Marc Ratkovic
LightNews LightNews
marcratkovic.bsky.social
Marc Ratkovic
@marcratkovic.bsky.social
110 followers 290 following 45 posts
Chair of Political Science, Social Data Science
Departments of Political Science and Data Science
University of Mannheim
Political Science, Data, Causal Inference, recently AI and Linguistics
Posts Replies Media Videos
Pinned
marcratkovic.bsky.social
Marc Ratkovic @marcratkovic.bsky.social · 7d
My group has a 3-year position available for a PhD student or post-doc interested in India, ML models/data pipelines/python, and crime and protecting children. Must be eligible to work in Germany. DM or email me for details. Formal calls later, trying here first. #poliscisky #datasciencesky
Reposted by Marc Ratkovic
tommylofstedt.bsky.social
We are #hiring a #phd student to develop novel #Bayesian #MachineLearning methods and algorithms. A fully funded PhD position at @umeauniversitet.bsky.social. See full ad and apply here: www.umu.se/en/work-with...
PhD student in Computing Science with a focus on Machine Learning
www.umu.se
December 19, 2025 at 2:00 PM
Reposted by Marc Ratkovic
thorberger.bsky.social
Everyone talks about the "credibility revolution", but I think one of the most valuable shifts in econ over the last decade has been the rise of rigorous descriptive historical work like this in top journals:
January 18, 2026 at 11:45 AM
Reposted by Marc Ratkovic
dgsoziologie.bsky.social
Professur für Soziologie
Universität Zürich
soziologie.de
January 22, 2026 at 8:45 AM
Reposted by Marc Ratkovic
environmentalpol.bsky.social
New article!

What’s capitalism – generational differentiation in the climate movement by Daphne Fietz.

doi.org/10.1080/09644016.2025.2609418
January 22, 2026 at 9:51 AM
marcratkovic.bsky.social
I keep seeing US, China, Russia discussed as the Great Powers. But Russia is poor and backwards, GDP < Germany nominally, the same or so in PPP, aging weapons systems, it's running out of soldiers. GDP WAY < EU. What puts it on that list? It seems belligerent, sure, but manageable.
January 22, 2026 at 11:41 AM
Reposted by Marc Ratkovic
jamiecummins.bsky.social
Comparing registrations to published papers is essential to research integrity - and almost no one does it routinely because it's slow, messy, and time-demanding.

RegCheck was built to help make this process easier.

Today, we launch RegCheck V2.

🧵

regcheck.app
RegCheck
RegCheck is an AI tool to compare preregistrations with papers instantly.
regcheck.app
January 22, 2026 at 11:05 AM
Reposted by Marc Ratkovic
hos-asa.bsky.social
#OTD 1925 Annie T Randall b (d 16 Aug 2021) Mathematical statistician at NIH Mental Health Biometrics branch. A trailblazer for African American women in federal science agencies, the ASA Annie T. Randall Innovator Award recognizes contributions of early-career statistical innovators.
January 22, 2026 at 11:15 AM
Reposted by Marc Ratkovic
causalab.org
The Methods Series @ki.se is back!

@epiellie.bsky.social kicks off the 2026 speaker lineup next Tuesday, "How do we actually draw a causal graph? A proposal for evidence-based best practices."

📆 Jan 27, 2027
⏰ 15.00 CET/9.00 EST
📍 Online

Register to attend 👇
stats.sender.net/forms/e7JD1d...
January 27, 2026 Methods Talk: How do we actually draw a causal graph? A proposal for evidence-based best practices.
January 21, 2026 at 9:28 PM
marcratkovic.bsky.social
If the manuscript would fail in any undergraduate or graduate seminar, it should not be accepted into NeurIPS. Maybe slow science is the answer? Closer peer review? This is a real issue.
January 21, 2026 at 9:57 PM
Reposted by Marc Ratkovic
adrianbowyer.bsky.social
"In 2024, the total installed electricity capacity of the planet—every coal, gas, hydro, and nuclear plant and all of the renewables—was about 10 terawatts. The Chinese solar supply chain can now pump out 1 terawatt of panels every year."
January 21, 2026 at 6:31 PM
Reposted by Marc Ratkovic
spacemarschall.net
Are you looking for a postdoc position working on asteroids and comets? Would it help if it's in a beautiful place in southern France? If yes, then consider applying to the following position of my colleague Benoit Carry at the Observatory in Nice:
euraxess.ec.europa.eu/jobs/402212

#planetSci
Post-doc on properties of Solar system small bodies
The small bodies of our Solar system (comets and asteroids) are the remnants of the building blocs that accreted to form the planets. The suite of events that occurred during planet formation left pri...
euraxess.ec.europa.eu
January 21, 2026 at 4:38 PM
Reposted by Marc Ratkovic
andrewknightfella.bsky.social
January 21, 2026 at 11:21 AM
Reposted by Marc Ratkovic
stefanmueller.bsky.social
Connected_Politics Lab Seminars, Spring Trimester ⤵️

- 28 Jan: @jessicadicocco.bsky.social
– 11 Feb: Akitaka Matsuo
– 25 Feb: @malojan.bsky.social, @luissattelmayer.bsky.social, and Noémie Piolat
– 4 Mar: @vivifabrien.bsky.social
– 1 Apr: @elisaadamico.bsky.social
– 15 Apr: @miriamsorace.bsky.social
Overview of seminar series, details available at: https://www.ucd.ie/connected_politics/events/
January 21, 2026 at 11:21 AM
Reposted by Marc Ratkovic
pengzell.bsky.social
JOB! I'm hiring a postdoc for 2 years on my ERC MaMo project.

