AI 中文总结
该研究针对皮尤美国趋势调查面板数据，发现大型语言模型模拟交叉身份时，其响应未呈现真实身份的加总特性，且会系统性舍弃种族、宗教等关键驱动因素，仅能单次代表单一身份。
AI 中文摘要
大型语言模型正越来越多地被用作合成调查受访者，有望以低成本获取稀有的交叉群体。我们针对皮尤研究中心15波美国趋势调查面板中的每一个真实交叉亚群，测试了标准人口统计角色方法，共涉及8个模型的2100万条模拟响应分布。在真实受访者中，亚群观点近似于其单一身份构成部分的加总，但随着身份交叉，其独特性会增长2.5倍。模拟受访者未表现出这种构成特性：在75%-82%的亚群中，单一特征对双特征角色响应的解释力优于加总组合，而第三个特征几乎无贡献。这种失效在我们测试的所有提示策略中均存在。此外，模型保留的特征几乎是随机选择的，仅系统地舍弃种族和宗教——这是观点最强的真实驱动因素。合成样本提供交叉角色，但每次仅代表一种身份。
英文摘要
Large language models are increasingly used as synthetic survey respondents, promising cheap access to rare intersectional populations. We test standard demographic-persona methods against every real intersectional subgroup across 15 waves of Pew's American Trends Panel -- 21 million simulated response distributions from eight models. In real respondents, subgroup opinion is approximately the additive sum of its single-identity components, yet grows 2.5x more distinctive as identities intersect. Simulated respondents show no such composition: a single feature explains a two-feature persona's responses better than the additive combination in 75-82% of subgroups, and a third feature adds almost nothing. This collapse survives every prompting strategy we test. Additionally, the feature models retain is chosen nearly blindly -- except that they systematically discard race and religion, the strongest real drivers of opinion. Synthetic samples offer intersectional personas but represent one identity at a time.
