University
Our group had two O-ring materials and two opinions. Rather than keep arguing, I built a bench: a manifold, two gauges, a needle valve and a stopwatch.
It pressurises a sealed volume to 6 bar and logs the decay by hand at fixed intervals. Not sophisticated, but it is repeatable, and repeatable beat clever here.
Forty runs later we had a leak rate for each material at three temperatures, and the argument was over. The bench is still on the shelf in the lab.
Reading the decay at thirty-second intervals.
Forty runs, written down as they happened.