Distribution-First Population Simulation: Collapse, Calibration, and Recall in Non-WEIRD LLM Persona Modeling

Tools that simulate a town or a market with language models often run every resident as a separate agent, each one a call to the model with its own persona.¹ This eight-page preprint by Gürkan Özkan of Istanbul Technical University tests what that does to the spread of opinion in the population.¹ It uses real microdata from the Turkish wave of the World Values Survey, a long-running international survey of values, because most such tests use American, English-language data.¹ It was posted to arXiv in July 2026 and has not been peer reviewed.¹

Each of 2,414 real respondents became a character, and the model answered for each one separately.¹ Across four scenarios and five random seeds the agents crowded onto the model's default answer: the share giving the most common answer rose from 0.36 to 0.69, and the measured diversity of answers roughly halved.¹ The paper calls this collapse and finds it worst where a question has one obviously "rational" answer.¹ The result showed on Qwen and again on GLM.¹

The alternative tested is to ask the model once for the whole population's distribution of answers (a technique called verbalized sampling) and then hand answers out to the characters.¹ That raised fidelity by 7 to 10 points, but overshot: the answers went from too narrow (about half the human spread) to too wide (about 1.3 times).¹ In a booking task on one model, 80% of personas chose the cheapest option; income shifted that only a little, with the comfortable option rising from 0% of low-income personas to 32% of high-income ones.¹

An election backtest shows two different failures. Predicting a past national election in Turkey from people's values, the one-call distribution matched the official result closely, but the author attributes that to the model remembering a well-known result, not simulating it.¹ The agent-per-person route gave the largest party 88.0% of the vote against a true 42.6%.¹ The author's summary: "one remembers, the other collapses; neither simulates."¹ The paper also finds that claims about subgroups and individuals were contaminated by the model's recall.¹ It is a single-author preprint on one country, and its figures have not been independently checked.