Christopher Barrie
@cbarrie
Followers
3K
Following
7K
Media
655
Statuses
5K
Assistant Professor of Sociology @nyuniversity
Joined December 2009
We have a new version of our 𝕡𝕣𝕠𝕞𝕡𝕥𝕤𝕥𝕒𝕓𝕚𝕝𝕚𝕥𝕪 paper now on arxiv! This is a significant update that testa *a lot* more data, suggests post-processing techniques, outlines how to compare across models, and tests with new models...
📈〽〰How stable are your LM prompts?〰〽 📈 In this new preprint with @ellipal13 and @pettertornberg, we provide a way to answer this question: https://t.co/sMlw2YzIFP
3
4
25
Yes but can they do it on a rainy night in Grimsby? Also no.
0
0
1
#OpenAccess from @polanalysis - Measuring Media Criticism with ALC Word Embeddings - https://t.co/n1Trwthfnr - @cbarrie, Neil Ketchley, @aasiegel & Mossaab Bagdouri #FirstView
0
9
19
And you can read the paper here:
cambridge.org
Measuring Media Criticism with ALC Word Embeddings
0
0
1
All data and code to replicate our analyses are open-access here:
1
0
1
Taken together, we think our work points to new possibilities for monitoring media freedom in real time—especially in places where there are challenges to the use of more conventional techniques.
1
0
0
And can also be used for types of counterfactual estimation common to causal analytical frameworks. Here, we ask what would have happened to media criticism in the absence of a coup in Egypt by using Tunisia as our comparison case in a synthetic DiD design:
1
0
0
And the method is versatile! 1. We show that the approach: 2. Works with sparse data Generalizes across multiple languages
1
0
0
What we found: In Egypt, criticism surged after Mubarak fell in 2011—then collapsed after the 2013 military coup. In Tunisia, criticism spiked during the revolution and stayed higher during its democratic transition. Stable autocracies? Almost no change.
2
0
0
Our method uses à la carte (ALC) word embeddings to analyze news articles and estimate how critical media coverage is toward executive leaders—across countries, languages, and time periods. h/t @arthur_spirling, Brandon Stewart, and Pedro Rodriguez for developing this method
1
0
2
Why this matters: Media criticism—the ability to question or challenge those in power—is fundamental to democracy. But most existing measures rely on expert surveys or costly indices that miss subtle or short-term changes.
1
0
0
Excited to share our new paper in Political Analysis co-authored with Neil Ketchley, @aasiegel, and Mossaab Bagdouri! 📢 We introduce an open-source, computationally light method for measuring media criticism—a key marker of media freedom and democratization.
2
19
68
I'm delighted to advertise a job as centre manager, helping me build Oxford's new Centre for Advanced Social Science Methods (CASSM). We start this academic year and aim to attract brilliant new faculty in computational, experimental and qual methods. https://t.co/8W0awktaUn
1
8
20
Nothing to worry about just several thousand papers that are now by definition not replicable... On that note, if you'd like to read our latest preprint: "Replication for Language Models" see below
"with GPT-5 we're actually deprecating all our previous models"
5
11
64
quickly becoming one of the most significant & under-discussed dynamics in AI research
Nothing to worry about just several thousand papers that are now by definition not replicable... On that note, if you'd like to read our latest preprint: "Replication for Language Models" see below
1
1
8
Nothing to worry about just several thousand papers that are now by definition not replicable... On that note, if you'd like to read our latest preprint: "Replication for Language Models" see below
"with GPT-5 we're actually deprecating all our previous models"
5
11
64
⚽ 🏠 sorry 4 all of you
0
0
1