Synthetic Personalities: How Well Can LLMs Mimic Individual Respondents Using Socio-Economic Microdata?
2026-06-03 • Computers and Society
Computers and SocietyArtificial IntelligenceHuman-Computer Interaction
AI summaryⓘ
The authors studied how to create detailed digital replicas of individual people using existing customer data from sources like surveys and loyalty programs. They tested many ways to build these ‘twins’ using different language models, information amounts, data encoding methods, and reasoning styles. They found that more detailed information improves the replicas up to a point, and using raw past responses instead of summaries helps with accuracy. Their best models predicted answers about 79% correctly, showing that current challenges are more about data quantity and model choices than data design. This means digital twins can be made effectively from the kinds of data companies already have.
Large Language ModelsDigital TwinsMarket ResearchSocio-Economic Panel (SOEP)Shannon EntropyEmbeddingsReasoning ModesCustomer Relationship Management (CRM)Hold-out EvaluationFisher-z Correlation
Authors
Leonard Kinzinger, Jochen Hartmann
Abstract
LLM-based digital twins promise to scale and accelerate market research, but most published twins are either coarse persona bots conditioned on a few demographic questions or detailed individual-level twins built on purpose-collected surveys and interview transcripts. Neither setup speaks to the operationally most relevant case for marketing practice: building detailed individual twins from the pre-existing heterogeneous panel data that firms already accumulate through CRM systems, loyalty programs, and repeat surveys. We construct detailed individual-level twins from the German Socio-Economic Panel (SOEP) and evaluate them across a $3 \times 5 \times 2 \times 2$ construction-method grid that covers three open-weights LLMs, five cumulative information depths ranked by normalized Shannon entropy, two embedding methods, and two reasoning modes, scoring over 2.1 million twin responses on 500 participants and 183 held-out questions. Twin quality rises with information depth but with diminishing returns past the 75 percent entropy quartile, which acts as a cost-efficient Pareto point relative to the best-performing 100 percent cells. Switching the embedding from a narrative persona summary to a raw dialog history of past responses raises hold-out accuracy in every model-by-reasoning cell at the 100 percent depth, while an explicit thinking mode raises rank-order correlation without moving accuracy. Best-cell accuracy reaches 78.8 percent and Fisher-$z$ correlation reaches $r = 0.590$ on the SOEP held-out evaluation set. The findings suggest that twin-based market research is no longer gated by data design, but by item volume, model selection, and a small set of construction-level decisions that this paper now maps.