Looking for someone with strong quant methods, ongoing work close to the project's aims, and a desire to publish in sociology. Start flexible in the next 12 months.

Formal call out shortly, but contact me first.
Markets and Mobility: How Employers Structure Economic Opportunity Intergenerational mobility, measuring the ability to achieve economic success regardless of family background, is a critical reflection of a society’s commitment to equality of opportunity. Rising income inequality has raised concerns about the potential erosion of upward mobility. While education has traditionally been viewed as the path to mobility, its transformative power is facing challenges in a rapidly evolving job market. This project reorients the focus of intergenerational mobility research by highlighting the labor market as an arena for the reproduction of advantage. It employs a comparative approach, using administrative data from four countries: Sweden, Austria, England, and the United States. It also incorporates evidence from a broader set of nations through cross-national surveys, longitudinal household surveys, labor force surveys, secondary data, and digital trace data. The project employs cutting-edge empirical methods, including quasi- experimental designs, event studies, within-family comparisons, decomposition analyses, counterfactual simulations, and diagnostic checks to rigorously assess the extent of inequalities in the labor market. The research investigates how family background influences the sorting of individuals to employers and workplaces, accounting for education and occupation, and explores variations in career progression within and between employers. It comprehensively catalogues and assesses mechanisms shaping workplace inequality, contributing to the development of social closure theory. Additionally, the project evaluates intervention strategies, encompassing both employer practices and government actions, to promote fair opportunity in the labor market.
January 21, 2026 at 12:32 PM
Reposted by Marc Ratkovic
epsrjournal.bsky.social
Why did countries back shared EU debt during COVID-19🦠 after rejecting it in the euro crisis?

@tiagomramalho.bsky.social explains how a new “solidarity coalition” formed in 2020 and what it reveals about changing ideas of responsibility in the EU.
The letter and the bond: coalitions, anti-coalitions, and the adoption of the coronabond in the European Union | European Political Science Review | Cambridge Core
The letter and the bond: coalitions, anti-coalitions, and the adoption of the coronabond in the European Union
buff.ly
January 21, 2026 at 3:44 PM
marcratkovic.bsky.social
A nice update of the LaLonde study! Mixed data with an experimental benchmark and observational component. Our causal methods can work, if done properly.
paperposterbot.bsky.social ArXiv Paperboy (Stat.ME+Econ.EM) @paperposterbot.bsky.social · 21h
arXiv📈🤖
Reevaluating Causal Estimation Methods with Data from a Product Release
By Young, Ngatia, Dillon
Recent developments in causal machine learning methods have made it easier to estimate flexible relationships between confounders, treatments and outcomes, making unconfoundedness assumptions in causal analysis more palatable. How successful are these approaches in recovering ground truth baselines? In this paper we analyze a new data sample including an experimental rollout of a new feature at a large technology company and a simultaneous sample of users who endogenously opted into the feature. We find that recovering ground truth causal effects is feasible -- but only with careful modeling choices. Our results build on the observational causal literature beginning with LaLonde (1986), offering best practices for more credible treatment effect estimation in modern, high-dimensional datasets.
January 21, 2026 at 5:46 PM
Reposted by Marc Ratkovic
paperposterbot.bsky.social
arXiv📈🤖
Lost in Aggregation: The Causal Interpretation of the IV Estimand
By Tsao, Muandet, Eberhardt et al
Instrumental variable based estimation of a causal effect has emerged as a standard approach to mitigate confounding bias in the social sciences and epidemiology, where conducting randomized experiments can be too costly or impossible. However, justifying the validity of the instrument often poses a significant challenge. In this work, we highlight a problem generally neglected in arguments for instrumental variable validity: the presence of an ''aggregate treatment variable'', where the treatment (e.g., education, GDP, caloric intake) is composed of finer-grained components that each may have a different effect on the outcome. We show that the causal effect of an aggregate treatment is generally ambiguous, as it depends on how interventions on the aggregate are instantiated at the component level, formalized through the aggregate-constrained component intervention distribution. We then characterize conditions on the interventional distribution and the aggregate setting under which standard instrumental variable estimators identify the aggregate effect. The contrived nature of these conditions implies major limitations on the interpretation of instrumental variable estimates based on aggregate treatments and highlights the need for a broader justificatory base for the exclusion restriction in such settings.
January 21, 2026 at 4:27 PM
Reposted by Marc Ratkovic
paperposterbot.bsky.social
arXiv📈🤖
How Well Do LLMs Predict Human Behavior? A Measure of their Pretrained Knowledge
By Gao, Han, Liang
Large language models (LLMs) are increasingly used to predict human behavior. We propose a measure for evaluating how much knowledge a pretrained LLM brings to such a prediction: its equivalent sample size, defined as the amount of task-specific data needed to match the predictive accuracy of the LLM. We estimate this measure by comparing the prediction error of a fixed LLM in a given domain to that of flexible machine learning models trained on increasing samples of domain-specific data. We further provide a statistical inference procedure by developing a new asymptotic theory for cross-validated prediction error. Finally, we apply this method to the Panel Study of Income Dynamics. We find that LLMs encode considerable predictive information for some economic variables but much less for others, suggesting that their value as substitutes for domain-specific data differs markedly across settings.
January 21, 2026 at 4:36 PM
Reposted by Marc Ratkovic
sociologicalsci.bsky.social
NEW: AJ Alvero, Dustin S. Stoltz, Oscar Stuhler, Marshall A. Taylor, "Generative AI in Sociological Research: State of the Discipline" sociologicalscience.com/articles-v13...
sociologicalscience.com
January 20, 2026 at 6:03 PM
Reposted by Marc Ratkovic
rolandparis.bsky.social
This is one of the sharpest analyses of international affairs that I've heard from a Canadian leader - or any national leader - in a long time. And I suspect he wrote the main bits himself. paulwells.substack.com/p/the-carney...
The Carney doctrine
Open comment thread on the PM's Davos speech
paulwells.substack.com
January 20, 2026 at 5:13 PM
Reposted by Marc Ratkovic
minimaxir.bsky.social
Back in May, I also used the gambling metaphor to describe vibe coding.

