cs.CYNov 16, 2026

Necessary for What? Auditing Demographic Disclosure to LLM Synthetic Respondents

Authors: Aspen Wang, Yen Yu Shan, Lun-Wei Ku

Organizations: Institute of Information Science, Academia Sinica, Taipei, Taiwan

Abstract

Large language models (LLMs) increasingly answer surveys on behalf of real panelists, and each panelist's recorded demographic attributes (e.g., age, education, ethnicity) are often placed in the prompt on the premise that richer profiles make more faithful synthetic respondents. While data minimization limits personal data to what a purpose requires, a synthetic survey serves several purposes at once, from predicting individual answers to estimating how opinions are distributed and comparing groups. We ask which of these purposes require the respondent's record, and whether improved estimates support sound conclusions about groups. We vary whose profile a model receives, where it comes from, and how many fields it contains, in two U.S. surveys with eleven open-weight and three hosted models, and compare each model with a statistical model fit to the same information. We make four contributions. First, we show that the record adds little to individual predictions and that, for the open-weight models, most of its gain in population estimates remains when it is replaced by census data or a synthetic persona. Second, we find that group comparisons improve only with the respondent's own record, and that even then the models' group differences rest more on the demographic label in the prompt than on the individual respondent, which the usual check of reproducing known group differences does not reveal. Third, we find that the models make groups look more alike than they are, keeping differences in circumstances while flattening or reversing differences in attitudes, most sharply for religious affiliation and ethnic origin, both special categories of personal data. Fourth, we outline an audit that a survey organization can run on questions it has already fielded to decide which estimates may use the record.

Figures & tables

Appendix figures & tables12 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

Jul 6, 2026

Plausible but Not Valid: A Psychometric Audit of LLMs as Synthetic Survey Respondents

Large language models (LLMs) are increasingly used as synthetic survey respondents, but existing evaluations mostly ask whether individual answers look plausible. We ask a stricter, psychometric question: do LLMs preserve the joint distribution, latent structure, reliability, mediation pathways, and demographic effects of real human survey data? We introduce an open Lithuanian organisational-psychology dataset (n=263 employees; three validated instruments; 68 items, 12 subscales) and condition 37 LLMs on real respondent profiles under a persona-disclosure ladder, presentation, reasoning-effort and cross-language ablations, counterfactual demographic swaps, and a memorization probe. The resulting Psychometric Similarity Score (PSS) is anchored against five statistical baselines, a held-out human ceiling, and the real answers of the same respondents. LLMs reproduce the direction of human psychometric relationships, but statistical baselines with no language model match or beat them. The best LLM (PSS 0.81) is level with a Gaussian copula fit only on other respondents (0.82) and below the real answers of the same 100 people (0.89). A simple classifier separates every LLM sample from humans (median AUC 0.99), and verbatim recall of the items is rare (at most 4.7%) and unrelated to PSS. Counterfactual swaps move answers in the direction of human group differences and by a similar amount (role 0.49 and education 0.33 human SD, vs. 0.46 and 0.44 in humans). Gender swaps move them by 0.06 SD. Tucker's phi on UWES falls inside an item-permutation null for 10 of 37 models. Synthetic respondents compress the response scale (within-respondent SD 0.79 vs. 1.19), lose predictive validity on held-out humans (mean R^2=-0.14 vs. 0.28), and produce significant indirect effects on 4 of 7 mediation paths that are null in humans. LLM samples are therefore not a drop-in replacement for human survey data.
Aug 19, 2026

Readable, Faithful, Used: Three Dissociable Properties of Demographic Identity in a Language Model

Large language models are widely used to simulate survey respondents, yet their outputs are homogeneous and unfaithful to real inter-group differences, and whether this reflects what a model knows or uses has remained untested. Using representational similarity analysis against Pew American Trends Panel ground truth, we score demographic read-out locations in Mistral-7B and intervene causally across six attribute types. The internal geometry is faithful: attention-head read-outs dominate the standard residual read-out, reaching selection-corrected ρρ up to 0.63 -- about 70% of the measurement-reliability ceiling -- and one head, L11 H16, is significantly faithful across all six types, though race-based types stay weak and prompt-fragile, replicating in a second model family. Yet causal use does not track fidelity: the clearest causal pathway (p=0.002p=0.002) sits in one of the least faithful types, the most faithful type shows no correction-surviving effect, and full identity swaps in the prompt move predictions by under 2% of their error. A 128-dimensional probe on that head lands 21-31% closer to survey truth than the model's answers, yet recovers almost none of the per-question group ordering. Readable, faithfully arranged, and causally used are three dissociable properties of the same model; treating them as one claim is what keeps the "can LLMs simulate populations" debate unresolved.
Jul 28, 2026

When Synthetic Users Fail: A Cross-Domain Benchmark of LLM-Simulated Human Survey Responses

Large language models (LLMs) are increasingly used as synthetic users, stand-ins for human respondents whose simulated answers feed product, policy, and market decisions. We ask when this substitution is valid and when it fails, and package the answer as an evaluation framework for intelligent synthetic-user systems. A single protocol, run across four models spanning two families and an 8B-to-frontier capability range, is applied to two independent domains of real human-response data: U.S. general social attitudes (General Social Survey) and cross-cultural values (World Values Survey). Every model is benchmarked against a suite of non-LLM baselines fit on held-out human data. Under demographic prompting and the survey-simulation protocols we test, two failures replicate across both domains, all four models, and both families. First, at the individual level no LLM beats even the strongest baseline; on cross-cultural values every model falls well below it, and the gap survives distance-aware and proper scoring. Second, models systematically over-determine demographics, treating identity as far more predictive of attitudes than it is among real people, a distortion present for nearly every question-group combination and robust to a coding-invariant measure. Neither failure is remedied by a larger, more capable model. A decision-impact analysis shows why this matters in practice: on a segment-targeting task the models inflate between-segment gaps two to fourfold, would direct a team to the wrong segment in half of U.S. and most cross-cultural cases, and manufacture segment splits that do not exist in real people. We publicly release the cross-domain benchmark and validation framework, including all code and data, so that teams can determine in advance when synthetic-user evidence is safe for decision support and when it is not.