From fragmented survey evidence to transferable behavioral models: a framework for electric vehicle adoption
Understanding heterogeneous household responses is essential for modeling electric-vehicle (EV) adoption, yet survey evidence remains fragmented across studies, constructs, samples, and specifications. This study develops an evidence-integration workflow that organizes dispersed findings through a demographic-psychological-purchase-intention (D→P→PI) architecture. Using EV adoption as an application, we establish a common micro-level platform from the China Household Finance Survey, initialize demographic-to-psychological mappings through literature-informed construct alignment assisted by large language models, and refine parameters by reverse calibration against nine EV-adoption studies. Reverse calibration increases the mean cross-study reproduction score from 74 to 86 on a 0-100 scale; conditional leave-one-study-out structural reproduction yields a mean score of 82. The workflow integrates heterogeneous survey findings into a transparent, auditable behavioral layer for travel-behavior and transport-modeling research.
Authors
- Shanyuan Zhang
- Songmin Yu
Institutions
- Tianjin University (CN)
- Cosmote (Greece) (GR)
- Fraunhofer Institute for Systems and Innovation Research (DE)
Publication Details
- Journal
- Transportation Letters
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1080/19427867.2026.2764452
- Primary Topic
- Electric Vehicles and Infrastructure
- Type
- article
- Field-Weighted Citation Impact
- 0.00