 2025 at 6:57 AM
Reposted by Michael Schomaker
mcknaus.bsky.social
🚨Job alert🚨

Premium tenure-track (ass to full prof) in "𝐌𝐋 𝐌𝐞𝐭𝐡𝐨𝐝𝐬 𝐢𝐧 𝐁𝐮𝐬𝐢𝐧𝐞𝐬𝐬 𝐚𝐧𝐝 𝐄𝐜𝐨𝐧𝐨𝐦𝐢𝐜𝐬" @unituebingen.bsky.social made possible by @ml4science.bsky.social

Spread the word and feel free to reach out if you have any questions about the position or the environment.

Link: shorturl.at/QHZ5G
September 12, 2025 at 1:25 PM
Reposted by Michael Schomaker
margaritamb.bsky.social
NEW PAPER! (in press in EPIDEMIOLOGY)

Have you wondered:
- How to specify a target trial: as an ideal trial or something else?
- What biases do target trial emulations and actual RCTs share?
- How does it all relate to potential outcomes?

Read here! 👉 arxiv.org/abs/2405.10026
#EpiSky #CausalSky
The ideal trial: defining causal estimands that balance relevance and feasibility in target trial emulations and actual randomized trials
Causal inference is the goal of randomized trials and many observational studies. The first step in a formal causal inference framework is to define the causal estimand, and in both types of study thi...
arxiv.org
September 9, 2025 at 11:23 PM
Reposted by Michael Schomaker
causalab.bsky.social
The CAUSALab Methods Series @ki.se is back!

Fall 2025 lineup kicks off Sep 23 with Vanessa Didelez (BIPS), “Causal mediation and separable treatments in time-to-event analyses.”

All talks are virtual, except for Nov. 4, 2025 hybrid session.

Learn more & register:
hsph.harvard.edu/research/cau...
https://hsph.harvard.edu/research/causalab/seminars/
September 9, 2025 at 2:44 PM
Reposted by Michael Schomaker
ehwoza.bsky.social
[TRAILER]🎥 Addiction is a film created by high school learners from @IkamvaYouthSA Masiphumelele, Fish Hoek. It tells the stories of young people who have struggled with substance abuse.

Film by: Anovuyo Gazi, Mamoya Mohapi and Mmamei Mofokeng

Full video: youtube.com/watch?v=ob47...
Addiction | A Journey of Recovery and Hope
YouTube video by Eh!woza
youtube.com
August 29, 2025 at 10:02 AM
Reposted by Michael Schomaker
idiaz.bsky.social
I think missing data is a CI problem (counterfactual is "would have observed the data") but not the opposite. E.g., recasting mediation analyses etc as a missing data problems seems contrived.
July 31, 2025 at 3:12 PM
Reposted by Michael Schomaker
ehwoza.bsky.social
[TRAILER] Beyond The Fear unfolds through the voices of women like Zimkitha, her mother, and local resident Noluvo, each navigating a landscape shaped by fear, grief, and survival.

Part of the Learner Doccies pilot in Xhora Mouth with Bulungula Incubator.
www.youtube.com/watch?v=JnIQ...
Beyond the Fear | Living Through the Shadow of Violence
YouTube video by Eh!woza
www.youtube.com
July 25, 2025 at 11:25 AM
Reposted by Michael Schomaker
jurodemann.bsky.social
🎉 Honored to receive the IJAR Young Researcher Award (🥈)

The award is given to researchers who demonstrate excellence at an early stage of their scientific careers, sponsored by the International Journal of Approximate Reasoning (IJAR).

isipta25.sipta.org/awards
July 24, 2025 at 7:21 AM
Reposted by Michael Schomaker
rmcelreath.bsky.social
@xkcd.com again provides for your meta-science slides (link xkcd.com/3117). I do wish we'd stop talking about a "crisis", bc it's been here for generations. Ppl used to pipette by mouth ffs. Enrico Fermi won a Nobel prize for a false result. But that doesn't mean we can't do better going forward.
xkcd comic about replication crisis https://xkcd.com/3117
July 20, 2025 at 5:27 AM
Reposted by Michael Schomaker
mjtsmith.bsky.social
🚨 New paper in Statistics in Medicine!
We compare TMLE vs CV-TMLE in tricky simulations:
✅ Misspecification
✅ Limited overlap
✅ High dimensions
✅ Super Learner + cross-validation

Who performs best?
Read it: onlinelibrary.wiley.com/doi/10.1002/...
#CausalInference #TMLE #CVTMLE
Performance of Cross‐Validated Targeted Maximum Likelihood Estimation
Background Advanced methods for causal inference, such as targeted maximum likelihood estimation (TMLE), require specific convergence rates and the Donsker class condition for valid statistical esti.....
onlinelibrary.wiley.com
July 18, 2025 at 9:31 PM
Reposted by Michael Schomaker
spotlightnsp.bsky.social
MUST READ | South Africa is staging a sequel to Mbeki-era denialism, only this time, the science, solutions, and costs are clearer, argues Professor @francoisventer.bsky.social‬. @groundup.org.za

www.spotlightnsp.co.za/2025/07/16/f...
Francois Venter | Slow motion denialism: Our leaders are allowing the HIV response to collapse
South Africa is staging a sequel to Mbeki-era denialism, only this time, the science, solutions, and costs are clearer, argues Professor Francois Venter.
www.spotlightnsp.co.za
July 16, 2025 at 5:27 AM
Reposted by Michael Schomaker
margaritamb.bsky.social
1/ NEW R PACKAGE! For estimating the impact of potential interventions on multiple mediators in countering exposure effects (led by @cttc101.bsky.social)

- Paper👉 tinyurl.com/ye26jsps
- Package👉 tinyurl.com/yuh4kens

Thread shows published examples of how the method can be used! #EpiSky #CausalSky
tinyurl.com
July 10, 2025 at 1:30 AM
Reposted by Michael Schomaker
margaritamb.bsky.social
NOW PUBLISHED! Featured article + 6 Commentaries + Rejoinder: “On the Uses and Abuses of Regression Models: A Call for Reform of Statistical Practice and Teaching” with my colleague John Carlin
tinyurl.com/2z2tmkhh
@cebu-melbourne.bsky.social @vicbiostat.bsky.social #EpiSky #CausalSky #StatsSky
Rejoinder to Commentaries on: On the Uses and Abuses of Regression Models: A Call for Reform of Statistical Practice and Teaching
Click on the article title to read more.
onlinelibrary.wiley.com
July 2, 2025 at 1:56 AM
Reposted by Michael Schomaker
spotlightnsp.bsky.social
THE WAIT BEGINS | Six months of HIV protection in a single shot. When will it arrive in SA and at what cost? Find out: shorturl.at/AphjB

Join our mailing list: shorturl.at/Ntzoa
June 20, 2025 at 11:18 AM
Reposted by Michael Schomaker
arikreuter.bsky.social
Why is it different?

Do-PFN is a radical new approach to causal inference, replacing standard assumptions of a ground-truth causal model (Pearl) or structural assumptions (Rubin) with a prior over SCMs—our modeling assumptions lie in our synthetic data-generating process. [6/8]
June 10, 2025 at 9:33 AM
Reposted by Michael Schomaker
rmcelreath.bsky.social
also from the meme archives
June 4, 2025 at 7:38 AM
Reposted by Michael Schomaker
rmcelreath.bsky.social
I feel like all the debates in the 20th century were for nothing sometimes. People used to argue that Bayes couldn't work. The James-Stein estimator was a big deal! Everything uses gradients now and is inherently about regularization. Let the 20th century go, I say.
June 4, 2025 at 6:32 AM