How do you research what EV buyers actually care about vs. what they say they care about?

Classic stated vs. revealed preference problem — we've had good luck with conjoint analysis (specifically choice-based conjoint) because it forces trade-offs instead of letting people say everything matters. Pairing that with behavioral data from actual purchase records or lease agreements really exposes the gaps. Range anxiety always scores lower in surveys than it behaves in real purchase decisions, for example.