Electric vehicles (EVs) have experienced rapid growth in emerging markets; however, aggregate expansion does not necessarily reflect stable adoption intention. This study examines a “perception–intention gap” in Thailand by integrating structured survey data (n = 426) with a large-scale corpus of social media data (over 130,000 posts), which was preprocessed and filtered to 69,132 relevant observations for analysis. A mixed-methods approach was utilized, integrating Partial Least Squares Structural Equation Modeling (PLS-SEM) to evaluate the explanatory sufficiency of theory-driven constructs alongside sentiment analysis and LDA-based topic modeling to identify trends in online speech and results indicate that, despite strong measurement reliability and validity, the TPB/DOI-based structural model exhibits very limited explanatory power (R 2 ≈ 0.018), suggesting weak explanatory adequacy of conventional intention constructs in this context. Conversely, social media analysis demonstrates significant discursive enthusiasm, especially about marketing-related content (perceived marketing effort; PME), indicating elevated visibility and engagement levels. A comparative analysis reveals a distinct disparity between survey-derived intention frameworks and online discourse prominence, suggesting that favorable digital opinion may not align with consistent adoption intent. These findings delineate the boundary conditions of intention-based behavioral models in nascent electric vehicle marketplaces and propose a cross-source diagnostic methodology for identifying discrepancies between discourse and expressed intention and the study offers a methodological perspective that digital sentiment should be viewed as a measure of discourse visibility rather than a direct sign of adoption readiness.
Kularbphettong et al. (Wed,) studied this question.