iiinova — where your digital self keeps evolving

Reposted by @davidbroska.bsky.social

Julian Berger @officialberger.bsky.social · 22/09/2026

New paper out in @pnas.org 🚨 Human learning is an understudied but promising lever for boosting human–AI synergy @jasonburton.bsky.social @ralfkurvers.bsky.social @stefanherzog.bsky.social @dirkwulff.bsky.social

12213

davidbroska.bsky.social @davidbroska.bsky.social · 31/08/2026

Americans favor predistributive policies over redistributive policies. Our new paper is now out in PNAS 🚨 Paper: www.pnas.org/doi/10.1073/... We have also written a summary of the article for Kudos, a brief report on our brief report: link.growkudos.com/1eb04zt75ds

Americans favor predistributive over redistributive policies

175

Reposted by @davidbroska.bsky.social

NEP-SPO: Sports and Economics @repec-nep-spo.bsky.social · 13/07/2026

A many-designs study of 516 algorithms challenges the predictability of wisdom of the crowd forecast aggregation: WoCCAP consortium

d.repec.org

NEP/RePEc link

to paper

001

Reposted by @davidbroska.bsky.social

RigorLabs @rigorlabs.bsky.social · 29/06/2026

AUTOMATING COMPUTATIONAL REPRODUCIBILITY My colleague @philipjakobbln.bsky.social and I are currently engaged in a research project where we reproduce scientific results en masse. To that end, we built rigor.me, a platform for automatically reproducing papers using agents 🧵

1177

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Thank you for helpful feedback from @garritzmannj.bsky.social, @timo-sprang.net, and folks at @pascl_stanford! We build on important work by @ikuziemko.bsky.social, @nicolaslonguetmarx.bsky.social, Suresh Naidu, and Jacob Hacker.

130

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Preprint: osf.io/preprints/so... Preregistration, data, materials, and code on OSF: osf.io/kndq7 With @jonnekamphorst.bsky.social and @robbwiller.bsky.social!

130

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Bottom line: Public support for inequality-reducing policies depends not only on what the policies do materially, but also on whether they are seen as raising earnings and respecting work or as government transfers of resources across groups.

1122

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Caveats: this is contemporary U.S. opinion across six tested policies (EITC, Federal Minimum Wage increase, Progressive Income Taxation, PRO Act, Medicare For All Who Want It, America's College Promise Act). Other policies, wordings, or contexts could shift the pattern. Non-causal mediation results.

140

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

The control condition clarified what was driving the gap. Predistributive descriptions received support similar to the baseline, but redistributive descriptions received less support than baseline. Policies don’t receive a predistribution boost in support, but a redistribution penalty.

140

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Why? When the same policies were described in predistributive terms, people saw them as more respectful of hard work, more beneficial to deserving recipients, fairer, more beneficial to society, and less burdensome to taxpayers.

190

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

We find that policies described as predistributive received more support than policies described as redistributive, especially among Republicans.

130

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Study 2 randomly varied whether a policy was described in predistributive terms (raises earnings for lower-income workers), redistributive terms (shifts resources from higher- to lower-income individuals), or with a baseline description.

130

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Study 1: We pulled 31 nationally representative U.S. survey waves fielded between 2015 and 2024, classifying 83 fieldings of policy support questions as pre- or redistributive. Americans backed predistributive policies more. The gap: 14 points on a 0–100 support scale (0.34 SDs).

250

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Predistribution: government policies that reduce inequality before taxes and transfers by raising earnings for lower-income people (e.g., minimum-wage laws). Redistribution: government policies that reduce inequality after market outcomes, through taxes and transfers (e.g., safety-net benefits).

161

davidbroska.bsky.social @davidbroska.bsky.social · 28/05/2026

Americans say they want less inequality. Why don’t they support policies to reduce it? New paper: across 31 nationally representative survey waves (N=384,248) and a preregistered experiment (N=1,009), Americans favor PREdistributive over REdistributive policies 🧵 osf.io/preprints/so...

87228

davidbroska.bsky.social @davidbroska.bsky.social · 21/05/2026

8/ The control condition clarified what was driving the gap. Predistributive descriptions received support similar to the baseline, but redistributive descriptions received less support than baseline. Policies don’t receive a predistribution boost in support, but a redistribution penalty.

020

davidbroska.bsky.social @davidbroska.bsky.social · 21/05/2026

7/ Why? When the same policies were described in predistributive terms, people saw them as more respectful of hard work, more beneficial to deserving recipients, fairer, more beneficial to society, and less burdensome to taxpayers.

120

davidbroska.bsky.social @davidbroska.bsky.social · 21/05/2026

6/ Predistributive frames got higher support than redistributive ones, especially among Republicans.

120

davidbroska.bsky.social @davidbroska.bsky.social · 21/05/2026

5/ Study 2 randomly varied whether a policy was described in predistributive terms (raises earnings for lower-income workers), redistributive terms (shifts resources from higher- to lower-income individuals), or with a baseline description.

120

davidbroska.bsky.social @davidbroska.bsky.social · 21/05/2026

4/ But Study 1 compared different policies. The gap could reflect differences other than pre- vs redistribution. So we ran an experiment next.

120

Reposted by @davidbroska.bsky.social

Moritz Marbach @moritz-marbach.com · 15/04/2026

🚨New Paper in PNAS: "Refugee Labor Market Integration at Scale: Evidence from Germany’s Fast-Track Employment Program" www.pnas.org/doi/10.1073/... Ungated preprint osf.io/preprints/socarxiv/px9ew_v3 w/ J Hainmueller, D Hangartner, @niklas-harder.bsky.social & E Vallizadeh #econtwitter #econsky

pnas.org

Refugee labor market integration at scale: Evidence from Germany’s fast-track employment program | PNAS

Governments face persistent challenges in integrating refugees into the local labor market, and many past interventions have shown limited impact. ...

