Can large language models stand in for human participants? Many social scientists seem to think so, and are already using "silicon samples" in research. One problem: depending on the analytic decisions made, you can basically get these samples to show any effect you want. THREAD 🧵

#NewPaper is out! With #NucleoMEPOP coauthors Claudia Lagos and @karengheza.bsky.social, led by the amazing Isabel Pavez in @jmcquarterly.bsky.social ❤️ journals.sagepub.com/doi/10.1177/... Thread below 👇

Today's ACIP meeting may use VAERS data to make false claims about vaccine injury. To help sort out what's true from what's false, @stecula.bsky.social and I updated an explainer we wrote several years ago for @us.theconversation.com. theconversation.com/vaccine-deat...

Americans who say there's a lot of discrimination against people who are: In US illegally 57% Transgender 48% Gay or lesbian 37% Black 36% Muslim 34% Hispanic 34% Jewish 30% Immigrants in US legally 29% Women 22% Asian 21% Religious 17% Older 16% Evangelical 13% White 12% Men 10% Atheist 7%

I am pleased to share the latest issue of The International Journal of Press/Politics, which features Special Issue articles from the 2023 #IJPP conference, University of Edinburgh & several open access articles. Link journals.sagepub.com/toc/HIJ/curr...

Missed it in august! But this (with @sumitra.bsky.social) also is out at @bjpols.bsky.social We map how religious beliefs and identities affect the endorsement of falsehoods, and use an experiment in India to test. More to do on the contextual drivers of misinfo, but this is a start. Check it :)

New study examines how people evaluate online ads as "political": source cues, message strength (i.e., persuasive intent), and uncongenial messages. All addressed by the politically relevant media model (PRMM) and why asking if a media text is "political" may not be the best measure.

I collected some materials on critical AI from my perspective; hope it's useful: olivia.science/ai "CAIL is as an umbrella for all the prerequisite knowledge required to have an expert-level critical perspective, such as to tell apart nonsense hype from true theoretical computer scientific claims"

New paper on algorithmic gatekeeping & news engagement out at JOC! Lucien Heitz, @masurphil.bsky.social, @judith-moeller.bsky.social, @vanatteveldt.com & me ran a 1-week long field experiment to test if we can nudge people to read (and learn from) environmental news. doi.org/10.1093/joc/...

For the first time, a special issue of @poqjournal.bsky.social devoted to #qualitative research! The 2025 special issue: Qualitative Research: Advancing the Social & Behavioral Sciences - Articles using a quality approach to study the vulnerable, sensitive topics & QDA bit.ly/POQQRissue

ICYMI: If you’re working or are planning on working with Google Trends Data for research, make sure to read this article by @floriankeusch.bsky.social & co-authors. They explore common pitfalls, incl. construct validity and reliability, and show how to improve data quality to enhance robustness.

Fab researchers at the National Literacy Trust launched new research asking: What does AI mean for the future of literacy? New findings offer insight into how young people & teachers use genAI & whether literacy skills can support a more critical approach More: literacytrust.org.uk/ai-and-literacy

Why did COVID-19 memes turn politicians into both heroes and clowns? In our new study of jokes from five European countries, we show how pandemic humor made leaders look powerful — yet, just as lost as the rest of us. Read how politics became a spectacle through humor: doi.org/10.1177/1367...

💥New: Can we measure trust in scientific publications? ✍️ @jacoates.bsky.social‬ #AcademicSky #OpenScience #PeerReview blogs.lse.ac.uk/impactofsoci...

Young adults almost always online Japan 85% Spain 70% S Korea 68% Turkey 64% US 61% Argentina 59% Israel 57% UK 53% Canada 49% Brazil 47% India 28% Greece 27% Nigeria 15% https://www.pewresearch.org/short-reads/2025/09/08/most-adults-across-24-countries-are-online-at-least-several-times-a-day/

"The gap [in inflation expectations] between Democrats and Republicans is largest for partisans who are the most knowledgeable about politics. Greater numeracy appears to exacerbate the partisan gap, while greater economic knowledge mitigates it." digitalcommons.carleton.edu/cgi/viewcont...

AI deepfakes on social media are fueling calls for stronger protections — and a recent @comatbu.bsky.social survey shows big public support for regulation. How can we balance innovation with trust to keep online spaces safe and truthful in the AI age? www.bu.edu/com/articles...

Highest level of support since Gallup started asking this question a quarter century ago

Happy to share my co-authored piece with Tan Khai Ee, on one of the first empirical studies of Facebook ads in India. We examine how incumbency, personalization & issue focus shape campaign tone in subnational elections: Bihar (2020), West Bengal (2021) & UP (2022).

"both prebunking and situational credibility can effectively counter false information about voter and election fraud but provide some evidence that prebunking can be more effective"

Trump's city-crime narrative leans on encoded stereotypes of largely Black cities. OUT TODAY @polbehavior.bsky.social, @kevintmorris.bsky.social and I show Black cities were also the epicenter of the 2020 election fraud narrative. Republicans and Racially resentful Whites were most susceptible.

saturday morning thoughts: things really don't need to always be easy or quick or efficient. simplifying, reducing, speeding up, homogenizing, standardizing processes/tasks won't always make our lives better. complexity, diversity, struggle, challenge, friction can birth a lot of meaning.

