Proprietary models/platforms may block the replicability part. However, reliability and robustness are a measurement validity problem; social scientists are trying to find a consensus on it. I will be presenting on this issue tomorrow, and my work is auditing the instruments we use to audit models.

Important food for thought for the #IC2S2 community, as we convene the next several days. How replicable/reliable/robust is the science we do? How do we build collective capacity and frameworks that facilitates a science that is cumulative?

J. Nathan Matias@natematias.bsky.social · last wk.

When people seek guidance from science, they expect it to work more than once. I have TWO new articles on commercial forces & proprietary models that make science less reliable, with pointers to communities and ideas that might help. This post offers a summary: citizensandtech.org/2026/07/comm...

Heading to Columbus for ISA 2026 ✈️ This 2×2 has been living in my head for a while. Presenting tomorrow on why foreign policy experts engage across borders but still cluster along national lines. 650K+ interactions to untangle 😄 #ISA2026 #PoliticalNetworks

Great talk by John Wihbey @harvard.edu @shorensteinctr.bsky.social on his new book!

Negative sentiments about the opposing political party vary greatly among individuals. In POQ, Stapleton and Wolak find that those with high internal efficacy also possess higher levels of affective polarization. Read more: doi.org/10.1093/poq/...

Excited to share that @burakozturan.bsky.social and I hosting free workshops on using LLMs in social science research. We would love to have you join us if you're grad/early career and can make it to campus for both sessions.. for more: bliss.boston nov 14 & 21 at northeastern, apps close oct 10 👇

Fall 2025 Workshops on using LLMs for Social Sciences

BLISS (Boston LLMs Initiative for Social Sciences) is organizing a funded 2-day workshop series teaching social science researchers how to use large language models in their work, discuss limitations ...

forms.gle

Great work from @cetaceanneeded.bsky.social (on the job market!) using starter packs to examine joining and following patterns on BlueSky. React valuable insights for anyone looking to study this platform #polnet2025 #pacss2025

If you’re curious about what happened on the old platform after our migration, come by my presentation tomorrow (Wednesday) at 2:30 PM in Vingen 3+4. #ic2s2 @ic2s2.bsky.social

Useful data alert! Looking for audience measures of domains, based on party, gender, race, geography? Check out our new paper, led by @yang3kc.bsky.social, that uses a large panel of Twitter data linked to voter data to make demographic inferences at domain level. www.nature.com/articles/s41...

DomainDemo: a dataset of domain-sharing activities among different demographic groups on Twitter - Scientific Data

Scientific Data - DomainDemo: a dataset of domain-sharing activities among different demographic groups on Twitter

nature.com

“Abandoning dictionaries and embracing mechanized writing would erode our capacity for collective identity quite as much as the ability to express ourselves.”

Opinion | Don’t Throw Your Dictionary Away

nytimes.com

New Pew data captures striking divergence in U.S. favorability shifts (2024-2025): While 16 countries register significant declines—notably Mexico (-32pts) and Sweden (-28pts)—Turkey moves against the tide with a 7pt gain (18%→25%). Trump effect? Full report: pewrsr.ch/4kFqPH0

Thanks to all who joined our conversation on Social Capital in the Age of AI and the screening of Join or Die. 25 years after Bowling Alone, we explored how tech is reshaping trust, connection & civic life in the algorithmic era. @danielaldrich.bsky.social @hamishgibbs.bsky.social socialurban.net

Repost if you’ve participated in a Summer Institute in Computational Social Science. Let’s get #SICSS Bluesky going!

A new article by me and coauthors. Based on our experience as computational social scientists in academia, private, and public sectors, we explained how to train social science PhD as computational social scientists in three steps. twitter.com/CUP_PoliSci/... (open access)

Okay this is now up and running. Tag "cssky" in your posts to publish to the computational social science feed. bsky.app/profile/did:...