17732

davidbroska.bsky.social @davidbroska.bsky.social · 19/01/2026

First day of my research visit at Oxford! Many thanks to @zparolin.bsky.social for the generous invitation. I’ll be presenting on simulating human behavior with LLMs on Jan 28 at the @nuffieldcollege.bsky.social Sociology Seminar. If you’ll be around Oxford, feel free to reach out!

052

davidbroska.bsky.social @davidbroska.bsky.social · 19/12/2025

4/ Great to learn from your work @tsrauf.bsky.social, Jan Voelkel, Jamie Druckman, and @jeremyfreese.bsky.social !

120

davidbroska.bsky.social @davidbroska.bsky.social · 19/12/2025

3/ One aspect of this study I particularly liked was the data quality of the sample: selection did not depend on research outcomes, the samples were nationally representative, and careful vetting of experimental designs suggests that insignificant results are not attributable to quality issues

110

davidbroska.bsky.social @davidbroska.bsky.social · 19/12/2025

2/ Beyond sample size, researchers could reduce residual variance (and thus increase power) by controlling for pretreatment variables that are strongly correlated with the outcome (e.g., government trust and policy attitudes). Future research could assess the efficacy of this strategy systematically

100

davidbroska.bsky.social @davidbroska.bsky.social · 19/12/2025

Effect sizes in social science survey experiments are typically small, requiring large samples and budgets for sufficient statistical power. It gets even trickier because increasing n/$ yields only diminishing returns to statistical power. This study points to an important barrier to credibility👇 1/

141

davidbroska.bsky.social @davidbroska.bsky.social · 18/12/2025

We learned from and built on the terrific work of scholars who thought deeply about how to leverage LLMs for the social and behavioral sciences. Feedback welcome!

000

davidbroska.bsky.social @davidbroska.bsky.social · 18/12/2025

To simulate-then-validate, to validate-then-simulate, or to calibrate: that is the question. We discuss ways to simulate human responses to behavioral science experiments using LLMs and strategies to address their limitations. Check out our preprint!

120

Reposted by @davidbroska.bsky.social

Julian Garritzmann @garritzmannj.bsky.social · 23/09/2025

I'm super happy & proud to be able to work with this rock star group of a team (and academic guests)! 🧠🧠🧠💫🙏 @goetheuni.bsky.social @infer-frankfurt.bsky.social - and it's really nice to check out parts of Frankfurt that I hadn't been to as part of our team event

192

davidbroska.bsky.social @davidbroska.bsky.social · 02/09/2025

Thank you for linking us under this thread, Shira! I enjoyed reading the discussion. It's worth considering in more detail how Prediction-Powered Inference borrows from older ideas, such as the GREG estimate--and how PPI differs from, for instance, Bootstrap sampling that uses just one data source.

020

davidbroska.bsky.social @davidbroska.bsky.social · 30/07/2025

Check out this article on leveraging AI for conducting social science research!

010

davidbroska.bsky.social @davidbroska.bsky.social · 21/07/2025

Excited to continue learning about the latest #CSS at @ic2s2.bsky.social! I’ll be at the Social Prediction Session, presenting the mixed subjects design on combining human and LLM data in experiments. Paper with Michael Howes and @austin-van-loon.bsky.social. Come join us! doi.org/10.1177/0049...

doi.org

Sage Journals: Discover world-class research

Subscription and open access journals from Sage, the world's leading independent academic publisher.

041

Reposted by @davidbroska.bsky.social

Stanford Sociology @stanfordsoc.bsky.social · 25/06/2025

New in Sociological Methods & Research: Soc PhD candidate @davidbroska.bsky.social, @austin-van-loon.bsky.social, & Michael Howes show how combining human subjects and large language models can yield precise estimates at low cost, with implications for scientific productivity doi.org/10.1177/0049...

022

davidbroska.bsky.social @davidbroska.bsky.social · 17/05/2025

How can we leverage generative AI to advance social science methods and research? Daniel Karell and Thomas Davidson led a special issue in Sociological Methods & Research to find out. Special kudos to them! journals.sagepub.com/doi/10.1177/...

000

davidbroska.bsky.social @davidbroska.bsky.social · 23/04/2025

Mixed feelings about silicon subjects (LLM predictions of human behavior) as replacements for human subjects? Consider the mixed subjects design. 🚨Now published at Sociological Methods and Research🚨 doi.org/10.1177/0049...

041

davidbroska.bsky.social @davidbroska.bsky.social · 03/04/2025

Congratulations, Chagai! Wishing you the best—though we’ll miss the causal inference powerhouse at @pascl-stanford.bsky.social. Good luck!

130

Reposted by @davidbroska.bsky.social

Oscar Stuhler @oms279.bsky.social · 31/03/2025

Happy to share my new paper with Cat Dang Ton and @eollion.bsky.social on how to use generative LLMs for extracting information from textual data (conditionally accepted at Sociological Methods & Research) Here's a rundown.. osf.io/preprints/so...

34817

Reposted by @davidbroska.bsky.social

Austin van Loon @austin-van-loon.bsky.social · 18/02/2025

🚨 ACCEPTED AT SMR 🚨 Confused by colleagues who seem to want to study LLMs instead of humans? Frustrated by skeptics (e.g., myself 8 months ago) who dismiss LLMs as a potential source of data on human behavior? Check out our paper for a new way forward: osf.io/j3bnt_v3/

osf.io

OSF

142