If I may for a moment.....

Thrust only tells you static performance of an engine/prop combination. In a very indirect way you measure rpm.

You want to know what the engine is doing, and how it performs under a defined load.

Power of the engine known by knowing load and the rpm it turns that load.

An easy known load is a prop. APC props in particular are good because they are fairly consistant blade to blade
For your tests, you would actually use the very same prop.

An array of say 5 different size props might allow you to plot a performance curve based on load and rpm.

Put the engine on a solid test stand.
Pick a prop. Run the engine. Measure the rpm that prop turns on that engine, and what it also turns on your sample of engines.
The closer together you can complete the tests, the less temperature/weather will come in to play.

Exhaust system may come into play, and may sometimes be considered a variable.

Keep in mind, for a car engine, it is designed for 20K+ rpm and higher operation. Has little torque in its native state. So you will have to select sample props that allow the engine to get up to its usual operating rpm, or desired rpm range.

Also to consider here....
As aero guys, we are usually dealing with steady state, full throttle measurements. Cars rarely run at full throttle or max RPM for extended periods of time.

For a car, boat or even helicopter, engine acceleration is something that is often important. That is why the aero tuned pipes are designed differently than a car or heli pipe. Boats do tend to run WOT quit a bit, so the marine pipes are more similar to aero pipes.

What I might suggest, is looking into one of the Eagle Tree data systems. Combine that with your test stand and prop testing. Hands-free measurements and detailed recording of everything you want to know on engine performance.

They make great data collection systems for aircraft, cars and boats that are good in engine rpm and acceleration plots. You can also compair it against throttle position to see throttle response.

You can measure rpm, and plot that rpm vs time (acceleration) - and get a reference on torque and throttle response.

Worth noting, these will not be concrete HP or torque numbers, but if you are as an experiement evaluating and compairing the capabilities of a number of engines, you will get good, real-world, compairable, and consistant information.

Bob