Curating and sharing high-quality datasets is the foundation of good empirical science.
@ICWSM promotes dataset creation through a dedicated conference track published in the conference's regular proceedings.
Submit your dataset paper by Jan 15th!
CfP: https://t.co/zsdyL8VMZT
Our bias-corrected estimates of support for US presidential candidates based on @X polls were more accurate than traditional polls! 🎊🎉
We estimated 52% for Trump vs 48% for Harris. So far there are 72,641,564 votes for Trump (51.7%) and 67,957,895 for Harris (48.3%)!
Don't believe everything you see on social media.
Informal political polls conducted on X during both the 2016 and 2020 U.S. presidential elections were significantly skewed by questionable votes.
🔗 Details here: https://t.co/FnpSfxXwkB
@manningcics
🚨New Pub Alert 📢
I’m excited to share our newest systematic review!
We analyzed the data quality of hateful communication datasets for automated detection, the targeted identities, linguistic diversity, and involved researchers. Check it out: https://t.co/TRlGrcU5Rw
First paper🔬 on social polls!!!🗳️
Over 100k polls gauged support for 2020 US presidential candidates on @X, and gathered >20m votes. They favored Trump by >10% (on average). Many were used to reinforce voter fraud beliefs and were likely manipulated by bots and fake votes.
🧵1/5
@asbruckman It follows a different solution path to your question, but I found “unpleasant design” to be a refreshing critique to interventions that ignore or antagonize user goals https://t.co/7mRwe5XEGo