If you can't prove it...don't adjust for it. Just say you have no basis for an adjustment based on paired sales analysis.
I get what you're saying about that, and I've heard it before in classes and what not, but it never felt right to me. Common sense says that an elevator would generate an immensely positive market reaction from any mobility limited person; certainly not as great as the cost of the feature which is why you dont see more of them. But you can bet your sweet bippy, that a person with a bad hip or in a wheel chair is going to gravitate to a house and with an elevator and open their wallet wider than they would for a house without an elevator. Even if you're not mobility limited, its hard to see how an elevator would be a negative; maybe there's some maintenance, but that's probably proportionate to use so its a not likely a major factor
So, if one of our primary goals is to not be misleading, and you believe there's a difference even though you cant prove it, shouldn't we be making an adjustment, some adjustment, even if we cant prove the magnitude of it?
I mean at the very least you could say "There's no direct market support for an elevator...yada, yada, yada. But since the elevator obviously has value, it was adjusted based on typical Installation-cost/market-value rates for other high value unusual amenities like pools, excessive garage/shop space, etc, etc.
Just seems like doing nothing is more misleading than doing something with minimal basis.