Consumer knowledge, attitude, and practice toward electric vehicle adoption: a psychometric re-evaluation of the KAP framework with PLS-SEM analysis

Electric vehicle (EV) adoption research has applied the Knowledge–Attitude–Practice (KAP) framework extensively, yet the empirical separability of its three dimensions in EV contexts has not been rigorously examined. This study contributes a psychometric re-evaluation of KAP measurement for EV adoption in the urban Indian consumer context, revealing a central and previously underreported finding: Knowledge, Attitude, and Practice toward EVs may not constitute empirically distinct constructs, as evidenced by Heterotrait–Monotrait (HTMT) ratios exceeding 0.90 for all construct pairs. This construct fusion reframes the theoretical question from whether the sequential KAP logic holds to why the three dimensions remain psychologically fused in low-exposure markets, and what this implies for attitude formation theory, Consumer Learning Theory, and Innovation Adoption Theory. Analyses are conducted on 1,123 authentic survey responses using composite-based Partial Least Squares Structural Equation Modelling (PLS-SEM) with confirmatory validity assessment. Structural associations,rather than causal effects,between the composite dimensions are estimated: the Knowledge–Attitude association ( $$\beta = 0.847$$ , 95% CI [0.821, 0.871]) and the Attitude–Practice association ( $$\beta = 0.761$$ , 95% CI [0.728, 0.793]) are large and robust under bootstrapping, with Attitude accounting for 71.3% of the total Knowledge–Practice relationship. Alternative model comparisons (three-factor, higher-order, single-factor) show that the data are consistent with a unidimensional EV Orientation structure, reinforcing the construct fusion finding. Pathmox segmentation reveals theoretically predicted heterogeneous structural associations across consumers with and without prior EV exposure, grounded in Psychological Distance Theory and Consumer Learning Theory. Common method variance diagnostics, full-collinearity VIF tests, subsample replication, and reverse-causality model tests are reported. These findings challenge prevailing assumptions about KAP’s factorial structure in EV adoption research and provide a blueprint for measurement instrument refinement in nascent-familiarity markets.