Sol Messing

PDF Code Follow Hannah Waight, Eddie Yang, Yin Yuan, Molly Roberts, Brandon Stewart, Josh Tucker and I published a paper in Nature (2026) showing that state-controlled media in LLM training data influences how those models talk about politics.

Early projections for 2024 based on previous Presidential and House returns slighly favor Republicans.

TLDR: [UPDATED SEPT 30] Yesterday, Science published a letter I wrote arguing that there is little evidence of algorithmic bias in Facebook’s feed ranking system that would serve to increase ideological segregation, also known as the “Filter Bubble” hypot…

Last Friday (2023-03-31) Twitter released what it calls “the algorithm,” which appears to be a highly redacted, incomplete part of code that governs the “for you” home timeline ranking system.

TL/DR Summary BlueSky has a chance to dethrone twitter right now, but that path is narrow.

It’s becoming clear that the 2020 polls underestimated Trump’s support by anywhere from a 4-8 point margin depending on your accounting–a significantly worse miss than in 2016, when state polls were off but the national polls did relatively well.

According to the latest polling research, Trump’s chances of hanging on to power beyond 2020 look pretty dismal.

PDF Follow Inspired by Donald Trump’s surprise victory over Hillary Clinton in the 2016 general election, Sean Westwood, Yphtach Lelkes and I set out to interrogate the question of whether elecion forecasts—particularly probablistic forecasts—might create…

PDF Follow On January 17, 2020 my team at Facebook launched one of the largest social science data sets ever constructed.

PDF Follow The Impression of Influence: Legislator Communication, Representation, and Democratic Accountability Princeton University Press, 2015.

Do you remember the night of Nov 8, 2016?

ggpairs My last post railed against the bad visualizations that people often use to plot quantitive data by groups, and pitted pie charts, bar charts and dot plots against each other for two visualization tasks.

Regression models are a cornerstone of modern social science.

This post presents a replication of Messing et al. (2016, study 2), which showed that exposure to darker images of Barack Obama increased stereotype activation, as indicated by the tendency to finish incomplete word prompts---such as “W E L _ _ _ _”---in…

Earlier this month, we published an early access version of our paper in ScienceExpress (Bakshy et al. 2015), “Exposure to ideologically diverse news and opinion on Facebook.”

Yesterday a few of us on Facebook’s Data Science Team released a blogpost showing how candidates are campaigning on Facebook in the 2014 U.S. midterm elections.

Data can often be usefully conceptualized in terms affiliations between people (or other key data entities).

After my post on making dotplots with concise code using plyr and ggplot, I got an email from my dad who practices immigration law and runs a website with a variety of immigration resources and tools.