Reproduce “AI Simulations Grounded in Rich Self-Reports Increase Accuracy and Reduce Distortion” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Joon Sung Park, Carolyn Q. Zou, Jonne Kamphorst, Niles Egan, Aaron Shaw, Benjamin Mako Hill, Carrie Cai, Meredith Ringel Morris and 3 more

The authors test whether giving AI simulations a person’s survey answers, interview, or both improves held-out predictions and population patterns. In a quota sample of 1,052 U.S. adults, combined interview-and-survey input raised accuracy on 150 General Social Survey questions to 72.8%, from 60.4% with demographic input. It closed a median 53.5% of eight distortion gaps, but not hyper-rationality; mismatches remained.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.