Testing procedure
 

What would be the best testing procedure to find significant results from aero mods? The other day, I was trying to test some mods made out of cardboard and duct-tape, and the results were inconclusive. I did A-B-A testing with the B runs as the mod and A being no change to the vehicle and did two A runs, followed by three B runs, and ending with two A runs. The changes in fuel economy seemed negligible and inconsistent. I think the natural variability in each of the runs caused there to be no significant difference. Would coast-down testing be better than just watching the fuel economy display (Scanguage) while holding a constant speed? Would I be able to see results better? Is there another way to test? I don't have cruise control in my car so the speed that I hold while using the fuel economy method is only as good as my ability to slightly alter my foot position as needed.
If someone could detail coast-down testing in their response, I would greatly appreciate it.

What kind of speed were you driving at for the runs? Things like wind, slight undetectable grades on different parts of roads, etc, will make testing difficult. You should try a high speed coast-down test also, where you get to a high speed, go neutral, and time the drop from X speed to Y speed.

I was going at 55 mph. The average wind speed was 8 mph. I think the wind might have been a huge factor.

I'm sure wind was a factor, but higher speeds could also be helpful. 75mph! For science!