"Designing non-anthropomorphic AI chatbots doesn’t mean making them difficult to interact with. It means stripping away the illusions of personality and cognition that suggest the AI is something it is not. "

“We used data comprised of 24,009 participants nested within 42 countries. Multilevel models largely supported the competing claims of social dominance and social identity theories over SJT.”

New Quinnipiac Poll — 50% of voters say Israel is committing genocide in Gaza. Includes *77%* of Democrats, 51% of Independents, and 20% of Republicans. 60% of voters oppose sending more military aid to Israel. Includes 75% of Democrats, 66% of Independents, and 37% of Republicans.

The findings show that calls for "free speech" often challenge moderation decisions, but reactions differ based on who raises the complaint.   🔗 Explore how these insights could reshape social media practices: journals.sagepub.com/doi/full/10....

"Until we recognise that the debate about AI is not just about what machines can do but also about how humans should value education and knowledge, it will remain mired in confusion"

In case you missed that paper

new in early view at Political Psychology -->

Today we release BAD VAXX-a new game that PREBUNKS misinfo about vaccines. Step into the shoes of "Ann McDotal" who loves to float scary anecdotes & "Dr Forge" the pseudosciensist. We find the game boosts people's ability to better *discern* information manipulation! www.nature.com/articles/s41...

“synthetic sample” is antithetical to not only public opinion research but to the very idea of opinion counting & normative democratic politics which rely on principles such as equality, freedom, and human agency for opinion change, expression, and non-expression. Each opinion poll is also

🎉 Delighted to share new research out today: The Political Consequences of the Mental Load 📄 European Sociological Review doi.org/10.1093/esr/jcaf019 @europeansocreview.bsky.social What happens to political engagement when the “mental load” of running a household gets too heavy? 🧵

From the @theconversation.com today by @maggielevenstein.bsky.social & @johnkubale.bsky.social “Data that taxpayers have paid for and rely on is disappearing – here’s how it’s happening and what you can do about it.”‬ #academicsky #data #umich @icpsr.bsky.social @umisr.bsky.social

'Vending machine' is really useful, because so many of the words available to describe genAI have inherent implications of 'alive', even if only minimally - 'spit', 'regurgitate', 'hallucinate' etc. I have found this very frustrating. (continues)

Have just devoured Jill Lepore's If Then – an amazing book and so well suited to the current moment. So many easter eggs for media/comm folks and an era of the academy now truly unimaginable. So envious of the bookcraft! Thanks @adrishaw.bsky.social for recommending wwnorton.com/books/978163...

Public opinion on social spending in the U.S. has long been known to shift in parallel across policy areas — but why? Here's an article with Steven van Hauwaert and Ryan Carlin that provides some answers, and it's open access: cup.org/3HdCwX3

The narrowing of public information—from book bans to censorship and legal pressures on media—is threatening democracy. @comatbu.bsky.social, we’re tackling this through research, education, and supporting local news. Full letter: sites.bu.edu/crc/2025/08/... #MediaLiteracy #Democracy #LocalNews

Here is a non-paywalled article about the Lancet study. As if I needed another reason to be furious about AI… www.medpagetoday.com/gastroentero...

Want to get a sense of my book with @ekvraga.bsky.social but not ready to commit? @oxfordacademic.bsky.social has released Chapter One free and downloadable for the next month. academic.oup.com/book/60489/c...

🚨 New publication, "Examining Character Assassination in Patient-Clinician Interactions About Chronic Pain" by myself, @timesnewromann.bsky.social, Rachel Tucker, and Madeline J. Moore out today in Journal of Health Communication! Read it here: www.tandfonline.com/eprint/P93NN...

What bums me out most about such doomsdayist claims about the demise of the scientific endeavor is that the exaggerated framing in the abstract and the media coverage is in direct conflict of the study limitations already discussed within the paper.

Academic freedom is under threat. Senior scholars from 11 countries examine the impact of political interference, censorship, and self-censorship in journalism studies. ➡️Read the commentary: tandfonline.com/doi/full/10.... 🌍We invite you to SIGN this article here: bit.ly/afsignatory

"About one in six (15%) social media users (14% of the public overall) say they regularly get health information and advice from social media influencers."

Are AI deepfakes always polarizing? Our 2024 India elections study (N=1,627): News about out-party use of deepfakes didn't increase affective polarization AI approval → lower polarization Impact depends on context — fears may be overstated. @ozankuru.bsky.social Full paper: shorturl.at/oQci2

Our 2nd Delhi School on Political Comm wrapped up with rich discussions & strong participation. Grateful to @dkreiss.bsky.social, plenary speakers, participants, mentors & partners for making it possible. Together, we’re building a vibrant Global South research space. #PolComm @polcomm.bsky.social

interesting. i noted this a few years ago. the problem has gotten worse since then. (misleading y axis on this chart, though , imo). www.economist.com/united-state...