arXiv cs.CY· Lai Wei, Yifei Zhu·· 11 小时前AI 评分34
研究：用 AI 预测增强而非取代社会调查
Using AI Predictions to Augment Rather Than Replace Surveys
AI 导读
一项基于 prediction-powered inference（PPI）框架的研究发现，AI 预测无法取代社会调查，但能提升其统计效率。在 GSS 和 NHIS 的 83 项社会结果上、以 ACS 为辅助数据，监督式预测器估计均值时带来约 7-12% 的有效样本量增益，LLM 增益较小但仍为正。
正文 · 原文
Abstract:LLMs can predict human survey responses with nontrivial accuracy. What this means for the future of social surveys remains hotly contested. Optimists anticipate a coming era of synthetic respondents; skeptics see a fragile and overhyped substitute. This article advocates a position: AI predictions cannot replace surveys, but they can augment them. Using prediction-powered inference (PPI), a statistical framework rooted in survey methodology and known in sociology as "mixed subjects design", the study tests whether AI predictions can improve the statistical efficiency of social surveys when combined with census-scale auxiliary information. Applying the framework to 83 social outcomes in the General Social Survey (GSS) and National Health Interview Survey (NHIS), with American Community Survey (ACS) being the auxiliary file, we find that supervised predictors yield effective-sample-size gains of about 7-12 percent when estimating the mean, while LLM gains are smaller but positive.
| Subjects: | Computers and Society (cs.CY) |
| Cite as: | arXiv:2611.06743 [cs.CY] |
| (or arXiv:2611.06743v1 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2611.06743 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yifei Zhu Mr. [view email]
[v1]
Wed, 4 Nov 2026 16:24:24 UTC (2,874 KB)
来源：arXiv cs.CY · arxiv.org