@sodalabs.bsky.social
04.03.2025 23:26 — 👍 0 🔁 0 💬 0 📌 0@sodalabs.bsky.social
04.03.2025 23:26 — 👍 0 🔁 0 💬 0 📌 0
With Alexey Chernikov, Caitlin Brown(@caitbrownecon.bsky.social ) and Denni Tommasi (@dennitommasi.bsky.social )
Paper: soda-wps.s3-website-ap-southeast-2.amazonaws.com/RePEc/ajr/so...
Our #AI model, trained to revisit the footage, scored ~70% on those same questions—significantly beating the hardest human calls. The paper explains the method; next, we’ll dig deeper into the broader economics across multiple survey waves. Stay tuned!
04.03.2025 23:26 — 👍 0 🔁 0 💬 1 📌 0
With the recent spotlight on replication in development economics studies( @i4replication.bsky.social )
), data reliability is key. Typically, we trust field surveys, but in tough cases, human survey answers were right only ~50% of the time.
We are thrilled to share our #AAAI25 acceptance! We spent months with @fao.org perfecting yes/no #FoodSafety questions for vendors in Kolkata—translated and filmed on-site. Why film? So we could replay the entire process afterwards and verify every step. #EconTwitter #EconSky
04.03.2025 23:26 — 👍 5 🔁 0 💬 2 📌 0