An excellent question, to which there is no perfect answer. Here are some thoughts:
- an issue for me is whether the prospective buyer wants to drive it to "see if he likes it". That's BS. Decide whether you like a specific vehicle before you go looking to buy one. Then you are only checking condition and operation of the specimen at hand. Expect your buyer to do that.
- when I am test driving a car (a specialty car, not a run-of-the-mill beater), I ask the owner to take me for a drive. Allows me to focus on the nuances of the car and how the owner drives it. Only then do I drive it myself. I am driving it to test it, not to see if I like it. Could the same thing be applied to a bike, riding two-up? (I hate the thought, but is it an option?)
- If I was selling a specialty car ( a performance car of some sort) or a unique bike, in principle I would not allow any test pilots. Get in, strap in and pay attention to the ride. I'll show you what it does. When you own it, then you can do whatever you want with it. If it were a bike, I'd show the seller what it does. Find a suitable place and demonstrate hard acceleration and deceleration with gear selection up and down.
- Trust your instinct; if the buyer strikes you in a positive manner, you would be inclined to trust him/her with a test drive. Test drive should be short, around the block sorta thing, not a five mile loop. Hold something of the buyer's; drivers license, car registration and keys, for example. Verify previous bike ownership and riding history best you can. If you are not confident of a buyer's riding ability, I would not allow a test drive.
At the end of the day, it is not entirely reasonable to expect a buyer to pull the trigger without a test drive. You just need to find a way to limit the risk. If the buyer can't appreciate that, maybe you don't have the right buyer.
Just my two-bits worth...