Claude Opus 4.5 has caused to me to reevaluate things. minimaxir.com/2025/05/llm-...
Vibe coding with coding agents like Claude Code or Cursor is something I have little desire to even experiment with. On paper, coding agents should be able to address my complaints with LLM-generated code reliability since it inherently double-checks itself and it’s able to incorporate the context of an entire code project. However, I have also heard the horror stories of people spending hundreds of dollars by accident and not get anything that solves their coding problems. There’s a fine line between experimenting with code generation and gambling with code generation. Vibe coding can get me 80% of the way there, and I agree there’s value in that for building quick personal apps that either aren’t ever released publicly, or are released with disclaimers about its “this is released as-is” nature. But it’s unprofessional to use vibe coding as a defense to ship knowingly substandard code for serious projects, and the only code I can stand by is the code I am fully confident in its implementation.
January 20, 2026 at 5:13 AM
marcratkovic.bsky.social
Democracy needs to mean something beyond simply the franchise. As a corollary, democracy does not defend itself. And there is much in it worth fighting for.
anneapplebaum.bsky.social Anne Applebaum @anneapplebaum.bsky.social · 2d
The Prime Minister of Belgium speaks:
January 20, 2026 at 1:56 PM
Reposted by Marc Ratkovic
georgiatomova.bsky.social
If you are interested in all things ✨causal inference✨, please join our multidisciplinary Causal Inference Interest Group (CIIG). We host monthly seminars featuring speakers with various academic backgrounds and research interests.

Links below.

cc @clscohorts.bsky.social @pwgtennant.bsky.social
January 20, 2026 at 12:35 PM
Reposted by Marc Ratkovic
cepeo-ucl.bsky.social
this morning, our director @lindseymacmillan.bsky.social will be part of a panel discussing 'How can policy tackle regional inequalities?' with @andyburnham.bsky.social
@xiaoweixu.bsky.social
@theifs.bsky.social

ifs.org.uk/events/how-c...
How can policy tackle regional inequalities? | Institute for Fiscal Studies
At this policy conference, researchers will present new evidence on the scale and causes of regional inequalities.
ifs.org.uk
January 20, 2026 at 9:02 AM
Reposted by Marc Ratkovic
paperposterbot.bsky.social
arXiv📈🤖
A General Form of Covariate Adjustment in Randomized Clinical Trials
By
arXiv:2306.10213v2 Announce Type: replace Abstract: In randomized clinical trials, adjusting for baseline covariates can improve credibility and efficiency for demonstrating and quantifying treatment effects. This article studies the augmented inverse propensity weighted (AIPW) estimator, which is a general form of covariate adjustment that uses linear, generalized linear, and non-parametric or machine learning models for the conditional mean of the response given covariates. Under covariate-adaptive randomization, we establish general theorems that show a complete picture of the asymptotic normality, {efficiency gain, and applicability of AIPW estimators}. In particular, we provide for the first time a rigorous theoretical justification of using machine learning methods with cross-fitting for dependent data under covariate-adaptive randomization. Based on the general theorems, we offer insights on the conditions for guaranteed efficiency gain and universal applicability {under different randomization schemes}, which also motivate a joint calibration strategy using some constructed covariates after applying AIPW. Our methods are implemented in the R package RobinCar.
January 19, 2026 at 7:06 PM