Leveraging Interview-Informed LLMs to Model Survey Responses: Comparative Insights from AI‑Generated and Human Data

Mixed methods research integrates quantitative and qualitative data but faces challenges in aligning their distinct structures, particularly in examining measurement characteristics and individual response patterns. Advances in large language models (LLMs) offer promising solutions by generating synthetic survey responses informed by qualitative data. This study investigates whether LLMs, guided by personal interviews, can reliably predict human survey responses, using the Behavioral Regulations in Exercise Questionnaire (BREQ) and interviews from after-school program staff as a case study. Results indicate that LLMs capture overall response patterns but exhibit lower variability than humans. Incorporating interview data improves response diversity for some models (e.g., Claude, GPT), while well-crafted prompts and low-temperature settings enhance alignment between LLM and human responses. Demographic information had less impact than interview content on alignment accuracy. Item-level analysis revealed higher discrepancies for negatively worded questions, suggesting LLMs struggle with emotional nuance. Person-level differences indicated varying model performance across respondents, highlighting the role of interview relevance over length. Despite replicating individual item trends, LLMs faltered in reconstructing the test’s psychometric structure. These findings underscore the potential of interview-informed LLMs to bridge qualitative and quantitative methodologies while revealing limitations in response variability, emotional interpretation, and psychometric fidelity. Future research should refine prompt design, explore bias mitigation, and optimize model settings to enhance the validity of LLM-generated survey data in social science research. The R code and the supplementary materials are available on the OSF platform (DOI:10.17605/OSF.IO/AFQG3).