Upon some reflection on the theory as stated in the video, I no longer believe that a small vacuum chamber can test this effect. It can be argued by proponents of the theory that the craft is pushing off of a wall of the chamber, rather than the atmosphere, and that nothing would be demonstrated either way.
I think this explanation still requires Newton's Third Law as a premise; it merely moves the discussion of action-reaction pairs from rockets and exhaust particles, to exhaust particles and walls. If we're talking about exhaust "pushing" the rocket off of a wall, then we're still talking about Newton's Third. The rocket pushes on the wall, and the wall pushes back. But if Newton's Third Law is correct, then thrust is perfectly understandable via the action-reaction pair of exhaust particles and the rocket that accelerated them. Exhaust particles cannot be accelerated out of the nozzle without the application of a force, and the application of that force must correspond to a force of equal magnitude in the opposite direction.
It's also not clear to me, contrary to Newton's explanation of thrust, how the process of pushing off the wall in this manner works. A particle leaves the nozzle and collides with a wall. Then what? How does the particle of exhaust hitting a wall accelerate the vehicle it just left behind?
I would like to see one more experiment tried, with a barrier that is not attached to the vehicle. For instance, what would happen if he put his outstretched palm following the exhaust pipe at a distance of two inches as the car sped away? Would the car speed up since his hand is more solid than the atmosphere? I think that might be more conclusive evidence.
Do you have any comments on the validity of such an experiment?
I definitely think that this a better experiment than the one in the video. I think the guy in the video did the equivalent of attaching parachutes directly behind a jet engine and then arguing that it proves jet engines don't work: the force of the air on the paper is opposite the direction of thrust.
I'm no physicist, but it sounds like a good experiment to me if you control it rigorously. The conditions of each trial should be identical except for the placement of the barrier.