On a test stand the propellor is basically running stalled, so the drag is at a maximum, and the coefficient of drag doesn't change much with rpm. That is why the hp vs rpm graph is pretty much a straight line. Kavan(Germany) used to sell a set of calibrated test props, which were simply strips of aluminum bar of different widths and lengths. You could get horsepower simply by running the engine on a series of props, measuring the rpm and air conditions(to correct the hp to standard conditions), and plotting hp. vs rpm. A bit more time consuming than a nice dynamometer, but cheap and fairly reliable. You can compare engines using similar test props cut down from commercial props. You won't get calibrated hp., but you can compare engines and measure pretty accurately which one has more hp, and the effects of changes like head clearance, sleeve shimming, etc. If you put the test props on an electric motor and measure the volts and amps at varying rpms you can calibrate them, just like the Kavan props. Use some durable props, not wood, so the performance doesn't get affected too much by dirt and knicks.
Once the plane is in the air the propellor is no longer stalled and measuring prop performance is much trickier. An electric plane with a data logger can get some pretty good data on rpm, volts and amps, much easier than collecting the same data on a glow engine. Once again, old Wild Bill beat us all to the punch. One of his CL articles detailed a method of measuring the speed of the plane at several different altitudes. If you can also measure the rpm with an audio tach (so you know the engines hp isn't changing) you can develop a lift/drag curve for the complete airplane. The big trick though is flying the plane at a constant angle of elevation. It gets really tough with a combat plane that can fly level at 80 degrees. He used the technique to prove that really small rat racers were at a disadvantage because they couldn't fly high enough long enough to pass. Flying high to pass pushed the drag up so high that the plane slowed down. Most rat flyers found the same thing in races. The smallest planes were slightly faster in level flight, but larger planes won races because they could pass if needed.