Very interesting results. Though, I believe the test is flawed because all the props were driven by the same power system at the rpm/torque that system was capable of producing with the prop selection tested. I don't see the fairness in testing a prop that would normally be ran at 12-16,000 rpm against a prop that operates at 6-8,000 rpm when only one of the props is being ran near it's designed rpm range.
I believe the authors pared down the propellers published to those that should operate through the range of the motor. The raw data, which is available, encompasses far more propeller sizes than is published in the paper.