New paper with @LamarPierce1 and Scott Wiltermuth came out today in Organization Science! We show paying MORE initially for durable good like a house ($400k vs $399k) may make sense because of intergenerational anchoring around round numbers. Check it out! pubsonline.informs.org/doi/abs/10.128…

This paper is ushering in using AI to better do social science. I predict that the reverse applies too! As AI capabilities further advance, quantitative social science methods will be increasingly relevant to understanding artificial agents, and their interactions

Hey look - it's me talking about one of my favorite research topics!

Chris Bail (chris_bail_duke 🧵) Yay! Thanks for having me! And for the interested, a preprint of the paper, forthcoming at Political Analysis, can be found here: arxiv.org/abs/2209.06899 // Nancy Fulda Joshua Gubler David Wingate Ethan Busby Chris Rytting

Such a fantastic conference! Can’t wait to get together with this group again next year!

It’s officially online! Our paper, “Out of One, Many: Using Language Models to Simulate Human Samples” is published Political Analysis. This paper is for anyone who is trying to figure out what ChatGPT and other recent Artificial Intelligence advances mean for social science research

This is such a cool paper and part of a very timely research agenda. Follow this work🚀

1/ Can AI chat assistants improve conversations about divisive topics such as gun control? I was very fortunate to join a team of outstanding researchers who designed an experiment intended to answer this question: arxiv.org/abs/2302.07268

2/ Our team recruited 1,574 Americans to participate in an online chat with someone who does not share their view. After a few messages, half of these people were offered rephrasing suggestions via GPT-3—which was prompted using evidence-based insights from social science.

3/ People whose chat partners accepted rephrasings of their messages by GPT-3 rated conversations as higher quality and less divisive than those in a control condition where no such intervention was offered.

4/ Much more research is needed to determine whether this intervention could scale to large platforms. But we hope our work will inspire future studies of how AI can be used to incentivize more pro-social behavior on social media.

5/ If you like the sound of this research, please follow my truly amazing collaborators Lisa Argyle Ethan Busby Joshua Gubler Vin Howe Chris Rytting David Wingate who are pioneers of research in this area and were very generous to let me come along for the ride!

So excited to share our latest project merging AI and social science to improve political conversations!

New article PNASNews with Joshua Gubler , Lisa Argyle , Chris Bail (chris_bail_duke 🧵) , David Wingate , Taylor Sorensen , Vin Howe, and Chris Rytting on how AI tools can improve democratic discourse. It's open access at this link: doi.org/10.1073/pnas.2… We explore how an AI chat assistant can act...

🚨Check out our new article: can GPT-3 facilitate more productive conversations about politics?🚨

As I've said many times before, so much of the credit for this work goes to my amazing collaborators Lisa Argyle Vin Howe David Wingate Joshua Gubler Ethan Busby Chris Rytting make sure to check out their other work, it's amazing!

Congratulations to Alex Coppock whose book, “Persuasion in Parallel,” has won the Robert E. Lane Best Book Award from the APSA Political Psychology Section! Thanks to @wustlPOLISci and Joshua Gubler for serving on the committee.

Congratulations to Alexandra Filindra whose book, “Race, Rights, and Rifles,” has received honorable mention for the Robert E. Lane Best Book Award from the APSA Political Psychology Section! Thanks to @wustlPOLISci and Joshua Gubler for serving on the committee.

Interested in using Large Language Models for Social Science Research? Here are links to two additional, recently-published articles by our team: pnas.org/doi/10.1073/pn… journals.sagepub.com/eprint/VAN3ZEF…